<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>NotProvided.eu</title>
<atom:link href="https://www.notprovided.eu/feed/" rel="self" type="application/rss+xml"/>
<link>https://www.notprovided.eu/</link>
<description>Enterprise and technical SEO by Jan-Willem Bobbink</description>
<lastBuildDate>Fri, 13 Nov 2026 12:34:29 GMT</lastBuildDate>
<language>en</language>
<item>
<title>GSC Terminal: read Search Console like a trading chart</title>
<link>https://www.notprovided.eu/gsc-terminal-read-search-console-like-a-trading-chart/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1356</guid>
<pubDate>Fri, 13 Nov 2026 12:34:29 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[This morning John Mueller dropped a one-line question in Slack: “Is there a Search Console → Trading View adapter?” I checked. There isn’t. Plenty of extensions add line and bar charts to Search Console, and TradingView’s only route for custom data (Pine…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">This morning John Mueller dropped a one-line question in Slack: &#8220;Is there a Search Console → Trading View adapter?&#8221;</p>



<p class="wp-block-paragraph">I checked. There isn&#8217;t. Plenty of extensions add line and bar charts to Search Console, and TradingView&#8217;s only route for custom data (Pine Seeds) no longer accepts new repositories. But the question stuck, because he&#8217;s right that it should exist.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="410" height="90" src="https://www.notprovided.eu/assets/00-john-mueller-question.png" alt="" class="wp-image-1357" srcset="https://www.notprovided.eu/assets/00-john-mueller-question.png 410w, https://www.notprovided.eu/assets/00-john-mueller-question-300x66.png 300w" sizes="auto, (max-width: 410px) 100vw, 410px" /></figure>



<p class="wp-block-paragraph">The Search Console performance chart is a line that goes up and down every weekend. That is most of what it tells you. Is last week&#8217;s dip a trend break, or just Sunday? Did the core update hurt you, or did it only make your traffic more volatile? A single smoothed line can&#8217;t answer either question.</p>



<p class="wp-block-paragraph">Traders solved this decades ago. A candlestick chart shows where a period opened, where it closed, and how far it swung in between. Add a couple of moving averages and you see trend changes before they show up in your monthly report.</p>



<p class="wp-block-paragraph">So I built it the same day: <strong>GSC Terminal,</strong> a Chrome extension that replaces the Search Console performance report with a TradingView-style chart. Same property, same filters, same data. Just a far better way to read it.</p>



<h2 class="wp-block-heading">What it does</h2>



<p class="wp-block-paragraph">Open any performance report in Search Console. A second after Google draws its own chart, GSC Terminal takes over the page below the Search Console header.</p>



<ul class="wp-block-list">
<li><strong>Same view, longer history.</strong> It picks up the property, report, search type and every filter you set in Search Console, and loads up to 16 months of daily data for it.</li>



<li><strong>Candles, Heikin Ashi, line or area</strong> on daily, weekly or monthly bars.</li>



<li><strong>Clicks, impressions, CTR and position.</strong> Position uses an inverted scale, so up and green always mean better.</li>



<li><strong>Indicators:</strong> moving averages with crossover signals, Bollinger Bands, RSI, a volume pane and a year-over-year overlay.</li>



<li><strong>Event markers</strong> for your own notes (&#8220;core update&#8221;, &#8220;migration&#8221;) plus Search Console&#8217;s own data annotations.</li>



<li><strong>The new Generative AI report</strong> works too, which has no API equivalent.</li>



<li><strong>No OAuth, no Google Cloud project, no API quota.</strong> It uses the session you are already logged in with, including every <code>/u/N/</code> account.</li>
</ul>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="575" src="https://www.notprovided.eu/assets/01-weekly-clicks-candles-1024x575.png" alt="" class="wp-image-1358" srcset="https://www.notprovided.eu/assets/01-weekly-clicks-candles-1024x575.png 1024w, https://www.notprovided.eu/assets/01-weekly-clicks-candles-300x169.png 300w, https://www.notprovided.eu/assets/01-weekly-clicks-candles-767x431.png 767w, https://www.notprovided.eu/assets/01-weekly-clicks-candles-1536x863.png 1536w, https://www.notprovided.eu/assets/01-weekly-clicks-candles-2048x1151.png 2048w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<h2 class="wp-block-heading">Installing it</h2>



<p class="wp-block-paragraph">GSC Terminal isn&#8217;t in the Chrome Web Store, so you load it as an unpacked extension. It takes two minutes. Open source via <a href="https://github.com/jbobbink/gsc-terminal">https://github.com/jbobbink/gsc-terminal</a></p>



<ol class="wp-block-list">
<li>Download the zip: <a href="https://www.notprovided.eu/area51/gsc-terminal-1.2.0.zip" target="_blank" rel="noopener">gsc-terminal-1.2.0.zip</a>.</li>



<li>Unzip it somewhere permanent, for example <code>Documents/gsc-terminal</code>. Chrome loads the extension from that folder, so don&#8217;t delete it afterwards.</li>



<li>Go to <code>chrome://extensions</code> and switch on <strong>Developer mode</strong> (top right).</li>



<li>Click <strong>Load unpacked</strong> and pick the folder that contains <code>manifest.json</code>.</li>



<li>Optional: pin GSC Terminal via the puzzle-piece icon.</li>



<li>Reload any Search Console tabs you already had open. Extensions don&#8217;t run in tabs that were open before installing.</li>
</ol>



<p class="wp-block-paragraph">It works in Google Chrome. It should also run in other Chromium browsers, but I only tested Chrome.</p>



<p class="wp-block-paragraph">Updating works the same way: replace the folder contents, click the reload icon on the extension&#8217;s card, and reload Search Console.</p>



<h2 class="wp-block-heading">Using it inside Search Console</h2>



<p class="wp-block-paragraph">The workflow is simple: Search Console decides <strong>what</strong> you look at, GSC Terminal decides <strong>how</strong> you look at it.</p>



<ol class="wp-block-list">
<li><strong>Open a performance report.</strong> Search results, Discover, News or Generative AI. The terminal takes over automatically.</li>



<li><strong>Need a different filter?</strong> Click <strong>◧ GSC</strong> in the toolbar, or press <strong>Alt+T</strong>. You&#8217;re back in the normal report. Set a query, page, country or device filter with Google&#8217;s own controls.</li>



<li><strong>Click 📈 Terminal</strong> (bottom right). The terminal reloads with the new filters. The active filters show next to the site name, for example <code>query ∋ empleo</code>.</li>



<li><strong>Switch property</strong> with the dropdown at the top. It navigates Search Console to that property, so you stay in the same report.</li>
</ol>



<p class="wp-block-paragraph">Once you switch back to the normal view, the terminal stops taking over pages automatically until you click 📈 Terminal again. That way it doesn&#8217;t jump in while you&#8217;re building a filter.</p>



<p class="wp-block-paragraph">Inside the terminal:</p>



<figure class="wp-block-table"><table class="has-fixed-layout"><tbody><tr><th>Control</th><th>What it does</th></tr><tr><td>Clicks / Impr / CTR / Pos</td><td>Metric on the chart. Keys <code>c</code>, <code>i</code>, <code>r</code>, <code>p</code>.</td></tr><tr><td>1D / 1W / 1M</td><td>Bar size. Keys <code>1</code>, <code>2</code>, <code>3</code>.</td></tr><tr><td>Candles / HA / line / area</td><td>Chart type. HA is Heikin Ashi: smoothed candles that make trend runs easier to see.</td></tr><tr><td>Vol, MA, BB, RSI, YoY, Sig, Ev</td><td>Toggle indicators, crossover signals and event markers.</td></tr><tr><td>3M / 6M / 12M / 16M</td><td>History loaded.</td></tr><tr><td>⋯ menu</td><td>Export the candles as CSV, manage events, load demo data.</td></tr></tbody></table></figure>



<p class="wp-block-paragraph">Double-click any bar to add an event, such as &#8220;Core update&#8221; or &#8220;Title rewrite&#8221;. Events are stored per property in your browser.</p>



<h2 class="wp-block-heading">How to read the candles</h2>



<p class="wp-block-paragraph">A stock candle summarises trades within a period. Search data has no trades, so I had to define what open, high, low and close mean for SEO. The goal: the body shows the trend, the wicks show the volatility.</p>



<figure class="wp-block-table"><table class="has-fixed-layout"><tbody><tr><th>Part</th><th>Meaning in GSC Terminal</th></tr><tr><td>Close</td><td>The value for the bar. On weekly and monthly bars, clicks and impressions are <strong>averages per day</strong>, so a 4-week month and a 5-week month compare fairly. CTR is total clicks / total impressions; position is weighted by impressions.</td></tr><tr><td>Open</td><td>The previous bar&#8217;s close. A green body means this week beat last week.</td></tr><tr><td>High / low (wicks)</td><td>The best and worst single day inside the bar.</td></tr><tr><td>Volume</td><td>Total impressions in the bar (total clicks when you chart impressions).</td></tr><tr><td>Grey bar</td><td>Data Google still marks as incomplete. Don&#8217;t panic about it yet.</td></tr></tbody></table></figure>



<p class="wp-block-paragraph">Two patterns to watch:</p>



<ul class="wp-block-list">
<li><strong>Long wicks, small bodies.</strong> Your weekly average is stable, but the days inside the week swing hard. On a B2B site that&#8217;s usually weekday versus weekend. When the wicks suddenly get longer without a seasonal reason, something changed in how Google serves you.</li>



<li><strong>A run of same-coloured bodies.</strong> Three or four red weekly candles in a row is a trend, not noise. That&#8217;s the moment to dig in, not when the monthly report lands.</li>
</ul>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="575" src="https://www.notprovided.eu/assets/03-weekly-position-inverted-1024x575.png" alt="" class="wp-image-1359" srcset="https://www.notprovided.eu/assets/03-weekly-position-inverted-1024x575.png 1024w, https://www.notprovided.eu/assets/03-weekly-position-inverted-300x169.png 300w, https://www.notprovided.eu/assets/03-weekly-position-inverted-767x431.png 767w, https://www.notprovided.eu/assets/03-weekly-position-inverted-1536x863.png 1536w, https://www.notprovided.eu/assets/03-weekly-position-inverted-2048x1151.png 2048w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="575" src="https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-1024x575.png" alt="" class="wp-image-1360" srcset="https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-1024x575.png 1024w, https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-300x169.png 300w, https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-767x431.png 767w, https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-1536x863.png 1536w, https://www.notprovided.eu/assets/04-weekly-clicks-heikin-ashi-2048x1151.png 2048w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<h2 class="wp-block-heading">Indicators that are actually useful for SEO</h2>



<p class="wp-block-paragraph">I don&#8217;t believe in technical analysis for predicting stock prices. Applied to your own search data it&#8217;s a different story: the indicators simply make trends and anomalies visible faster.</p>



<ul class="wp-block-list">
<li><strong>Moving averages (MA).</strong> A fast and a slow average per bar size: 7/28 days on daily bars, 4/13 weeks on weekly bars, 3/12 months on monthly. On daily bars the windows cover whole weeks, so weekend dips cancel out instead of dragging the line around.</li>



<li><strong>Crossover signals (Sig).</strong> A <strong>GX</strong> marker appears when the fast average crosses above the slow one, <strong>DX</strong> when it crosses below. On weekly clicks a DX is a solid trigger to check what happened. Change the MA periods under Filters if they are too sensitive for your site.</li>



<li><strong>Bollinger Bands (BB).</strong> A 20-bar average with a band of two standard deviations around it. A close outside the band is an anomaly worth a look: a ranking jump, a lost featured snippet, a tracking issue.</li>



<li><strong>RSI.</strong> Momentum on a 0 to 100 scale. A rising RSI while clicks still look flat often means growth is starting before it shows in the totals.</li>



<li><strong>YoY overlay.</strong> The value from exactly 52 weeks earlier as a dashed line, aligned on weekday. The legend shows the year-over-year change for whatever bar you hover. You need 16 months loaded to get a full year of overlap.</li>



<li><strong>Events (Ev).</strong> Your own markers plus Search Console&#8217;s annotations. Put a core update or a deploy on the chart and the cause-and-effect discussion with a client gets a lot shorter.</li>
</ul>



<p class="wp-block-paragraph">My default setup: weekly candles, MA and Sig on, volume on, YoY on. For a migration or a core update I switch to daily bars with Bollinger Bands.</p>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="575" src="https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-1024x575.png" alt="" class="wp-image-1361" srcset="https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-1024x575.png 1024w, https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-300x169.png 300w, https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-767x431.png 767w, https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-1536x863.png 1536w, https://www.notprovided.eu/assets/02-daily-clicks-bollinger-rsi-2048x1151.png 2048w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<h2 class="wp-block-heading">The Generative AI report</h2>



<p class="wp-block-paragraph">In June 2026 Google added a separate performance report for AI Overviews and AI Mode. It only shows impressions, and as far as I can tell there&#8217;s no way to get it through the Search Console API.</p>



<p class="wp-block-paragraph">GSC Terminal reads it straight from the page, so it works there too. On the AI report the chart switches to impressions, the other metrics are greyed out, and the volume pane shows impressions. History starts when Google started collecting the data, around mid-May 2026 for the properties I checked.</p>



<p class="wp-block-paragraph">For GEO work this is the most interesting view in the tool. AI impressions are volatile, and a weekly candle with a 4/13 moving average separates a real shift in AI visibility from a noisy week far better than Google&#8217;s line does.</p>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="575" src="https://www.notprovided.eu/assets/05-generative-ai-impressions-1024x575.png" alt="" class="wp-image-1362" srcset="https://www.notprovided.eu/assets/05-generative-ai-impressions-1024x575.png 1024w, https://www.notprovided.eu/assets/05-generative-ai-impressions-300x169.png 300w, https://www.notprovided.eu/assets/05-generative-ai-impressions-767x431.png 767w, https://www.notprovided.eu/assets/05-generative-ai-impressions-1536x863.png 1536w, https://www.notprovided.eu/assets/05-generative-ai-impressions-2048x1151.png 2048w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<h2 class="wp-block-heading">How it works, and the catch</h2>



<p class="wp-block-paragraph">Search Console&#8217;s interface loads its chart data through an internal endpoint. When you open a report, the extension notes which request the page made: property, report type, search type and filters. It then makes that same request with a longer date range, inside your own logged-in session. The chart itself is TradingView&#8217;s open-source <a target="_blank" rel="noopener" href="https://github.com/tradingview/lightweight-charts">Lightweight Charts</a> library.</p>



<p class="wp-block-paragraph">That design has three consequences:</p>



<ul class="wp-block-list">
<li><strong>Your data stays in your browser.</strong> Nothing goes to a server of mine. There&#8217;s no account, no tracking, and the extension only requests what the page could already load.</li>



<li><strong>It sees exactly what you see.</strong> Every property and every <code>/u/N/</code> Google account you have access to works, without setting up OAuth.</li>



<li><strong>It depends on an undocumented format.</strong> If Google changes how the interface loads its data, the terminal shows an error until I update it. The normal report keeps working, and you can always switch back with Alt+T.</li>
</ul>



<p class="wp-block-paragraph">The extension also has an optional API mode with your own OAuth client, which runs in a separate tab. The README in the zip explains how to set that up.</p>



<h2 class="wp-block-heading">Try it</h2>



<p class="wp-block-paragraph">Download GSC Terminal here: <a target="_blank" rel="noopener" href="https://www.notprovided.eu/area51/gsc-terminal-1.2.0.zip">gsc-terminal-1.2.0.zip</a>. It&#8217;s free, and it never sends your data anywhere.</p>



<p class="wp-block-paragraph">It started as an answer to John&#8217;s question, but it I already see some nice ways of using it usefully.</p>
]]></content:encoded>
</item>
<item>
<title>ChatGPT doesn’t know your 2027 health insurance premium. Now it does.</title>
<link>https://www.notprovided.eu/chatgpt-doesnt-know-your-2027-health-insurance-premium-now-it-does/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1351</guid>
<pubDate>Fri, 30 Oct 2026 10:00:34 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[Millions of Dutch people now ask ChatGPT first, before they search. In March 2026, Newcom counted 7.2 million Dutch people aged 18 to 65 using AI in daily life, up from about 5 million a year earlier. ChatGPT is by far the…]]></description>
<content:encoded><![CDATA[
<h2 class="wp-block-heading">The Netherlands asks ChatGPT</h2>



<p class="wp-block-paragraph">Millions of Dutch people now ask ChatGPT first, before they search. In March 2026, Newcom counted 7.2 million Dutch people aged 18 to 65 using AI in daily life, up from about 5 million a year earlier. ChatGPT is by far the most-used tool (<a target="_blank" rel="noopener" href="https://www.emerce.nl/nieuws/72-miljoen-nederlanders-gebruiken-ai-hype-vaste-praktijk">Emerce</a>). In September 2026 it handled 82.94% of AI chatbot visits in the Netherlands (<a target="_blank" rel="noopener" href="https://gs.statcounter.com/ai-chatbot-market-share/all/netherlands">Statcounter</a>).</p>



<p class="wp-block-paragraph">Now it is November, and every Dutch household gets the same question: should I switch health insurance? Insurers publish their 2027 premiums by 12 November, and you have until 31 December to switch. That makes it one of the biggest yearly money decisions people bring to a chatbot.</p>



<p class="wp-block-paragraph">The problem: ChatGPT does not have this year&#8217;s data.</p>



<h2 class="wp-block-heading">Why ChatGPT gets it wrong</h2>



<p class="wp-block-paragraph">People ask for the latest numbers, and the model answers from old ones. A language model knows what was in its training data, and that stops at a cutoff months before you ask. The 2027 premiums did not exist when any current model was trained.</p>



<p class="wp-block-paragraph">This is not only about premiums. The market itself changes every year:</p>



<ul class="wp-block-list">
<li>ZEKUR stops on 1 January 2027 and its customers move to Univé (<a href="https://www.unive.nl/zorgverzekering/zekur/veelgestelde-vragen" target="_blank" rel="noopener">Univé</a>).</li>



<li>Anderzorg merges into Menzis on 1 January 2027 (<a href="https://www.anderzorg.nl/menzis" target="_blank" rel="noopener">Anderzorg</a>).</li>



<li>Contracts between insurers and hospitals, physiotherapists or mental health providers change every year. With a budget policy, that contract decides whether you pay the bill yourself.</li>
</ul>



<p class="wp-block-paragraph">Even the basics move. The mandatory eigen risico goes from € 385 in 2026 to € 400 in 2027 (<a target="_blank" rel="noopener" href="https://zoek.officielebekendmakingen.nl/stcrt-2026-34372.html">Staatscourant</a>). A model trained on last year&#8217;s data will tell you € 385, with full confidence.</p>



<p class="wp-block-paragraph">Web search does not fully fix this. ChatGPT then cites whatever pages rank, and some comparison pages for 2027 still list Anderzorg as a separate insurer. When the model has no number, it may make one up.</p>



<h2 class="wp-block-heading">So I built a plugin</h2>



<p class="wp-block-paragraph"><a href="https://www.zorgverzekering.chat" data-type="link" data-id="https://www.zorgverzekering.chat">zorgverzekering.chat</a> is now a plugin for ChatGPT. It gives ChatGPT live access to every Dutch health insurance policy, so it stops guessing. Ask about premiums, coverage or contracts, and ChatGPT calls my server for the current numbers.</p>



<p class="wp-block-paragraph">Under the hood it is a remote MCP server with 12 read-only tools and six skills. The skills tell ChatGPT how to run a comparison: ask the right intake questions, call the tools, show a table of yearly costs, and cite the source.</p>



<figure class="wp-block-table"><table class="has-fixed-layout"><tbody><tr><th>You ask</th><th>What the plugin does</th></tr><tr><td>What will my household pay in 2027?</td><td>Calculates yearly costs per policy, including eigen risico and costs not covered</td></tr><tr><td>Which supplementary insurance covers braces and physio?</td><td>Searches all supplementary and dental packages by coverage</td></tr><tr><td>Does my policy have a contract with my hospital or physio?</td><td>Checks the contract status per care provider and policy, with date and origin</td></tr><tr><td>What changes in my policy next year?</td><td>Compares a policy across years and quotes the policy terms</td></tr><tr><td>What is the eigen risico, and when must I switch?</td><td>Returns the statutory amounts and deadlines</td></tr></tbody></table></figure>



<p class="wp-block-paragraph">Some rules I built in from day one:</p>



<ul class="wp-block-list">
<li><strong>Every number has a source.</strong> Each answer cites the insurer&#8217;s document. Amounts from the model&#8217;s memory are not allowed.</li>



<li><strong>Information, not advice.</strong> The plugin compares and calculates. You decide and sign up with the insurer yourself.</li>



<li><strong>Privacy by design.</strong> The tools are stateless and hosted in the EU. Nothing about your health is stored. For expected care use, the tools accept only coverage codes and counts, never diagnoses or medication.</li>



<li><strong>Dutch and English.</strong> Expats get the Dutch terms explained.</li>
</ul>



<p class="wp-block-paragraph">The same server also works in Claude, Gemini CLI, Cursor and VS Code.</p>



<h2 class="wp-block-heading">Millions of data points, one question away</h2>



<p class="wp-block-paragraph">The plugin sits on 1.58 million contract statuses: for each of 25,909 care providers, whether each of 83 basic policies has a contract with them. Each status has nine fields, so that is close to 15 million data points.</p>



<figure class="wp-block-table"><table class="has-fixed-layout"><tbody><tr><th>Dataset</th><th>Size</th></tr><tr><td>Contract statuses (care provider × policy × type of care)</td><td>1,584,725</td></tr><tr><td>Care providers in the register</td><td>26,522, at 35,710 locations</td></tr><tr><td>Coverage rules in supplementary and dental packages</td><td>5,238</td></tr><tr><td>Source documents (premium lists, policy terms, coverage guides)</td><td>607</td></tr><tr><td>Group schemes (municipalities, associations)</td><td>438</td></tr><tr><td>Supplementary and dental packages</td><td>390</td></tr><tr><td>Basic policies</td><td>83, from 38 insurer labels</td></tr></tbody></table></figure>



<p class="wp-block-paragraph">The contract data covers ten types of care. The biggest are physiotherapy (575,796 statuses), GPs (367,343) and mental health care (325,355). Pharmacies, hospitals, clinics, district nursing, rehabilitation, exercise therapy and lab diagnostics are in there too.</p>



<p class="wp-block-paragraph">I collect this data from the insurers&#8217; own publications and contract search tools. Every status carries the date it was retrieved and an expiry date, so the plugin knows when it needs a fresh lookup. The 2027 premiums go in as soon as insurers publish them, by 12 November.</p>



<p class="wp-block-paragraph">All of it is one question away. You do not need to open 38 websites and 607 PDFs. You ask ChatGPT, and it asks my server.</p>



<h2 class="wp-block-heading">How to use it</h2>



<p class="wp-block-paragraph">In ChatGPT, open the plugin directory, search for Zorgverzekering.chat and add it. Then just ask your question in Dutch or English.</p>



<p class="wp-block-paragraph">Prefer to connect it yourself, or using another client? The MCP server needs no login:</p>



<pre class="wp-block-code"><code>https:&#47;&#47;mcp.zorgverzekering.chat/mcp</code></pre>



<p class="wp-block-paragraph">In ChatGPT, turn on Developer mode under Settings &gt; Apps &amp; Connectors &gt; Advanced, then create a connector with that URL. In Claude, Cursor, VS Code or Gemini CLI, add it as a remote MCP server.</p>



<p class="wp-block-paragraph">Some questions to start with:</p>



<ul class="wp-block-list">
<li>&#8220;Wat is de goedkoopste basisverzekering voor volgend jaar?&#8221;</li>



<li>&#8220;Welke aanvullende verzekering vergoedt een beugel voor mijn kind en 12 keer fysiotherapie?&#8221;</li>



<li>&#8220;Is my hospital contracted by the cheapest Dutch health insurance policy?&#8221;</li>



<li>&#8220;We are a family of four in Utrecht. What is the cheapest combination for 2027, including braces and physio?&#8221;</li>



<li>&#8220;Wat verandert er aan mijn polis tussen 2026 en 2027? Citeer de voorwaarden.&#8221;</li>
</ul>



<h2 class="wp-block-heading">Why I think this matters beyond insurance</h2>



<p class="wp-block-paragraph">For twenty years, getting your information to people meant ranking a page. In AI assistants, the better position is being the source the assistant calls. A page can be outdated, paraphrased wrongly or ignored. A tool returns the current number with its source, every time someone asks.</p>



<p class="wp-block-paragraph">Health insurance is a good test case: the data changes every year, the stakes are real, and millions of people ask the same questions in the same six weeks. If it works here, it works for any market where people need data from this week, not from the last training run.</p>



<p class="wp-block-paragraph">Try it this switching season, and tell me what it gets wrong. You can reach me at help@zorgverzekering.chat.</p>



<h2 class="wp-block-heading">Sources</h2>



<ul class="wp-block-list">
<li><a href="https://www.emerce.nl/nieuws/72-miljoen-nederlanders-gebruiken-ai-hype-vaste-praktijk" target="_blank" rel="noopener">Emerce: 7,2 miljoen Nederlanders gebruiken AI</a> (Newcom AI Monitor, March 2026)</li>



<li><a href="https://gs.statcounter.com/ai-chatbot-market-share/all/netherlands" target="_blank" rel="noopener">Statcounter: AI chatbot market share in the Netherlands</a> (September 2026)</li>



<li><a href="https://zoek.officielebekendmakingen.nl/stcrt-2026-34372.html" target="_blank" rel="noopener">Staatscourant 2026, 34372</a> (eigen risico 2027)</li>



<li><a href="https://www.unive.nl/zorgverzekering/zekur/veelgestelde-vragen" target="_blank" rel="noopener">Univé: ZEKUR veelgestelde vragen</a></li>



<li><a href="https://www.anderzorg.nl/menzis" target="_blank" rel="noopener">Anderzorg wordt Menzis</a></li>



<li>Data counts: zorgverzekering.chat database, retrieved 28–29 October 2026</li>
</ul>
]]></content:encoded>
</item>
<item>
<title>ChatGPT plugin for Google Search Console</title>
<link>https://www.notprovided.eu/chatgpt-plugin-for-google-search-console/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1345</guid>
<pubDate>Sun, 20 Sep 2026 17:56:43 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[Most “AI for SEO” workflows still come down to the same three steps: export a CSV from Search Console, drop it into a chat window, and hope the model adds the numbers up correctly. It often does not. A 16-month export does…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">Most &#8220;AI for SEO&#8221; workflows still come down to the same three steps: export a CSV from Search Console, drop it into a chat window, and hope the model adds the numbers up correctly. It often does not. A 16-month export does not fit in a context window, and a model handed exactly 1,000 rows has no way of knowing whether that was all of them or the first page.</p>



<p class="wp-block-paragraph">So I removed the export step. GSC Wizard is now available as an app inside ChatGPT: <a target="_blank" rel="noreferrer noopener" href="https://chatgpt.com/plugins/gsc-wizard"><strong>chatgpt.com/plugins/gsc-wizard</strong></a></p>



<h2 class="wp-block-heading">What you can ask</h2>



<p class="wp-block-paragraph">You sign in with Google, and the app reads only the properties already connected to your account. From there you type what you would otherwise click together in a dashboard:</p>



<ul class="wp-block-list">
<li>&#8220;Give me a performance summary for example.com&#8221;</li>



<li>&#8220;Which queries gained or lost rankings versus last month?&#8221;</li>



<li>&#8220;Find pages that are decaying&#8221;</li>



<li>&#8220;Where am I cannibalizing keywords?&#8221;</li>



<li>&#8220;Show my CTR versus position curve&#8221;</li>



<li>&#8220;Inspect https://example.com/pricing in Google&#8221;</li>



<li>&#8220;Generate a full SEO report&#8221;</li>
</ul>



<p class="wp-block-paragraph">Behind that are more than a hundred tools: search analytics, ranking changes, anomaly and change-point detection, forecasting, opportunity scoring, URL inspection and bulk inspection, IndexNow submissions, content groups and topic clusters. Bing Webmaster Tools and GA4 are covered too, so &#8220;how did this page do in Google, Bing and conversions&#8221; is one question instead of three exports.</p>



<h2 class="wp-block-heading">Why this beats pasting a CSV</h2>



<p class="wp-block-paragraph">Every analysis is computed server-side against the warehouse that already holds your daily GSC data. ChatGPT never sees the raw rows; it gets a finished result, usually a few dozen rows, rendered as an interactive card or table.</p>



<p class="wp-block-paragraph">That matters for three reasons. Speed: an aggregate over millions of rows returns in about a second. Accuracy: the numbers come out of a SQL aggregation, not out of a language model doing arithmetic on a wall of text. Context: a full report costs a fraction of the tokens, so the conversation stays usable afterwards. Where a tool does cap its output, it says so explicitly and tells the model how to fetch the next page, which is the part silent truncation gets wrong.</p>



<h2 class="wp-block-heading">Setting it up</h2>



<p class="wp-block-paragraph">Open the app, sign in with Google, ask your question. It needs a GSC Wizard account with Search Console connected, and new accounts start on a trial. Access is scoped to your own account, and you can revoke it whenever you like.</p>



<p class="wp-block-paragraph">Not a ChatGPT user? The same server speaks MCP, so Claude, Gemini, Grok and anything else that talks MCP can connect to <code>mcp.gscwizard.com</code> and get identical answers.</p>



<p class="wp-block-paragraph">If you try it, I would like to hear which questions it answers well and which ones it fumbles. That feedback is what shapes the next set of tools.</p>
]]></content:encoded>
</item>
<item>
<title>Four SaaS bets I’m running in 2026</title>
<link>https://www.notprovided.eu/four-saas-bets-im-running-in-2026/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1255</guid>
<pubDate>Mon, 27 Apr 2026 14:05:41 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[After two decades of SEO consulting, I’ve leaned hard into shipping product. Here’s a quick tour of the four SaaS businesses I’m actively running right now, what they do, and who they’re built for. Glippy started life as a Chrome extension and…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">After two decades of SEO consulting, I&#8217;ve leaned hard into shipping product. Here&#8217;s a quick tour of the four SaaS businesses I&#8217;m actively running right now, what they do, and who they&#8217;re built for.</p>



<h2 class="wp-block-heading">Glippy MCP Server</h2>



<p class="wp-block-paragraph">Glippy started life as a Chrome extension and a desktop app for GEO (Generative Engine Optimization) checks. The MCP Server is the version built for people who already live inside an AI assistant. Install it once, paste a license key, and any MCP-compatible client (Claude, Cursor, VS Code, Windsurf, Zed, Cline, Warp) gets nine tools and 240+ checks across 16 categories, with bulk sitemap analysis up to 50,000 URLs per run.</p>



<p class="wp-block-paragraph">The point isn&#8217;t a fixed dashboard. The point is that the analysis lands in your chat, where you can cross-reference your own content, generate custom reports, turn findings into tickets, or pipe results into other agents. It&#8217;s a one-time purchase: $49 personal, $99 pro, $249 agency (50 activations, useful if you&#8217;re running an agent swarm). Best fit: site migrations, multi-domain portfolios, agency reporting at scale, competitor benchmarking, large ecommerce or SaaS sitemap audits, CI/CD GEO gates, and content ops prioritization.</p>



<p class="wp-block-paragraph">Link: <a href="https://www.glippy.dev/mcp/">glippy.dev/mcp/</a></p>



<h2 class="wp-block-heading">GSC Wizard</h2>



<p class="wp-block-paragraph">Google Search Console gives you numbers. GSC Wizard turns them into answers. It&#8217;s a layer on top of GSC that surfaces the analyses you&#8217;d normally rebuild in a spreadsheet every month: content decay maps, keyword cannibalization, CTR benchmarking against position curves, page poaching (positions 4-20 that are ripe for a small push into the top 3), topic clusters, hreflang and country-level performance, migration before/after compares, ML-powered traffic forecasting, and SEO A/B testing with statistical significance.</p>



<p class="wp-block-paragraph">Twelve built-in reports, all interactive, all designed to point at a specific action rather than another chart to stare at. Currently in a 6-weeks beta , no credit card.</p>



<p class="wp-block-paragraph">Link: <a href="https://www.gscwizard.com/">gscwizard.com</a></p>



<h2 class="wp-block-heading">Botsanalyser</h2>



<p class="wp-block-paragraph">Botsanalyser is the one I get the most &#8220;finally&#8221; reactions to. GA4 filters every bot out by design, and on Shopify, managed WordPress, Webflow, and Squarespace you can&#8217;t reach the raw server logs. So GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Bytespider, and the dozens of other AI crawlers showing up every week are completely invisible to most marketing teams.</p>



<p class="wp-block-paragraph">Botsanalyser is a plugin (WordPress, WooCommerce, Shopify, Magento at launch, Webflow and Squarespace next, universal script tag for everyone else). It captures every bot hit, identifies the user-agent, stores the data on Cloudflare, and shows it in a dashboard, a Monday digest email, and a REST API with native connectors to GA4, Looker Studio, BigQuery, HubSpot, and Segment. No client JS, no Core Web Vitals impact, no cookies, no PII, GDPR-safe by design.</p>



<p class="wp-block-paragraph">Pre-launch, summer 2026, with 500+ people already on the waitlist.</p>



<p class="wp-block-paragraph">Link: <a href="https://www.botsanalyser.com/">botsanalyser.com</a></p>



<h2 class="wp-block-heading">Peecockpit</h2>



<p class="wp-block-paragraph">Peecockpit is the most ambitious of the four and probably the one that&#8217;s hardest to pitch in one line, so here&#8217;s the honest version: most GEO tools surface one signal. Peecockpit joins six per-URL data sources (access logs, GSC URL Inspection, GSC Wizard, Peec AI, Glippy, GA4) on the canonical URL, then routes every gap to the team that can actually fix it (SEO, PR, or Content).</p>



<p class="wp-block-paragraph">Under the hood it orchestrates four interoperating MCP servers (Peec, Ahrefs, GSC Wizard, Glippy) plus wrapped Ahrefs Firehose, Majestic, and Botsanalyser APIs, and collapses what would otherwise be 3-7 expert MCP calls into one department-shaped answer per page. Every action emits a tracked ID and gets re-checked at T+14 and T+28, so &#8220;we shipped a fix, did it work?&#8221; stops being a guess. Built for in-house teams who want a full-funnel GEO operating system, and for agencies that want one inbox of severity × MRR across every client.</p>



<p class="wp-block-paragraph">Link: <a href="https://www.peecockpit.pro/">peecockpit.pro</a></p>



<hr class="wp-block-separator has-alpha-channel-opacity"/>



<p class="wp-block-paragraph">Four products, one through-line: take the work I&#8217;ve been doing for clients for years and turn it into something a marketing team or an AI agent can run on their own. If any of these scratch an itch you&#8217;ve got, ping me on <a href="https://www.linkedin.com/in/jbobbink">LinkedIn</a> or hit the waitlist.</p>
]]></content:encoded>
</item>
<item>
<title>Transform Any Listing Page Into a Professional Article in Minutes</title>
<link>https://www.notprovided.eu/my-new-listicle-wizard-app/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1224</guid>
<pubDate>Tue, 13 Jan 2026 20:22:49 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[Creating compelling listicle content used to mean hours of research, writing, and formatting. Whether you’re curating the best hotels in Paris, comparing software tools, or rounding up top job opportunities, the process was tedious. Until now. Listicle Wizard is a revolutionary content…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">Creating compelling listicle content used to mean hours of research, writing, and formatting. Whether you&#8217;re curating the best hotels in Paris, comparing software tools, or rounding up top job opportunities, the process was tedious. Until now.</p>



<h2 class="wp-block-heading">Introducing Listicle Wizard</h2>



<p class="wp-block-paragraph">Listicle Wizard is a revolutionary content generation platform that transforms any listing page from any website into a polished, SEO-optimized article. Then publishes it to your preferred platform with a single click.</p>



<p class="wp-block-paragraph"><em>Ready to transform your content workflow? Try Listicle Wizard and experience the future of content creation.</em> <a href="https://listicle.dev">https://listicle.dev</a> </p>



<h2 class="wp-block-heading">How It Works</h2>



<p class="wp-block-paragraph">The magic happens in four simple steps:</p>



<p class="wp-block-paragraph"><strong>Step 1: Paste Your Source URL</strong> Start with any listing, search, or category page. Whether it&#8217;s an Amazon search result, Booking.com hotel listings, Indeed job postings, or Zillow property searches, Listicle Wizard works with virtually any website.</p>



<p class="wp-block-paragraph"><strong>Step 2: Choose Your Template</strong> Select from nine pre-built templates optimized for specific industries: travel guides, product comparisons, job roundups, restaurant reviews, real estate listings, software comparisons, course recommendations, event guides, and health content. Need something unique? Create a custom prompt.</p>



<p class="wp-block-paragraph"><strong>Step 3: Select Your Items</strong> Our proprietary link-scoring engine intelligently extracts and ranks all items from the page. Simply select which ones you want to include in your article: anywhere from 3 to 15 items.</p>



<p class="wp-block-paragraph"><strong>Step 4: Generate and Publish</strong> Watch as AI transforms your selections into a beautifully structured article. Edit any section, regenerate specific parts, reorder with drag-and-drop, then publish directly to your platform of choice.</p>



<h2 class="wp-block-heading">What Makes Listicle Wizard Different</h2>



<h3 class="wp-block-heading">Universal Web Scraping</h3>



<p class="wp-block-paragraph">Unlike tools limited to specific platforms, Listicle Wizard works with thousands of websites across every industry. No coding, browser extensions, or technical setup required.</p>



<h3 class="wp-block-heading">Experienced SEO Powered Quality</h3>



<p class="wp-block-paragraph">Powered by advanced AI, every article is professionally written and SEO-optimized. The content isn&#8217;t just scraped and reformatted, it&#8217;s intelligently crafted into engaging narratives that readers love.</p>



<h3 class="wp-block-heading">One-Click Publishing to 11+ Platforms</h3>



<p class="wp-block-paragraph">Connected directly to WordPress, Shopify, Wix, Webflow, Ghost, Blogger, Squarespace, and automation platforms like Zapier, Make, and N8N. No more copy-pasting or manual formatting.</p>



<h3 class="wp-block-heading">Multi-Language Support</h3>



<p class="wp-block-paragraph">Generate content in over 25 languages. Source pages are automatically detected, and articles can be created in any supported language, perfect for reaching global audiences.</p>



<h2 class="wp-block-heading">Perfect For Content Creators, Marketers, and Affiliates</h2>



<p class="wp-block-paragraph">Whether you&#8217;re running an affiliate marketing site, managing content for clients, or building a niche blog, Listicle Wizard eliminates the bottleneck of content creation. What previously took hours now takes minutes.</p>



<p class="wp-block-paragraph">Imagine publishing daily roundups of trending products, weekly hotel guides for popular destinations, or comprehensive software comparison articles, all without the manual research and writing grind.</p>



<h2 class="wp-block-heading">Start Creating Today</h2>



<p class="wp-block-paragraph">Stop spending hours on content that could take minutes. Listicle Wizard combines intelligent web scraping, powerful AI generation, and seamless publishing into one streamlined workflow. Your next viral listicle is just four steps away.</p>



<p class="wp-block-paragraph"><em>Ready to transform your content workflow? Try Listicle Wizard and experience the future of content creation.</em> <a href="https://listicle.dev">https://listicle.dev</a> </p>



<p class="wp-block-paragraph"></p>
]]></content:encoded>
</item>
<item>
<title>Shopify SEO challenges: platform architecture limitations and workarounds</title>
<link>https://www.notprovided.eu/shopify-seo-challenges-platform-architecture-limitations-and-workarounds/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1204</guid>
<pubDate>Tue, 02 Dec 2025 09:01:33 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[Shopify stores face significant SEO limitations rooted in the platform’s fundamental design choices, from rigid URL structures that cannot be changed to duplicate content issues affecting virtually every product page. While Shopify provides basic SEO features out-of-the-box (auto-generated sitemaps, canonical tags, SSL)…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">Shopify stores face significant SEO limitations rooted in the platform&#8217;s fundamental design choices, from rigid URL structures that cannot be changed to duplicate content issues affecting virtually every product page. While Shopify provides basic SEO features out-of-the-box (auto-generated sitemaps, canonical tags, SSL), merchants must understand these constraints to work around them effectively. <strong>The most critical issue is that products accessible via multiple collection URLs create duplicate content by default</strong>, diluting link equity and confusing search engines about which page to rank.</p>



<p class="wp-block-paragraph">This matters because Shopify powers over 4 million stores globally, and many merchants assume ecommerce SEO is fully handled by the platform. In reality, several SEO problems require theme code modifications, third-party apps, or workarounds that aren&#8217;t immediately obvious. The good news: most issues can be mitigated once you understand why they occur.</p>



<h2 class="wp-block-heading">Every product creates duplicate URLs through collection paths</h2>



<p class="wp-block-paragraph">The most pervasive SEO problem on Shopify stems from how the platform generates product URLs. When a product is assigned to multiple collections, Shopify automatically creates separate URLs for each path. A &#8220;Blue T-Shirt&#8221; in three collections generates four accessible URLs: <code>/products/blue-t-shirt</code> (canonical), <code>/collections/mens-clothing/products/blue-t-shirt</code>, <code>/collections/summer-sale/products/blue-t-shirt</code>, and <code>/collections/all/products/blue-t-shirt</code>.</p>



<p class="wp-block-paragraph">The technical culprit is Shopify&#8217;s Liquid templating code. Theme files use the filter <code>| within: current_collection</code>, which dynamically generates collection-based URLs whenever products appear on collection pages. <strong>Internal links on collection pages point to these non-canonical URLs by default</strong>, meaning most internal link equity flows to duplicate pages rather than the canonical product URL.</p>



<p class="wp-block-paragraph">While Shopify adds canonical tags pointing to the proper <code>/products/</code> URL, canonicals are hints rather than directives. Google may ignore them, especially when overwhelming internal linking signals point elsewhere. One documented case found a store with <strong>millions of indexed pages when it should have had approximately 200</strong>, caused by filter parameters creating exponential URL variations like <code>?filter.v.option.color=red</code> being indexed as separate pages.</p>



<p class="wp-block-paragraph">The fix requires editing theme code, specifically, locating the product grid template (in Dawn: <code>main-collection-product-grid.liquid</code>) and removing the <code>| within: current_collection</code> filter so links point directly to canonical URLs. This single change addresses the platform&#8217;s most widespread SEO problem.</p>



<h2 class="wp-block-heading">Shopify enforces rigid URL structures that cannot be modified</h2>



<p class="wp-block-paragraph">Unlike WordPress or Magento, Shopify enforces mandatory URL prefixes with no option for customization. Products must live at <code>/products/</code>, collections at <code>/collections/</code>, and pages at <code>/pages/</code>. <strong>You cannot create hierarchical category structures</strong> like <code>/clothing/shirts/blue-tee</code> or place products at the root domain like <code>/blue-tee</code>.</p>



<p class="wp-block-paragraph">This limitation impacts several SEO strategies. Sites cannot mirror their content taxonomy in URL paths, making it harder for search engines to understand site structure through URLs alone. Blog URLs force redundant structures, naming your blog &#8220;blog&#8221; creates paths like <code>/blogs/blog/article-title</code>. For migrations from other platforms with custom URL structures, extensive 301 redirects become necessary, and the redirect limit of <strong>100,000 on standard plans</strong> can be restrictive for larger catalogs.</p>



<p class="wp-block-paragraph">The only workaround for complete URL control is Shopify Plus with headless commerce using Hydrogen and Oxygen, which decouples the frontend entirely. Standard Shopify merchants must accept these constraints and compensate through strong internal linking hierarchies and breadcrumb navigation.</p>



<h2 class="wp-block-heading">Technical restrictions limit what SEO professionals can control</h2>



<p class="wp-block-paragraph">Shopify&#8217;s hosted nature means certain technical SEO elements remain partially or fully inaccessible. The platform now allows robots.txt customization through a <code>robots.txt.liquid</code> template file (added in 2021), but modifications require Liquid syntax and some objects remain restricted. Server log access, valuable for analyzing Googlebot crawl patterns, is completely unavailable.</p>



<p class="wp-block-paragraph"><strong>Checkout customization remains heavily restricted for non-Plus merchants</strong>, limited to basic appearance changes. Plus merchants historically accessed <code>checkout.liquid</code> directly, though Shopify is migrating to Checkout Extensibility as of August 2024. Code access follows a hierarchy: global Liquid objects are available everywhere, but template-specific variables only work in certain contexts, requiring careful implementation for SEO customizations.</p>



<p class="wp-block-paragraph">Site speed challenges compound these restrictions. While Shopify provides baseline performance through Cloudflare CDN, browser caching, and WebP image conversion, <strong>app bloat is the primary speed killer</strong>. Each installed app adds JavaScript, CSS, and HTTP requests. A study found Shopify Plus merchants achieved 20% speed improvements simply by optimizing Liquid code and removing unused apps. Customer-facing apps particularly impact Core Web Vitals, with no built-in limit on installations leading to accumulation over time.</p>



<h2 class="wp-block-heading">Theme structure creates header hierarchy and schema problems</h2>



<p class="wp-block-paragraph">Most Shopify themes, including the popular Dawn theme, wrap the store logo in an <code>&lt;h1&gt;</code> tag on the homepage, resulting in an empty or invalid H1 when SEO tools audit the site. The H1 should contain descriptive text about the page&#8217;s main topic, not an image. This requires manual correction by replacing the <code>&lt;h1&gt;</code> wrapper with a <code>&lt;div&gt;</code> in header.liquid and adding a proper text-based H1 to homepage content.</p>



<p class="wp-block-paragraph">Dawn theme (powering over <strong>800,000 stores</strong>) offers significant SEO advantages over older themes: 35% faster than Debut, approximately 86.7% Core Web Vitals pass rate versus 65-68% for legacy themes, and built-in structured data using the <code>structured_data</code> Liquid filter. However, default schema markup remains limited, missing fields required for Google Merchant Listing eligibility including shipping details, return policies, and aggregate ratings.</p>



<p class="wp-block-paragraph">Schema conflicts arise when themes generate microdata format while apps inject JSON-LD, creating <strong>multiple instances of Product schema</strong> that Google may misread. Before adding any schema apps, merchants should audit existing markup with Google&#8217;s Rich Results Test and enable &#8220;microdata filtering&#8221; in apps to prevent conflicts.</p>



<h2 class="wp-block-heading">Sitemaps and meta tags have limited customization options</h2>



<p class="wp-block-paragraph">Shopify auto-generates sitemaps at <code>/sitemap.xml</code> using a parent-child structure with separate files for products, collections, blogs, and pages. <strong>You cannot upload custom sitemaps</strong>, create separate image or video sitemaps, control URL priority, or set custom change frequencies. The only removal method requires setting an <code>seo.hidden</code> metafield on individual pages.</p>



<p class="wp-block-paragraph">Meta tag handling introduces additional constraints. Title tags are limited to <strong>70 characters</strong> (Google truncates beyond this), and meta descriptions should stay under 155-160 characters. More significantly, <strong>Shopify doesn&#8217;t natively support separate SEO meta fields for product variants</strong>, all variants share identical meta titles and descriptions. A store selling shirts in multiple colors cannot target &#8220;blue linen shirt&#8221; versus &#8220;red linen shirt&#8221; without workarounds involving variant metafields and Liquid code modifications or creating separate product pages.</p>



<p class="wp-block-paragraph">Collection page filtering creates near-duplicate pages through URL parameters, and filtered results should typically use <code>noindex, follow</code> meta robots tags to prevent index bloat. Multiple meta description tags can appear when themes and apps both inject them, confusing search engine indexing.</p>



<h2 class="wp-block-heading">App dependency creates ongoing costs and performance tradeoffs</h2>



<p class="wp-block-paragraph">Many essential SEO tasks require third-party apps, creating ongoing expenses and performance implications:</p>



<ul class="wp-block-list">
<li><strong>Schema markup customization</strong> (JSON-LD for SEO: $399/year)</li>



<li><strong>Bulk meta optimization</strong> (Smart SEO, Yoast: $0-$35/month)</li>



<li><strong>Image compression at scale</strong> (TinyIMG, Booster: $0-$69/month)</li>



<li><strong>Redirect management and 404 monitoring</strong> (various: $0-$99/month)</li>



<li><strong>Multilingual SEO with hreflang</strong> (requires external sitemap solutions)</li>
</ul>



<p class="wp-block-paragraph">The risks extend beyond cost. App discontinuation can break functionality without warning. Multiple apps frequently create conflicting implementations, several apps adding schema simultaneously cause validation errors. <strong>App bloat is a leading cause of poor Core Web Vitals scores</strong>, with some apps loading features even when unused. Best practice limits customer-facing apps to five essentials maximum with regular audits to remove unused installations.</p>



<h2 class="wp-block-heading">Shopify Plus unlocks advanced capabilities for enterprise SEO</h2>



<p class="wp-block-paragraph">Plus merchants gain significant SEO advantages unavailable on standard plans. Headless commerce through Hydrogen and Oxygen provides <strong>complete frontend URL structure control</strong>, custom meta tag implementation, and static site generation for near-instant load times. The Combined Listings app (Plus exclusive) creates SEO-friendly URLs for product variants, addressing a major limitation for variant-heavy stores.</p>



<p class="wp-block-paragraph">Higher API rate limits enable bulk SEO operations at scale, while priority hosting infrastructure delivers better baseline performance. According to Shopify&#8217;s 2024 Enterprise Commerce Report, Plus merchants achieving comprehensive SEO maturity see <strong>35-50% higher customer lifetime value</strong> compared to those focusing solely on paid acquisition. Enterprise implementations typically show 85-95% traffic recovery within 30 days post-migration, 10-15% improvement by month four, and 25-40%+ increases by month nine with strategic implementations.</p>



<p class="wp-block-paragraph">Recent platform updates include variant limits expanded from 100 to <strong>2,000 products</strong> (via GraphQL API 2024-04), domain-specific robots.txt rules through <code>request.host</code> for Shopify Markets, and Shopify Magic AI for generating SEO-optimized titles and descriptions.</p>



<h2 class="wp-block-heading">Comparing Shopify&#8217;s SEO capabilities to alternatives</h2>



<p class="wp-block-paragraph">Shopify intentionally trades SEO flexibility for ease of use and security. WordPress with WooCommerce offers complete URL customization, full robots.txt and .htaccess control, server log access, thousands of free SEO plugins (Yoast, Rank Math), and superior blogging capabilities. Magento provides similar complete control but requires significant technical resources.</p>



<p class="wp-block-paragraph">Shopify&#8217;s advantages include managed hosting eliminating server-side SEO concerns, automatic SSL implementation, guaranteed mobile responsiveness across themes, and platform-wide performance optimizations. <strong>For merchants prioritizing rapid deployment and ease of management over granular SEO control</strong>, Shopify often proves the right choice, provided they understand and address the platform&#8217;s inherent limitations.</p>



<h2 class="wp-block-heading">Conclusion: actionable priorities for Shopify SEO</h2>



<p class="wp-block-paragraph">The most impactful actions for Shopify merchants address structural issues first. <strong>Fix duplicate product URLs immediately</strong> by removing the <code>| within: collection</code> filter from theme templates, this single change eliminates the platform&#8217;s most common SEO problem. Correct homepage H1 tags by replacing logo wrappers with proper heading text. Audit existing schema markup before adding apps to prevent conflicts that confuse search engines.</p>



<p class="wp-block-paragraph">Accept URL structure limitations but compensate through strong internal linking hierarchies and breadcrumb navigation. Limit apps to essential tools only, conducting quarterly audits to remove unused installations dragging down site speed. Submit sitemaps to Google Search Console and monitor the Index Coverage report for unexpected URL inflation from filter parameters.</p>



<p class="wp-block-paragraph">For stores where SEO is mission-critical and variant-level optimization or custom URL structures are necessary, Shopify Plus with headless architecture or alternative platforms like WooCommerce warrant serious consideration. Standard Shopify provides &#8220;good enough&#8221; SEO for most ecommerce needs, but achieving excellent organic performance requires understanding why these issues occur and systematically addressing each one.</p>



<hr class="wp-block-separator has-alpha-channel-opacity"/>



<h2 class="wp-block-heading">References</h2>



<ol class="wp-block-list">
<li><strong>The Ultimate Guide to Shopify SEO in 2022</strong> &#8211; NOVOS<br><a href="https://thisisnovos.com/blog/the-ultimate-guide-to-shopify-seo/">https://thisisnovos.com/blog/the-ultimate-guide-to-shopify-seo/</a></li>



<li><strong>Advanced Shopify SEO Tips to Scale Organic Traffic</strong> &#8211; ConvertMate<br><a href="https://www.convertmate.io/blog/shopify-seo-for-beginners">https://www.convertmate.io/blog/shopify-seo-for-beginners</a></li>



<li><strong>How to Fix Shopify Duplicate Content</strong> &#8211; Go Fish Digital<br><a href="https://gofishdigital.com/blog/how-to-fix-shopify-duplicate-content/">https://gofishdigital.com/blog/how-to-fix-shopify-duplicate-content/</a></li>



<li><strong>How to Fix Duplicate Content on Shopify</strong> &#8211; Digismoothie<br><a href="https://www.digismoothie.com/blog/shopify-duplicate-content">https://www.digismoothie.com/blog/shopify-duplicate-content</a></li>



<li><strong>How to fix duplicate content in Shopify</strong> &#8211; Kahunam<br><a href="https://kahunam.com/articles/shopify/fix-duplicate-content-shopify/">https://kahunam.com/articles/shopify/fix-duplicate-content-shopify/</a></li>



<li><strong>An SEO Tutorial for Handling Shopify&#8217;s Duplicate Collections and Product Pages</strong> &#8211; Intellitonic<br><a href="https://intellitonic.com/blog/an-seo-tutorial-for-handling-shopifys-duplicate-collections-and-product-pages/">https://intellitonic.com/blog/an-seo-tutorial-for-handling-shopifys-duplicate-collections-and-product-pages/</a></li>



<li><strong>Crawl Budget: What Is It and Why It&#8217;s Important for SEO</strong> &#8211; Passion Digital<br><a href="https://passion.digital/blog/crawl-budget-what-is-it-and-why-its-important-for-seo/">https://passion.digital/blog/crawl-budget-what-is-it-and-why-its-important-for-seo/</a></li>



<li><strong>Weird Shopify SEO Issue &#8211; Millions of Indexed Pages</strong> &#8211; Ed Codes<br><a href="https://ed.codes/blog/weird-shopify-seo-issue-millions-of-indexed-pages">https://ed.codes/blog/weird-shopify-seo-issue-millions-of-indexed-pages</a></li>



<li><strong>How to fix your Shopify product URL structure for better SEO performance</strong> &#8211; Freddie Chatt<br><a href="https://freddiechatt.com/shopify-product-url-structure/">https://freddiechatt.com/shopify-product-url-structure/</a></li>



<li><strong>Shopify URLs: Optimize for SEO &amp; UX</strong> &#8211; BlueTuskr<br><a href="https://blog.bluetuskr.com/shopify-url-optimization">https://blog.bluetuskr.com/shopify-url-optimization</a></li>



<li><strong>Guide to Change Shopify URL Structure for SEO-friendliness</strong> &#8211; EComposer<br><a href="https://ecomposer.io/blogs/shopify-knowledge/shopify-url-structure">https://ecomposer.io/blogs/shopify-knowledge/shopify-url-structure</a></li>



<li><strong>Overcoming SEO limitations on Shopify &#8211; Part 1</strong> &#8211; Croud<br><a href="https://croud.com/blog/seo-limitations-on-shopify-how-to-overcome-them-pt-1/">https://croud.com/blog/seo-limitations-on-shopify-how-to-overcome-them-pt-1/</a></li>



<li><strong>7 Core Technical SEO Challenges For Shopify Websites &amp; How To Fix Them</strong> &#8211; blubolt<br><a href="https://blubolt.com/insights/7-technical-seo-challenges">https://blubolt.com/insights/7-technical-seo-challenges</a></li>



<li><strong>How to Fix an Empty H1 Tag on Your Shopify Homepage</strong> &#8211; Masuga<br><a href="https://www.gomasuga.com/articles/how-to-fix-empty-h1-header-on-shopify-homepage">https://www.gomasuga.com/articles/how-to-fix-empty-h1-header-on-shopify-homepage</a></li>



<li><strong>Missing H1 tags</strong> &#8211; Shopify Community<br><a href="https://community.shopify.com/c/shopify-design/missing-h1-tags/m-p/2576801">https://community.shopify.com/c/shopify-design/missing-h1-tags/m-p/2576801</a></li>



<li><strong>How To Create SEO-Friendly Shopify URLs &#8211; Complete Guide</strong> &#8211; Candidinfo<br><a href="https://www.candidinfo.com/blog/seo-friendly-shopify-urls/">https://www.candidinfo.com/blog/seo-friendly-shopify-urls/</a></li>



<li><strong>Avoiding Duplicate Content Issues On Your Shopify Store</strong> &#8211; OyeCommerz<br><a href="https://oyecommerz.com/blog/avoid-duplicate-content-issues-shopify/">https://oyecommerz.com/blog/avoid-duplicate-content-issues-shopify/</a></li>



<li><strong>Customize robots.txt</strong> &#8211; Shopify Developers<br><a href="https://shopify.dev/docs/storefronts/themes/seo/robots-txt">https://shopify.dev/docs/storefronts/themes/seo/robots-txt</a></li>



<li><strong>Resolving Shopify Duplicate Content Between Collection &amp; Product Pages</strong> &#8211; Amsive<br><a href="https://www.amsive.com/insights/seo/resolving-shopify-duplicate-content-between-collection-product-pages/">https://www.amsive.com/insights/seo/resolving-shopify-duplicate-content-between-collection-product-pages/</a></li>



<li><strong>Mastering Shopify Internal Linking: Enhance Your SEO and User Experience</strong> &#8211; Praella<br><a href="https://praella.com/blogs/shopify-insights/mastering-shopify-internal-linking-enhance-your-seo-and-user-experience">https://praella.com/blogs/shopify-insights/mastering-shopify-internal-linking-enhance-your-seo-and-user-experience</a></li>



<li><strong>Shopify URL Structure Optimization: Best Practices For SEO-Friendly URLs</strong> &#8211; Redshift Digital Marketing<br><a href="https://redshiftdm.com/shopify-url-structure-optimization/">https://redshiftdm.com/shopify-url-structure-optimization/</a></li>



<li><strong>Finding and submitting your sitemap</strong> &#8211; Shopify Help Center<br><a href="https://help.shopify.com/en/manual/promoting-marketing/seo/find-site-map">https://help.shopify.com/en/manual/promoting-marketing/seo/find-site-map</a></li>



<li><strong>Technical SEO Basics: Crawling, Indexing, and Ranking</strong> &#8211; Sherpas Design<br><a href="https://sherpas.design/blogs/e-commerce/technical-seo-basics-crawling-indexing-ranking">https://sherpas.design/blogs/e-commerce/technical-seo-basics-crawling-indexing-ranking</a></li>
</ol>
]]></content:encoded>
</item>
<item>
<title>The LLMs.txt Creator WordPress Plugin</title>
<link>https://www.notprovided.eu/introducing-the-llms-txt-creator-wordpress-plugin/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1107</guid>
<pubDate>Wed, 19 Feb 2025 20:23:13 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[UPDATE 06-05-2025: after testing this on more than 150 domains I have not seen active use by any of the LLMs bots. No hits visible in server logs. Hence I can conclude LLMs.txt are a waste of time, for now. Enhance your…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph"><strong>UPDATE 06-05-2025: after testing this on more than 150 domains I have not seen active use by any of the LLMs bots. No hits visible in server logs. Hence I can conclude LLMs.txt are a waste of time, for now.</strong></p>



<h3 class="wp-block-heading"><strong>Get Started Today!</strong></h3>



<p class="wp-block-paragraph">Enhance your website’s AI accessibility with the <strong>LLMs.txt Creator WordPress Plugin</strong>. Download and activate it now to ensure your content is AI-ready: <a href="https://www.notprovided.eu/area51/llms-txt-creator.zip">https://notprovided.eu/area51/llms-txt-creator.zip</a></p>



<p class="wp-block-paragraph">For more details or support, feel free to reach out to me via <a href="https://www.linkedin.com/in/jbobbink/#seo">LinkedIn</a> or in the comments.</p>



<p class="wp-block-paragraph"><strong>Update 21-02: more flexibility via settings</strong></p>



<p class="wp-block-paragraph">You can now update all fields in the LLMs.txt file except for the website information, which is handled via the default WordPress settings.</p>



<p class="wp-block-paragraph">If you don&#8217;t want specific posts or pages included, you can exclude them via a commaseparated lists of ID&#8217;s. If you want to exclude specific categories, you can use the multiselect field to exclude specific categories from being included. This only works for posts.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="648" height="629" src="https://www.notprovided.eu/assets/image-1.png" alt="" class="wp-image-1122" srcset="https://www.notprovided.eu/assets/image-1.png 648w, https://www.notprovided.eu/assets/image-1-300x291.png 300w" sizes="auto, (max-width: 648px) 100vw, 648px" /></figure>



<p class="wp-block-paragraph">For the OpenAI API connection you now also have more flexibility so you can make sure you get localized or even translated results back if you want, because you can update prompts.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="623" height="492" src="https://www.notprovided.eu/assets/image.png" alt="" class="wp-image-1121" srcset="https://www.notprovided.eu/assets/image.png 623w, https://www.notprovided.eu/assets/image-300x237.png 300w" sizes="auto, (max-width: 623px) 100vw, 623px" /></figure>



<p class="wp-block-paragraph"><strong>Update 20-02: Enhanced Summary Generation</strong></p>



<p class="wp-block-paragraph">I am excited to announce a new update to the LLMs.txt Generator plugin that improves how your site’s content is summarized in the generated llms.txt file. The latest changes ensure that your summaries are more accurate and informative by using a three-tiered approach:</p>



<ol class="wp-block-list">
<li><strong>OpenAI Generated Summary:</strong><br>When available, the plugin now first uses an OpenAI-generated summary. This state-of-the-art summary is tailored to capture the essence of your content in a concise and coherent way.</li>



<li><strong>WordPress Excerpt:</strong><br>If an OpenAI summary isn’t available for a particular post or page, the plugin will automatically fall back to the WordPress excerpt you’ve set. This ensures that a manually curated summary isn’t overlooked.</li>



<li><strong>First 30 Words of the Content:</strong><br>As a final fallback, if neither an OpenAI summary nor an excerpt is available, the plugin grabs the first 30 words of the content. This guarantees that every post or page in your llms.txt file will have a summary, giving users and search engines a quick snapshot of what your content is about.</li>
</ol>



<p class="wp-block-paragraph">This enhanced summary generation method not only improves the clarity of your site structure but also helps language models better understand your content. </p>



<p class="wp-block-paragraph">Update your plugin today and let us know what you think!</p>



<p class="wp-block-paragraph"><strong>Boost your website’s AI Accessibility with LLMs.txt</strong></p>



<p class="wp-block-paragraph">In today’s rapidly evolving AI-driven world, ensuring that search engines and Large Language Models (LLMs) understand your website is crucial. That’s where the <strong>LLMs.txt Creator WordPress plugin</strong> comes in. This simple yet powerful plugin automatically creates and updates an <code>llms.txt</code> file for your site, helping AI tools easily access structured information about your content. Also see https://llmstxt.org/</p>



<h3 class="wp-block-heading"><strong>What Is the LLMs.txt File?</strong></h3>



<p class="wp-block-paragraph">The <code>llms.txt</code> file acts as a guide for AI models, similar to <code>robots.txt</code> for search engines. It provides structured details about your site, including:</p>



<ul class="wp-block-list">
<li>The website’s title and description</li>



<li>A list of recent blog posts and pages</li>
</ul>



<p class="wp-block-paragraph">This allows AI-powered assistants, search engines, and automated tools to retrieve the most relevant information efficiently.</p>



<h3 class="wp-block-heading"><strong>How the LLMs.txt Creator Plugin Works</strong></h3>



<p class="wp-block-paragraph">The plugin automatically generates an <code>llms.txt</code> file and keeps it up to date. It:</p>



<ol start="1" class="wp-block-list">
<li><strong>Extracts Key Website Data</strong> – Retrieves your website name, description, and homepage URL.</li>



<li><strong>Lists Recent Content</strong> – Displays a structured list of the latest blog posts and pages.</li>



<li><strong>Updates Automatically</strong> – Whenever a new post or page is created, modified, or deleted, the <code>llms.txt</code> file refreshes accordingly.</li>



<li><strong>Provides Customization Options</strong> – Allows users to specify how many posts and pages should be listed in the file through a settings panel.</li>
</ol>



<h3 class="wp-block-heading"><strong>How to Use the Plugin</strong></h3>



<ol start="1" class="wp-block-list">
<li><strong>Install and Activate the Plugin</strong>
<ul class="wp-block-list">
<li>Go to <strong>Plugins &gt; Add New</strong> in your WordPress admin panel.</li>



<li>Upload the plugin as a ZIP file</li>



<li>Activate the plugin.</li>
</ul>
</li>



<li><strong>Configure the Settings</strong>
<ul class="wp-block-list">
<li>Navigate to <strong>Settings &gt; LLMs.txt Settings</strong> in your WordPress dashboard.</li>



<li>Adjust the number of recent posts and pages to include in the file.</li>



<li>Click <strong>Save Changes</strong> to apply updates.</li>
</ul>
</li>



<li><strong>Check the Generated File</strong>
<ul class="wp-block-list">
<li>The <code>llms.txt</code> file will be created automatically in your site’s root directory.</li>



<li>Visit <code>https://yourwebsite.com/llms.txt</code> to view the structured AI-readable content.</li>
</ul>
</li>
</ol>



<h3 class="wp-block-heading"><strong>Get Started Today!</strong></h3>



<p class="wp-block-paragraph">Enhance your website’s AI accessibility with the <strong>LLMs.txt Creator WordPress Plugin</strong>. Download and activate it now to ensure your content is AI-ready: <a href="https://www.notprovided.eu/area51/llms-txt-creator.zip">https://notprovided.eu/area51/llms-txt-creator.zip</a></p>



<p class="wp-block-paragraph">For more details or support, feel free to reach out to me via <a href="https://www.linkedin.com/in/jbobbink/#seo">LinkedIn</a> or in the comments.</p>



<h2 class="wp-block-heading">Known bugs</h2>



<ol class="wp-block-list">
<li>When your WordPress doesn&#8217;t live in the root folder, so in www.domain.com/blog/ for example, it may not work</li>



<li>The amount of generated summaries mentioned in the notices are not the actual amount of generated summaries but the value from the settings page. It needs to become dynamic.</li>
</ol>



<p class="wp-block-paragraph"></p>
]]></content:encoded>
</item>
<item>
<title>WordPress plugin: Track ChatGPT Hits</title>
<link>https://www.notprovided.eu/track-chatgpt/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1051</guid>
<pubDate>Sun, 10 Dec 2023 16:38:54 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Analytics</category>
<category>Miscellaneous</category>
<category>Technical SEO</category>
<description><![CDATA[Due to high demand I decided to make a user friendly version of tracking known OpenAI bot hits. This WordPress plugin tracks URL requests by the ChatGPT / OpenAI bots and direct user actions by tracking request made by specific user agents…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">Due to high demand I decided to make a user friendly version of tracking known OpenAI bot hits. This WordPress plugin tracks URL requests by the ChatGPT / OpenAI bots and direct user actions by tracking request made by specific user agents.</p>



<p class="wp-block-paragraph">You can download the plugin at <a href="https://www.notprovided.eu/area51/track-chatgpt-v_1_0.zip">https://www.notprovided.eu/area51/track-chatgpt-v_1_0.zip</a> &#8211; You can simply upload the folder <em>track-chatgpt</em> to your plugin directory via sFTP or use the WordPress plugin interface to upload the ZIP. </p>







<p class="wp-block-paragraph">There are currently two known useragents and a small set of <a href="https://platform.openai.com/docs/plugins/production">IP addresses</a> which can be used to check if it are valid requests by OpenAI ChatGPTs. The plugin will show if the request where from a verified source (valid requests) or not.</p>



<p class="wp-block-paragraph">The plugin shows a graph of the hits during the past 28 days. It has a download functionality to download the full dataset at once.</p>



<p class="wp-block-paragraph">REST API is also enabled, so you can connect via <em>/wp-json/chatgpt-tracker/v1/download-data/</em> and use automated exports to an external database to include the data into your monitoring dashboards. Its a simple dump of the full dataset. I may update this feature in the future if there is enough interest for it.</p>



<p class="wp-block-paragraph">Any feature requests? PM me on <a href="https://www.linkedin.com/in/jbobbink/#seo">LinkedIn</a> or <a href="https://twitter.com/jbobbink">Twitter</a> or leave a comment at the bottom. </p>



<p class="wp-block-paragraph"><strong>For updates (if useragents or IPs change for example), follow me on LinkedIn or Twitter. </strong>I&#8217;m trying to get the plugin into the official WordPress repository as soon as possible which enables auto-updates too.</p>



<h2 class="wp-block-heading">Tracking on the Edge</h2>



<p class="wp-block-paragraph">When there is a CDN or caching in between the requests and the server, you may want to use a worker to put it in between and catch the requests. Quick example;</p>



<pre class="wp-block-code"><code>// Cloudflare Worker Script to track requests with User-Agent "GPTBot" or "ChatGPT-User"

addEventListener('fetch', event =&gt; {
  event.respondWith(handleRequest(event.request));
});

async function handleRequest(request) {
  // Extract the User-Agent from the request headers
  const userAgent = request.headers.get('User-Agent');

  // Check if the User-Agent is "GPTBot" or "ChatGPT-User"
  if (userAgent === "GPTBot" || userAgent === "ChatGPT-User") {
    // Handle the tracking logic here
    // Log the request, count it, send to analytics, etc.
    console.log(`Detected bot request from ${userAgent}:`, new Date().toISOString());

    // Insert logic for tracking, alerting, or handling the bot request. You could sent the request into Google Analytics for example
  }

  // Proceed with the request
  return fetch(request);
}
</code></pre>



<h2 class="wp-block-heading">Frequently Asked Questions</h2>



<p class="wp-block-paragraph"><strong>How can I test if the tracker works?</strong></p>



<p class="wp-block-paragraph">The easiest way to test the plugin is to go to the <a href="https://chat.openai.com/">ChatGPT 4 interface</a> and request it to summarize one of your latest URLs on your website. Make sure it actually requests the URL. It could be Bing has already crawled and stored the contents so that will be used instead of visiting the live URL. Make sure it actually shows it is <strong>browsing:</strong></p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="961" height="210" src="https://www.notprovided.eu/assets/chatgpt4-browsing-a-swebsite.gif" alt="" class="wp-image-1060"/></figure>



<p class="wp-block-paragraph">You can also set your own browser with a useragent containing GPTBot or ChatGPT. You will notice those hits will be documented as invalid since the IP address will not match OpenAI&#8217;s ones.</p>



<p class="wp-block-paragraph"><strong>Does it have any privacy related impact?</strong></p>



<p class="wp-block-paragraph">No, it doesn&#8217;t impact any privacy related matters since the plugin only tracks and documents user-agents and IP adresses from validated sources.</p>



<p class="wp-block-paragraph"><strong>What is the difference between crawling and browsing behaviour?</strong></p>



<p class="wp-block-paragraph">More information about the behaviour of the different bots can be found on <a href="https://platform.openai.com/docs/plugins/bot">ChatGPT-User browsing</a> and <a href="https://platform.openai.com/docs/gptbot">GPTBot crawling</a></p>



<p class="wp-block-paragraph"><strong>Can I use this data in external reporting?</strong></p>



<p class="wp-block-paragraph">Yes, you definitely can. You can use the REST API. The plugin has a specific endpoint enabled.</p>



<h2 class="wp-block-heading">Known issues / Feature requests</h2>



<ol class="wp-block-list">
<li>Feature: Referral traffic: do people click on the mentioned URLs from chat.openai.com to your website</li>



<li>Issue: When your site is using a CDN like Cloudflare, it reports their IP addresses</li>
</ol>



<h2 class="wp-block-heading">Changelog</h2>



<p class="wp-block-paragraph">= 1.0 =</p>



<ul class="wp-block-list">
<li>Added a download functionality</li>



<li>Added a simple graph plotting the last 28 days of hits.</li>
</ul>



<p class="wp-block-paragraph">= 0.5 =</p>



<ul class="wp-block-list">
<li>Basic functionality</li>
</ul>
]]></content:encoded>
</item>
<item>
<title>Dealing with user intent in a time where Google depends on AI – SMX Advanced Berlin 2023</title>
<link>https://www.notprovided.eu/dealing-with-user-intent-in-a-time-where-google-depends-on-ai-smx-advanced-berlin-2023/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1018</guid>
<pubDate>Thu, 14 Sep 2023 16:00:00 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[During the core updates Google rolls out every year, SEOs often discuss the impact of vague concepts like EEAT, content relevancy and the number of external links. In many cases the reason of loosing visibility in search engines is simple: Google interpreted…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">During the core updates Google rolls out every year, SEOs often discuss the impact of vague concepts like EEAT, content relevancy and the number of external links. In many cases the reason of loosing visibility in search engines is simple: Google interpreted a search query different, causing a shift in conceived user intent. Target pages do not align with the intended user intent, hence dropping out of the top 5.</p>







<div class="embed" data-embed="https://docs.google.com/presentation/d/e/2PACX-1vR0LyuF3ndWLJM0a8AJ3PXmL4xLDQM2_YJYgo6wGXiML5uIGSOSNUVzvSvdMr16swFcHByhGhzvTDXg/embed?start=false&amp;loop=false&amp;delayms=3000" data-embed-title="Google Slides slides" style="aspect-ratio:960 / 569"><button class="embed__play" type="button"><svg viewBox="0 0 24 24" fill="currentColor" aria-hidden="true"><path d="M8 5.5v13l10.5-6.5z"/></svg><span>Load the slides from Google Slides</span></button><p class="embed__note">Google Slides may set cookies once it loads. <a href="https://docs.google.com/presentation/d/e/2PACX-1vR0LyuF3ndWLJM0a8AJ3PXmL4xLDQM2_YJYgo6wGXiML5uIGSOSNUVzvSvdMr16swFcHByhGhzvTDXg/pub" rel="noopener" target="_blank">Open on Google Slides</a></p></div>



<p class="wp-block-paragraph">All scripts used can be found on <a href="https://colab.research.google.com/drive/1dL7tHt-9xJwiQ8YSnEmTA7z8kLHKNpvw?usp=sharing" target="_blank" rel="noreferrer noopener">https://colab.research.google.com/drive/1dL7tHt-9xJwiQ8YSnEmTA7z8kLHKNpvw?usp=sharing </a></p>



<p class="wp-block-paragraph">User intent is often simplified: what kind of content types show up for a specific set of user intents. Traditionally we have been categorizing keywords in four different types: (content from GPT4; Prompt: <em><strong>What are search intents?</strong></em>)</p>



<p class="wp-block-paragraph"><em>There are typically four primary categories of search intent:</em></p>



<p class="wp-block-paragraph"><em>Informational Intent: The user is looking for more information on a particular topic, question, or idea. Examples of such queries might be &#8220;how does photosynthesis work?&#8221; or &#8220;history of the Eiffel Tower.&#8221;</em></p>



<p class="wp-block-paragraph"><em>Navigational Intent: The user knows which site they want to visit and is using the search engine to get there. For instance, typing &#8220;Facebook login&#8221; or &#8220;OpenAI website&#8221; indicates that the user is likely trying to navigate to those specific sites.</em></p>



<p class="wp-block-paragraph"><em>Transactional Intent: The user is ready to buy or is looking to perform some type of web-based activity, like shopping. Searches like &#8220;buy iPhone 12&#8221; or &#8220;best prices for running shoes&#8221; typically fall under this category.</em></p>



<p class="wp-block-paragraph"><em>Commercial Investigation: The user is looking to make a purchase soon but wants to compare and review products before making a final decision. Examples might include &#8220;best DSLR cameras 2021&#8221; or &#8220;top-rated smartphones.&#8221;</em></p>



<p class="wp-block-paragraph">&nbsp;Looking to some random search result pages across languages and regions shows we need a much more specific set of intents. Using <a href="https://en.wikipedia.org/wiki/Large_language_model">Large Language Models</a> (LLMs) to predict search intents is possible, but it is just guessing. Lots of relevant context is missing which results in non actionable output. An example based on an input of four different keywords:</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="875" height="568" src="https://www.notprovided.eu/assets/gpt4-search-intent-guessing.jpg" alt="" class="wp-image-1019" srcset="https://www.notprovided.eu/assets/gpt4-search-intent-guessing.jpg 875w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-300x195.jpg 300w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-768x499.jpg 768w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-320x208.jpg 320w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-640x415.jpg 640w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-360x234.jpg 360w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-720x467.jpg 720w, https://www.notprovided.eu/assets/gpt4-search-intent-guessing-800x519.jpg 800w" sizes="auto, (max-width: 875px) 100vw, 875px" /></figure>



<p class="wp-block-paragraph">If you try some other prompts, like “What is the search intent for the keyword “rank tracking software”? the model actually acknowledges it is guessing. Look at all those words pointing to the uncertainty:</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="820" height="589" src="https://www.notprovided.eu/assets/llm-guessing-intents.jpg" alt="" class="wp-image-1020" srcset="https://www.notprovided.eu/assets/llm-guessing-intents.jpg 820w, https://www.notprovided.eu/assets/llm-guessing-intents-300x215.jpg 300w, https://www.notprovided.eu/assets/llm-guessing-intents-768x552.jpg 768w, https://www.notprovided.eu/assets/llm-guessing-intents-320x230.jpg 320w, https://www.notprovided.eu/assets/llm-guessing-intents-640x460.jpg 640w, https://www.notprovided.eu/assets/llm-guessing-intents-360x259.jpg 360w, https://www.notprovided.eu/assets/llm-guessing-intents-720x517.jpg 720w, https://www.notprovided.eu/assets/llm-guessing-intents-800x575.jpg 800w" sizes="auto, (max-width: 820px) 100vw, 820px" /></figure>



<p class="wp-block-paragraph">But we are a bunch of lucky people: according to GPT4, search engines have become much better in determining what their users expect. Prompt: <em>“What are search intents?”</em> gives us a nice hint were to start looking when we need a systematical approach to search intent.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="834" height="200" src="https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent.jpg" alt="" class="wp-image-1021" srcset="https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent.jpg 834w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-300x72.jpg 300w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-768x184.jpg 768w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-320x77.jpg 320w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-640x153.jpg 640w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-360x86.jpg 360w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-720x173.jpg 720w, https://www.notprovided.eu/assets/search-engines-are-better-in-search-intent-800x192.jpg 800w" sizes="auto, (max-width: 834px) 100vw, 834px" /></figure>



<p class="wp-block-paragraph">Lets agree that Google as a platform is much better in this than 3<sup>rd</sup> party tools that have less user based metrics and datasets. Even if Google’s algorithms are not doing a good job, we still want to create the highest chance of ranking in the top spots. So let’s go back to the actual search engine result pages and use that as a starting point to create a better tailored system:</p>



<ol class="wp-block-list" type="1">
<li>Create a database with keywords and ranking URLs</li>



<li>Collect body content and classify URLs</li>



<li>Identify best ranking content type</li>



<li>Create follow up tasks</li>



<li>(Bi-) Weekly &amp; Monthly repetition</li>
</ol>



<h2 class="wp-block-heading">Create database with keywords and ranking URLs</h2>



<p class="wp-block-paragraph">One thing I can be very clear about, do not use browse plugins like <a href="https://voxscript.awt.icu/index.html">VoxScript</a> or <a href="https://www.keymate.ai/">Keymate.ai</a> in combination with a GPT model to scrape Google search engine result pages (SERPs). That is a painfully slow and expensive process. It may work for one-time checks but goes wrong for repetitive tasks. There are dedicated API’s from standalone services available that can check SERPs in a low cost and faster setup. They take care of randomized IP’s and browser fingerprints and have a focused approach on getting the data for you.</p>



<p class="wp-block-paragraph">Generally speaking their performance is fine for small batches like a few thousands keywords. Just to name a few to get an idea for the costs involved:</p>



<figure class="wp-block-table"><table><tbody><tr><td><strong>Provider</strong></td><td><strong>Avg. Costs</strong></td></tr><tr><td>Valueserp.com &nbsp;</td><td>$0,50 / 1000</td></tr><tr><td>Dataforseo.com &nbsp;</td><td>$0,60 / 1000</td></tr><tr><td>Oxylabs.io &nbsp;&nbsp;</td><td>$2,80 / 1000</td></tr><tr><td>Serpstack.com &nbsp;&nbsp;</td><td>$6,00 / 1000</td></tr><tr><td>Serpapi.com </td><td>$8,33 &nbsp;/ 1000</td></tr></tbody></table></figure>



<p class="wp-block-paragraph">The first step is to create a database containing the top 30 URLs ranking for the set of relevant keywords you or your client wants to rank for. For the sake of simplicity of the examples, we just want to focus on the organic ranking URLs and we do not include SERP elements like knowledge graph blocks, featured snippets, Google Ads etcetera. It is recommended for your final process to include feature snippets since they show a (sub)intent related to the <em>informational intent</em>.</p>



<p class="wp-block-paragraph">Using the <a href="https://www.valueserp.com/docs/search-api/overview">SERP API</a> from Valueserp.com we can use the following script to get the top 30 URLs per keyword:</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="991" height="597" src="https://www.notprovided.eu/assets/keyword-ranking-python.jpg" alt="" class="wp-image-1034" srcset="https://www.notprovided.eu/assets/keyword-ranking-python.jpg 991w, https://www.notprovided.eu/assets/keyword-ranking-python-300x181.jpg 300w, https://www.notprovided.eu/assets/keyword-ranking-python-768x463.jpg 768w, https://www.notprovided.eu/assets/keyword-ranking-python-320x193.jpg 320w, https://www.notprovided.eu/assets/keyword-ranking-python-640x386.jpg 640w, https://www.notprovided.eu/assets/keyword-ranking-python-360x217.jpg 360w, https://www.notprovided.eu/assets/keyword-ranking-python-720x434.jpg 720w, https://www.notprovided.eu/assets/keyword-ranking-python-800x482.jpg 800w" sizes="auto, (max-width: 991px) 100vw, 991px" /></figure>



<p class="wp-block-paragraph">Once you have the SERP data collected, I advice to recheck rankings every two weeks. Relevant shifts are only visible if you zoom out for longer periods and usually happen around core updates. You can always manually update the dataset if you know a core algorithm update has been rolled out and rankings have shifted considerably. The example below shows how relatively stable the top10 rankings are for a 6 month period until the core update hits the SERPs:</p>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="285" src="https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-1024x285.jpg" alt="" class="wp-image-1022" srcset="https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-1024x285.jpg 1024w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-300x84.jpg 300w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-768x214.jpg 768w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-320x89.jpg 320w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-640x178.jpg 640w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-360x100.jpg 360w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-720x201.jpg 720w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-1080x301.jpg 1080w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software-800x223.jpg 800w, https://www.notprovided.eu/assets/SERP-volatility-rank-tracking-software.jpg 1148w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<p class="wp-block-paragraph">I have calculated volatility numbers for different timeframes based on historical datapoints to check the most relevant period to use. It turns out checking the ranking URLs every <strong>2 weeks</strong> is fine. Yes, you will miss the occasional newcomer but 93% of the relevant changes are visible when using a 2 week period.</p>



<h2 class="wp-block-heading">Collecting body content</h2>



<p class="wp-block-paragraph">Since we now have a nice dataset with keyword and ranking URLs, we can start looking to the content on those URLs. For the examples below I will use a manual input of URLs but you can easily connect the database from above. Ask GPT4 to update the Python script for you if you are lazy 🙂</p>



<p class="wp-block-paragraph">The quickest way is to use Python with <em>BeautifulSoup</em> to request a URL, scrape the body content and go to the next URL. Unfortunately that doesn’t work as well as it used to do due to many CDN solutions have built-in scraper / bots detection algorithms running. To circumvent that partly, it is important to fire up a real browser and mimic real users. The easiest solution is Selenium which also enables client side JS rendering. Since you can set all browser variables and combine it with <a href="https://pypi.org/project/selenium-stealth/"><em>selenium-stealth</em></a> or <a href="https://github.com/ultrafunkamsterdam/undetected-chromedriver"><em>undetected_chromedriver</em></a> to have higher success rates. In 90-95% of the URL requests this prevents getting a “are you a robot, dude?” or “Forbidden” response.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="694" height="739" src="https://www.notprovided.eu/assets/selenium-python.jpg" alt="" class="wp-image-1035" srcset="https://www.notprovided.eu/assets/selenium-python.jpg 694w, https://www.notprovided.eu/assets/selenium-python-282x300.jpg 282w, https://www.notprovided.eu/assets/selenium-python-320x341.jpg 320w, https://www.notprovided.eu/assets/selenium-python-640x681.jpg 640w, https://www.notprovided.eu/assets/selenium-python-360x383.jpg 360w" sizes="auto, (max-width: 694px) 100vw, 694px" /></figure>



<p class="wp-block-paragraph">For websites that have set up higher level of protection (usually well ranking websites have better anti bot measures in place unfortunately) it is recommended to pay a little extra for dedicated services like like <a href="https://www.zenrows.com/">ZenRows</a> or <a href="https://www.webscrapingapi.com/">WebScraping API</a> for about $50 / month for 80-100K URL requests. If you zoom in to the top 5 ranking pages to determine most successful content types, you need to be sure you have actually collected the content of those 5 URLs.</p>



<p class="wp-block-paragraph">In the colab you can find a basic script. This script can be extended with an additional loop to use the external APIs to request URLs that return a blocked message initially.</p>



<p class="wp-block-paragraph">To make things more effective, you should also built in a check to see if a specific URL has already been scraped for another keyword. This prevents scraping individual URLs multiple times.</p>



<h3 class="wp-block-heading">Additional notes for collecting body content</h3>



<p class="wp-block-paragraph">In live environments I have built in some output cleaning like removing comment sections and preventing inclusion of content from <em>header, nav</em> and <em>footer</em> elements. Some default CMS have “skip to&nbsp; main content” links and a table of contents, remove those.</p>



<p class="wp-block-paragraph">Be aware that some of the models that you can use for classifying have an input limit of 4096 tokens and having the cleaned up content as input performs better with the classifying process.</p>



<h2 class="wp-block-heading">Classify content into content types</h2>



<p class="wp-block-paragraph">Most LLM models are not specifically trained for classifying a piece of content based on a set of previous unknown rules (we come up with the input of a number of content types). &nbsp;So you can try inputting a prompt like “Give me the search intent for the following keywords: nike, Jordan 4, sneakers, air forces’ but the results are not something you should depend on in future analysis because it is missing relevant content.</p>



<p class="wp-block-paragraph">Since you need to process lots of content pieces, latency can be of importance. A nice monitoring dashboard for current latency averages for popular LLMs can be found at <a href="https://gptforwork.com/tools/openai-api-and-other-llm-apis-response-time-tracker">https://gptforwork.com/tools/openai-api-and-other-llm-apis-response-time-tracker</a> &nbsp;Classifying 15000 URLs / content pieces with a latency of 10s per request means a total processing time of 42 hours.</p>



<figure class="wp-block-image size-full"><img loading="lazy" decoding="async" width="832" height="747" src="https://www.notprovided.eu/assets/python-classifiying-content.jpg" alt="" class="wp-image-1036" srcset="https://www.notprovided.eu/assets/python-classifiying-content.jpg 832w, https://www.notprovided.eu/assets/python-classifiying-content-300x269.jpg 300w, https://www.notprovided.eu/assets/python-classifiying-content-768x690.jpg 768w, https://www.notprovided.eu/assets/python-classifiying-content-320x287.jpg 320w, https://www.notprovided.eu/assets/python-classifiying-content-640x575.jpg 640w, https://www.notprovided.eu/assets/python-classifiying-content-360x323.jpg 360w, https://www.notprovided.eu/assets/python-classifiying-content-720x646.jpg 720w, https://www.notprovided.eu/assets/python-classifiying-content-800x718.jpg 800w" sizes="auto, (max-width: 832px) 100vw, 832px" /></figure>



<h3 class="wp-block-heading">Hosting your own LLM</h3>



<p class="wp-block-paragraph">Latency was my biggest annoyance during building my first setup, I started using locally hosted LLMs. You can easily set up <a href="https://gpt4all.io/index.html">GPT4ALL</a> on your local PC. It can be loaded up with a few well known models out of the box. The greatest thing is you can easily switch models or use your own model. Not every existing model is suited for classifying: on <a href="https://github.com/Troyanovsky/Local-LLM-Comparison-Colab-UI">https://github.com/Troyanovsky/Local-LLM-Comparison-Colab-UI</a> you can find an overview and scores given for 68 different models. Models where tested for specific tasks like translating, summarizing, answering, rewriting, sentiment analysis etcetera.</p>



<h3 class="wp-block-heading">Finetuning models</h3>



<p class="wp-block-paragraph">Models available are not specifically trained and suited for classifying content into a unique set of content types like we SEOs need. To help them out, we can finetune the current models without needing to fully train and built out a full model. <a href="https://platform.openai.com/docs/guides/fine-tuning/preparing-your-dataset">https://platform.openai.com/docs/guides/fine-tuning/preparing-your-dataset</a> shows how to prepare your own dataset (manually classified content dataset in this case) and how to finetune an existing model. OpenAI recommends 50-100 training examples, but I have done test with 500-1000 examples to finetune <em>gpt-3.5-turbo</em> to get to a usable model for my setup:</p>



<figure class="wp-block-table"><table><tbody><tr><td>&nbsp;<strong>URL</strong></td><td><strong>GPT 4</strong></td><td><strong>Finetuned GPT 3.5T</strong></td></tr><tr><td>https://startupbonsai.com/rank-tracking-tools/</td><td>Comparative lists</td><td>Comparative lists</td></tr><tr><td>https://bloggingwizard.com/rank-tracking-tools/</td><td>Product pages</td><td>Comparative lists</td></tr><tr><td>https://diggitymarketing.com/best-rank-tracker-tool/</td><td>Guides</td><td>Comparative lists</td></tr><tr><td>https://www.searchenginejournal.com/seo-tools/rank-tracking-tools/</td><td>Guides</td><td>Guides</td></tr><tr><td>https://searchengineland.com/keyword-rank-tracking-software-6-tools-compared-387503</td><td>Comparative lists</td><td>Comparative lists</td></tr><tr><td>https://nealschaffer.com/best-rank-tracker-tool/</td><td>Guides</td><td>Guides</td></tr><tr><td>https://www.link-assistant.com/rank-tracker/</td><td>Product pages</td><td>Commercial</td></tr><tr><td>https://www.g2.com/software/rank-tracking</td><td>Commercial</td><td>Comparative lists</td></tr><tr><td>https://www.ranktracker.com/</td><td>Product pages</td><td>Product pages</td></tr><tr><td>https://www.adamenfroy.com/rank-tracking-software</td><td>Reviews</td><td>Comparative lists</td></tr></tbody></table></figure>



<h2 class="wp-block-heading">Speeding up processes</h2>



<p class="wp-block-paragraph">All the above scripts run in a linear setup. They process one SERP, one URL, one piece of content and then process the output. Once the output has been processed, it goes to the next SERP, URL etcetera. Going one by one is not most time effective, so have a look at possibilities for running tasks in parallel. A good start is reading up on the differences (and limitations) of <a href="https://www.geeksforgeeks.org/difference-between-multithreading-vs-multiprocessing-in-python/">Multithreading and Multiprocessing in Python</a></p>



<h2 class="wp-block-heading">Create follow up tasks</h2>



<p class="wp-block-paragraph">A rich database with classified content is what we now have at our disposal. Now we can determine the best ranking content type per keyword. This is the most simple task: check for the most common content type amongst the top 5, excluding the target URL of your or your client’s website.</p>



<p class="wp-block-paragraph">If there is no clear winner (like two A types, two B types and one C type) or there are unique content types in the top 5, extend your selection to the top 10. If that doesn’t give a clear winner, extend to top 20.</p>



<p class="wp-block-paragraph">Upfront I already have clustered a set of relevant keywords to target one specific URL. We can compare the content type of the current or proposed target URL with the content type that has the highest chances for ranking high.</p>



<p class="wp-block-paragraph">If they align, you have to look to other SEO activities to improve rankings. External links, optimizing the existing content, internal linking etcetera. If content type does not align, you can go on with creating a content briefing for your text writers.</p>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="239" src="https://www.notprovided.eu/assets/python-generate-briefings-1024x239.jpg" alt="" class="wp-image-1037" srcset="https://www.notprovided.eu/assets/python-generate-briefings-1024x239.jpg 1024w, https://www.notprovided.eu/assets/python-generate-briefings-300x70.jpg 300w, https://www.notprovided.eu/assets/python-generate-briefings-768x180.jpg 768w, https://www.notprovided.eu/assets/python-generate-briefings-320x75.jpg 320w, https://www.notprovided.eu/assets/python-generate-briefings-640x150.jpg 640w, https://www.notprovided.eu/assets/python-generate-briefings-360x84.jpg 360w, https://www.notprovided.eu/assets/python-generate-briefings-720x168.jpg 720w, https://www.notprovided.eu/assets/python-generate-briefings-1080x253.jpg 1080w, https://www.notprovided.eu/assets/python-generate-briefings-800x187.jpg 800w, https://www.notprovided.eu/assets/python-generate-briefings.jpg 1253w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<p class="wp-block-paragraph">One example to show you shouldn’t use LLMs to write content, that’s all I want to say about it in this article. (Oh yes, I know you can use LLMs for specific text generation processes, but not for this). Out of the 5 tools listed in the blog, 3 of them didn’t exist. If you want to work with factual information, make sure you either include the facts in your prompt or ask a human being.</p>



<figure class="wp-block-image size-large"><img loading="lazy" decoding="async" width="1024" height="378" src="https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-1024x378.jpg" alt="" class="wp-image-1028" srcset="https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-1024x378.jpg 1024w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-300x111.jpg 300w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-768x284.jpg 768w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-1536x567.jpg 1536w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-320x118.jpg 320w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-640x236.jpg 640w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-360x133.jpg 360w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-720x266.jpg 720w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-1080x399.jpg 1080w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-800x296.jpg 800w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist-1280x473.jpg 1280w, https://www.notprovided.eu/assets/seomaster-pro-doesnt-exist.jpg 1784w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /></figure>



<h2 class="wp-block-heading">Bi Weekly &amp; Monthly repetition</h2>



<p class="wp-block-paragraph">As said above, a 2 week period often catches most ranking changes relevant to intent changes. Content changes on the actual ranking website are a bit different between niches and the type of content ranking. You may want to built in a check for last-modified headers or tags to prevent unnecessary scraping activities and make the whole system run faster.</p>



<p class="wp-block-paragraph">Summarized, I run this in the following configuration:</p>



<ol class="wp-block-list" type="1">
<li>(Bi-) Weekly repetition:<ol><li>Request top 30 rankings, check for new URLs<ol><li>Calculate best ranking URLs over the past 2 weeks</li></ol></li></ol><ol><li>Download content &amp; classify newly found URLs</li></ol>
<ol class="wp-block-list">
<li>Check if the top 5 and your project target URL has changed
<ol class="wp-block-list">
<li>If yes, adjust advice for content team</li>
</ol>
</li>
</ol>
</li>



<li>Monthly repetition
<ul class="wp-block-list">
<li><span style="color: initial;">Request top 30 rankings</span></li>



<li><span style="color: initial;">Download and classify all URLs</span></li>



<li><span style="color: initial;">Calculate best ranking URLs over the past 2 weeks</span></li>



<li>Classify content for best ranking URL of your project, if availabe</li>



<li><span style="color: initial;">Adjust content requirements and briefs if necessary</span></li>
</ul>
</li>
</ol>



<h2 class="wp-block-heading">Evaluation of the system</h2>



<p class="wp-block-paragraph">For my first setup I spent 22 hours for a 5000 keyword set. It turns out to save me lots of time (vs manual checks and briefs) which equals less costs for the client or more time available for other analysis. The great thing about this setup is that the data you collect is very valuable.</p>



<p class="wp-block-paragraph">There is lots of tweaking and additions possible in the system. Since you have the content of the best ranking websites available, you can extract all entities, topics, keywords, etcetera in there and see if your target URL also has these covered. Combine the dataset with external link data and you can push for automated link building briefs (eg. This glossary page requires at least 5 more external backlinks with a minimum requirement of X, Y, Z.)</p>



<p class="wp-block-paragraph">Looking to general trends within a industry, like blogs starting to rank instead of e-commerce category pages, may justify a change in SEO strategy and requires a change in internal UX and taxonomy structure. Those are decisions where I think the value of using AI stops and our human brains need to take over.</p>



<p class="wp-block-paragraph"></p>



<p class="wp-block-paragraph">All scripts used can be found on <a href="https://colab.research.google.com/drive/1dL7tHt-9xJwiQ8YSnEmTA7z8kLHKNpvw?usp=sharing" target="_blank" rel="noreferrer noopener">https://colab.research.google.com/drive/1dL7tHt-9xJwiQ8YSnEmTA7z8kLHKNpvw?usp=sharing </a></p>
]]></content:encoded>
</item>
<item>
<title>Search Seekers – How to keep growing organic traffic for already well ranking websites</title>
<link>https://www.notprovided.eu/search-seekers-how-to-keep-growing-organic-traffic-for-already-well-ranking-websites/</link>
<guid isPermaLink="false">https://www.notprovided.eu/?p=1008</guid>
<pubDate>Wed, 14 Jun 2023 10:28:17 GMT</pubDate>
<dc:creator>Jan-Willem Bobbink</dc:creator>
<category>Technical SEO</category>
<description><![CDATA[During Searchseekers.de I shared 10 #SEO tactics I use for websites that already rank well but have plenty of opportunities to grow. Any questions? Feel free to comment or DM! 1. Are you actually properly indexed? 2. Pruning your non valuable content…]]></description>
<content:encoded><![CDATA[
<p class="wp-block-paragraph">During&nbsp;<a href="http://searchseekers.de/">Searchseekers.de</a>&nbsp;I shared 10&nbsp;<a href="https://www.linkedin.com/feed/hashtag/?keywords=seo&amp;highlightedUpdateUrns=urn%3Ali%3Aactivity%3A7074690744849969152">#SEO</a>&nbsp;tactics I use for websites that already rank well but have plenty of opportunities to grow. Any questions? Feel free to comment or DM!</p>



<p class="wp-block-paragraph">1. Are you actually properly indexed?<br>2.&nbsp;Pruning your non valuable content<br>3.&nbsp;Examine the index strategies<br>4.&nbsp;Using Internal link value effectively<br>5.&nbsp;Crawl based optimization<br>6.&nbsp;Scaling content: covering all search intents<br>7.&nbsp;Focus on other Google based channels<br>8.&nbsp;Using subdomains effectively<br>9.&nbsp;Multi domain strategy<br>10.&nbsp;Utilizing communities</p>



<div class="embed" data-embed="https://docs.google.com/presentation/d/e/2PACX-1vQrhVWUM2X6TzmlgBJZd-oA1lTONvmr-Y4NhEBl6rr7AIW2m6gyyu5NRcOM63F_rL62X86U1CHMGj6j/embed?start=false&amp;loop=false&amp;delayms=3000" data-embed-title="Google Slides slides" style="aspect-ratio:960 / 569"><button class="embed__play" type="button"><svg viewBox="0 0 24 24" fill="currentColor" aria-hidden="true"><path d="M8 5.5v13l10.5-6.5z"/></svg><span>Load the slides from Google Slides</span></button><p class="embed__note">Google Slides may set cookies once it loads. <a href="https://docs.google.com/presentation/d/e/2PACX-1vQrhVWUM2X6TzmlgBJZd-oA1lTONvmr-Y4NhEBl6rr7AIW2m6gyyu5NRcOM63F_rL62X86U1CHMGj6j/pub" rel="noopener" target="_blank">Open on Google Slides</a></p></div>
]]></content:encoded>
</item>
</channel>
</rss>
