{"id":4207,"date":"2026-11-10T17:00:00","date_gmt":"2026-11-10T15:00:00","guid":{"rendered":"https:\/\/latevaia.ai\/blog\/ai-dashboard-to-analyze-your-website\/"},"modified":"2026-11-10T17:00:00","modified_gmt":"2026-11-10T15:00:00","slug":"ai-dashboard-to-analyze-your-website","status":"publish","type":"post","link":"https:\/\/latevaia.ai\/en\/blog\/ai-dashboard-to-analyze-your-website\/","title":{"rendered":"AI dashboard to analyze your website"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">Monday morning, the owner of an online store opens Google Analytics, then Search Console, then the store&#8217;s orders dashboard, and half an hour later has three tabs showing three different sales figures and no clear idea of what happened last week. <strong>There&#8217;s no shortage of data, it&#8217;s just scattered<\/strong> and nobody has time to pull it together.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This guide explains how to bring those sources together in a Looker Studio dashboard and how to add <strong>an AI layer that reads the numbers every week<\/strong>, compares them and sends you a short summary by email or Slack.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Why put your data in a dashboard<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Think about the last time someone on your team asked you whether the sales campaign worked. To answer with data, you need to cross-reference the visits that came in, where they came from and how many ended in an order, and each of those three things lives in a different tool. If cross-referencing them takes hours, the question gets answered by gut feeling, and <strong>gut-feeling decisions are the ones that cost you<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A single dashboard solves that in a rather unglamorous way, with <strong>the same metrics and definitions in the same place, updated automatically<\/strong>. When everyone looks at the same number, <strong>the argument about which figure is the right one is over<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">There&#8217;s a second reason. For an AI to analyze your data and flag what changes, that data has to be organized first, and <strong>the dashboard is that organization<\/strong>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Which metrics to include and which to leave out<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>The most common mistake in a first dashboard is putting everything in<\/strong>, because GA4 offers a huge number of metrics and it feels wrong to leave any out. The result is a screen full of charts nobody looks at.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We <strong>narrow it down to three families<\/strong>, with just a few metrics in each:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Traffic<\/strong>: users, sessions and the source channel (organic search, paid campaigns, social, direct). If you have Search Console, add Google clicks and impressions, which tell you how many people saw you before coming in.<\/li>\n\n\n\n<li><strong>Engagement<\/strong>: engaged sessions and most viewed pages. They tell you whether visitors find what they were looking for.<\/li>\n\n\n\n<li><strong>Conversion<\/strong>: the action that makes you money, which in a store is a purchase and on a services website is usually the contact form or a phone call. This is where the number of conversions, the conversion rate and, if you sell online, revenue go.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Leave out metrics that go up and down without changing any decision, such as total page views or the breakdown by browser, unless you&#8217;re investigating a technical problem. <strong>If a metric doubling<\/strong> or dropping by half wouldn&#8217;t make you change anything, it doesn&#8217;t belong on the main dashboard.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Before opening any tool, write down on paper the questions you want the dashboard to answer every Monday, such as \u201cwhere are orders coming from\u201d or \u201cwhich pages get traffic from Google but don&#8217;t convert\u201d. <strong>Every chart in the dashboard should answer one of them<\/strong>, and if a chart doesn&#8217;t answer any, it goes.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Connecting your sources to Looker Studio<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Looker Studio is Google&#8217;s free visualization tool, and for an SME it&#8217;s the logical starting point because it already has its own connectors for GA4, Search Console, Google Sheets and BigQuery. The online store is another story: WooCommerce, Shopify and PrestaShop <strong>don&#8217;t have a native Google connector<\/strong>, so you have to go through a third-party connector, often a paid one, or export orders to a spreadsheet that Looker Studio can read.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">GA4 connector, Google Sheets or BigQuery<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The native GA4 connector is the quickest option: you pick it, select the property and you have data. For a website with moderate traffic and a dashboard with only a few charts, it&#8217;s enough. Its limits show up as the dashboard grows, because every chart queries the Google Analytics API, the API has quotas, and a large dashboard with several people looking at it at the same time <strong>can throw quota errors or load slowly<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Exporting to Google Sheets is the middle ground. <a href=\"https:\/\/latevaia.ai\/en\/blog\/n8n-automation\/\">An n8n workflow<\/a> or a Make scenario, or a Sheets add-on, pulls the metrics you care about every night and drops them into a sheet, and Looker Studio reads from the sheet instead of querying GA4 on every load. <strong>This is the option we recommend most for an SME<\/strong>, because it&#8217;s cheap and keeps the data where the AI will read it later too. <strong>Its limit is size<\/strong>: a spreadsheet isn&#8217;t built for years of data per page per day.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">BigQuery is the option for when volume really matters, or when you want raw data, event by event. GA4 links to BigQuery from the Analytics admin, and from there you can cross-reference anything. In exchange, <strong>you need someone who writes SQL<\/strong>, and the cost depends on how much you store and query, so check Google Cloud pricing before turning it on.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Combining sources without the numbers drifting apart<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Looker Studio lets you blend sources in a single chart, and this is where <strong>many dashboards start lying without anyone noticing<\/strong>. GA4, Search Console and the store measure similar things with different definitions. A click in Search Console isn&#8217;t a session in GA4, and a purchase in GA4 isn&#8217;t an order in the store, because the store counts orders even when the customer rejected cookies, and it deducts returns and cancellations that GA4 doesn&#8217;t see.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The practical rule is to <strong>always blend on a simple, shared key<\/strong>, such as the date, with the same granularity in both sources, day with day or week with week. <strong>Also check each tool&#8217;s time zone<\/strong>, because if one cuts the day off at midnight Madrid time and another at midnight in a different zone, the daily totals will never match.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">And <strong>never add up metrics from two sources<\/strong> as if they were the same thing. For revenue, we <strong>take the store as the source of truth<\/strong>, because it sees returns and cancellations, and use GA4 only to see which channel it comes from. Write that decision down in a note inside the dashboard itself, so whoever looks at it later knows where each number comes from.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Building the dashboard: scorecards, charts and filters<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Once the sources are connected, building it is the most rewarding part. Open a new report in Looker Studio, add the data sources you&#8217;ve prepared and start at the top with <strong>a row of scorecards<\/strong>. Use one scorecard per figure: users, conversions, revenue and conversion rate. It&#8217;s worth turning on the comparison with the previous period on all of them, which shows in green or red whether the figure has gone up or down.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"986\" height=\"657\" src=\"https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585-986x657.jpg\" alt=\"building an analytics dashboard in Looker Studio with scorecards and charts\" class=\"wp-image-3833\" title=\"\" srcset=\"https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585-986x657.jpg 986w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585-480x320.jpg 480w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585-768x512.jpg 768w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585-380x253.jpg 380w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-2616585.jpg 1280w\" sizes=\"auto, (max-width: 986px) 100vw, 986px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Below the scorecards goes <strong>a time series chart<\/strong> with the daily or weekly trend of the metric that matters most to you; for a store, conversions or revenue. That&#8217;s where you see peaks, drops and seasonality at a glance, things a scorecard with a total doesn&#8217;t show.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The third layer is tables. One with the source channels and their conversions, another with the pages that get the most traffic from Google, using Search Console data. Sort them by the metric that matters and <strong>limit the rows<\/strong>; nobody reads an endless table.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Finally, filters. A date range control at the very top so anyone can change the dates and, if it makes sense in your case, a drop-down by channel or device type. Keep them few, because a dashboard full of filters always ends up being used with the default settings.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Before sharing it, <strong>compare a couple of figures from the dashboard<\/strong> with those in the original tool for the same week. If they don&#8217;t match, first check what we covered earlier, definitions or time zone; better for it to come up now than in a meeting.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Weekly AI summary in n8n or Make<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The dashboard saves you from pulling the data together, but someone has to open it and interpret it, and in an SME there are many weeks when nobody opens it. The AI piece is there so that <strong>the analysis comes to you<\/strong>.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"986\" height=\"657\" src=\"https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049-986x657.jpg\" alt=\"getting the weekly web analytics summary on your phone\" class=\"wp-image-3834\" title=\"\" srcset=\"https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049-986x657.jpg 986w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049-480x320.jpg 480w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049-768x512.jpg 768w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049-380x253.jpg 380w, https:\/\/latevaia.ai\/wp-content\/uploads\/2026\/11\/pixabay-5700049.jpg 1280w\" sizes=\"auto, (max-width: 986px) 100vw, 986px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">The workflow has three stages and is built the same way in <a href=\"https:\/\/latevaia.ai\/en\/blog\/n8n-vs-make-vs-zapier\/\">n8n as in Make<\/a>. <strong>A scheduled trigger launches it every Monday<\/strong> first thing in the morning. The first stage pulls the data for the week that has just ended and the one before, either with the Google Analytics node directly against GA4 or by reading the Google Sheets spreadsheet that feeds the dashboard, which has the advantage that <strong>the summary and the dashboard start from exactly the same numbers<\/strong>. The second stage passes that table to a language model, <a href=\"https:\/\/latevaia.ai\/en\/blog\/claude-chatgpt-gemini-ai-for-business\/\">Claude, ChatGPT or Gemini<\/a>, all of which have their own node in both orchestrators. The third sends the response by Gmail or posts it to a Slack channel.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">A prompt that doesn&#8217;t make up numbers<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">This is the main risk in the whole setup. A language model you ask to \u201canalyze this data\u201d can hand you a very convincing text with a percentage that appears nowhere, or with an explanation for the drop that sounds right and is pure guesswork. That&#8217;s why <strong>we don&#8217;t leave the calculations to the model<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Calculate the changes beforehand<\/strong>, in a code node or in the sheet itself, and give the model a finished table with the metric, this week&#8217;s value, last week&#8217;s and the difference. <strong>Its job is to write and point out what matters, not to do the math<\/strong>. In the prompt, tell it that it&#8217;s an SME&#8217;s analyst, that the table already includes the calculated change, and give it four rules:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Use only the figures in the table, without calculating or rounding anything.<\/li>\n\n\n\n<li>Don&#8217;t comment on small changes.<\/li>\n\n\n\n<li>If the table doesn&#8217;t explain a change, say that it can&#8217;t be known from this data, without guessing at causes.<\/li>\n\n\n\n<li>Reply in English, in five lines at most, with no headings or lists.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The rule about admitting it doesn&#8217;t know why something changed is the one that makes the biggest difference. Without it, the model tends to fill in the gaps, and a summary that blames a drop in sales on \u201cseasonality\u201d with no data to back it up <strong>is worse than no summary at all<\/strong>, because someone is going to believe it. It&#8217;s also worth <strong>checking the first few emails by hand<\/strong> against the dashboard until you trust the format.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">A five-line email<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A long email gets archived unread; <strong>a five-line one gets read on your phone<\/strong> before you reach the office. It should say what went up, what went down, which of those changes deserves attention and where to look, with a link to the dashboard for anyone who wants the details, something like this:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Conversions were up this week compared with last week, almost all from organic search. Store revenue didn&#8217;t follow: the average order value went down. Traffic from paid campaigns dropped and there&#8217;s no data here to explain why. Check whether a campaign was paused or ran out of budget. Breakdown by channel in the dashboard.<\/em><\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Cost and time of an AI dashboard<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Licenses are usually the cheapest part<\/strong>. Looker Studio is free, so are GA4 and Search Console, and Google Sheets comes with any Google account. The costs show up in the store connector, if you choose a third-party one; in the orchestrator, n8n or Make, with paid plans based on executions (n8n can also be self-hosted on your own server); and in the model calls, which don&#8217;t add up to much for a weekly summary of a small table. Prices change often, so check them on each tool&#8217;s website with your use case in mind.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In terms of time, our estimate for an SME with GA4 already properly set up is <strong>a few days of work<\/strong> to connect the sources, build the dashboard, reconcile the numbers and get the summary workflow running. What drags the setup out most often happens before you even open Looker Studio or n8n: <strong>discovering that GA4 wasn&#8217;t tracking conversions properly<\/strong> and that it has to be fixed before building on top of it.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Then there&#8217;s maintenance, which is light but not zero. Connectors change, Google renames metrics, a new campaign needs its channel added to the dashboard. Plan for a quick review every few months and make sure the Monday email keeps arriving, because when a workflow like this breaks, <strong>it usually fails silently<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If you want to know what would make sense to <a href=\"https:\/\/latevaia.ai\/en\/solutions\/automate-with-ai\/\">automate in your case<\/a> and what isn&#8217;t worth it, <a href=\"https:\/\/latevaia.ai\/en\/contact\/\">get in touch<\/a> and we&#8217;ll go through it with your data.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Monday morning, the owner of an online store opens Google Analytics, then Search Console, then the store&#8217;s orders dashboard, and half an hour later has three tabs showing three different sales figures and no clear idea of what happened last week. There&#8217;s no shortage of data, it&#8217;s just scattered and nobody has time to pull [&hellip;]<\/p>\n","protected":false},"author":5,"featured_media":3832,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[56],"tags":[],"class_list":["post-4207","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-web-analytics"],"acf":[],"_links":{"self":[{"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/posts\/4207","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/comments?post=4207"}],"version-history":[{"count":0,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/posts\/4207\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/media\/3832"}],"wp:attachment":[{"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/media?parent=4207"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/categories?post=4207"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/latevaia.ai\/en\/wp-json\/wp\/v2\/tags?post=4207"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}