An SEO dashboard is a screen that brings together the numbers from organic search in one place: how often a site appears on Google, how many people click on it, what position the pages are in and what visitors do once they arrive. The data isn't created there. It comes from Google Search Console, Google Analytics 4 and a rank tracker, and the dashboard lines them up side by side in columns.
That is exactly where the problem starts. Two columns next to each other seem to tell the same story, and they almost never do.
Two numbers on the same row that can't be compared
Take a real case. Macropix makes LED walls and LED screens, it is a Visilay client and its numbers are already public in the case study on manufacturing SEO. On 21 September 2026 the rank tracker monitoring the project showed two columns on the same row: the position recorded on google.it and the estimated traffic. Next to them, the Search Console connector reported real clicks and impressions.
| Search (Google Italy) | Recorded position | Traffic estimated by the tracker | Impressions (Search Console) | Clicks (Search Console) | Actual CTR |
|---|---|---|---|---|---|
| ledwall | 3 | 826 | 2,332 | 28 | 1.20% |
| led wall | 4 | 583 | 1,034 | 12 | 1.16% |
| ledwall prezzo (LED wall price) | 3 | 60 | 504 | 8 | 1.59% |
| ledwall da esterno (outdoor LED wall) | 3 | 73 | 281 | 6 | 2.14% |
| led wall indoor prezzi (indoor LED wall prices) | 3 | 3 | 70 | 2 | 2.86% |
| ledwall esterno (LED wall outdoor) | 2 | 16 | 37 | 1 | 2.70% |
| schermo led wall (LED wall screen) | 3 | 21 | 49 | 0 | 0% |
| schermi led pubblicitari (advertising LED screens) | 4 | 18 | 65 | 0 | 0% |
| Total | 1,600 | 4,372 | 57 |
1,600 against 57. Anyone looking at that screen in a meeting reads two big numbers and two small numbers on the same row and draws the wrong conclusion, such as "SEO doesn't convert" or "the tracker is wrong". Neither. One column estimates how many clicks that position should bring in an average month; the other counts how many clicks actually arrived in the window Search Console returned. They don't share a unit, a period or a source.
The second thing the table shows is even more interesting: position doesn't predict clicks. Four searches in third position give CTRs between 0% and 2.86%. On a second project, an Italian company that leases land for solar farms, two searches both in fifth position give a 12.5% CTR ("affitto terreno per agrivoltaico", land lease for agrivoltaics, 2 clicks from 16 impressions) and 0% ("affitto terreni fotovoltaico", solar land lease, 0 clicks from 24 impressions). Same position, opposite behaviour. If your dashboard shows average position as its main metric, you are looking at the variable that explains the least. CTR and zero-click searches explain far more.
Where each number comes from, and how late
Before choosing the charts, it is worth writing this table and keeping it next to the dashboard. It is the part almost nobody documents and the first one you need when a number doesn't add up.
| Source | What it really measures | Stated delay | History available | Limit to remember |
|---|---|---|---|---|
| Search Console | Clicks, impressions, average position on Google's SERP | Up to 48 hours, the last few days are preliminary | 16 months | Rare queries anonymised, dates in Pacific Time |
| Analytics 4 | Sessions and conversions from people who have already reached the site | From a few hours to 48 hours | Up to 14 months for events, depending on the setting chosen | Depends on cookie consent: people who decline aren't counted |
| Rank tracker | Position recorded by a crawler in a set location | Daily or on-demand updates | From the date you added the keyword | Estimated traffic is a calculation, not a measurement |
| Technical crawler | State of the pages at the time of the crawl | Only when you run it | None, except exports | It captures the site, not what Google does |
The 48 hours and 16 months aren't our estimates: Google states them in the documentation for the Search Console and Analytics 4 integration, where it says Search Console data becomes available in both tools 48 hours after collection and that Search Console keeps a maximum of 16 months, which is why Analytics reports stop there too.
Those 16 months have a practical consequence that we often see underestimated: a year-on-year comparison works for four months a year and then the starting data disappears. If you want to keep history beyond that limit you have to export it yourself, with the API or a scheduled save to Google Sheets or BigQuery. It isn't a job for a senior SEO, it's half a day once, and it's the difference between having a time series and starting again every sixteen months.
Four Search Console limits that quietly break a dashboard
Search Console is the only source that measures rather than estimates, which is why it sits at the centre of any serious dashboard. But it has precise boundaries, written in the official documentation, that no interface reminds you of while you look at a chart.
The number of rows isn't the same everywhere
The Performance report interface shows 1,000 rows. The Search Analytics API accepts a rowLimit parameter from 1 to 25,000, with a default of 1,000, and Google adds a sentence worth reading twice: the API is "not guaranteed to return all data rows but rather top ones". A dashboard built on the interface and one built on the API for the same site give different totals, and both are correct within their scope.
The chart total and the table total may not match
Google explains it in the guide to the Performance report: data grouped by query, country, device and date is aggregated by property, while data by page and search appearance is aggregated by page. A search that shows two pages from the same site counts as one impression for the property and two for the pages. If you put the two totals side by side on a dashboard without labelling them, you have just created a number nobody will be able to explain.
The same applies in Data Studio's Search Console connector, where the choice between the Site Impression and URL Impression tables is exactly that distinction: Google's documentation presents them as the two aggregation methods, and only the URL Impression table exposes the Discover and Google News search types.
Rare queries aren't there
Searches made by very few people are anonymised and don't appear in the query report. On a site with a lot of long tail, the sum of clicks by query is systematically lower than the property total. It isn't a dashboard bug, it's data Google doesn't hand over, and it belongs in the footnote of the client report before the client asks about it.
The dates aren't in your time zone
The Search Analytics API works with dates in Pacific Time, that is UTC minus 7 or 8 hours. A UK dashboard that compares a Search Console day with an Analytics 4 day set to Europe/London is comparing two windows eight hours apart. You won't see it over a month; you will on the single day of a campaign.
The row missing from almost every dashboard since 31 August 2026
Search Console has a performance report for generative AI, available to every site worldwide since 31 August 2026. Google's official guide sets out its scope unambiguously: it covers impressions earned inside AI Overviews and AI Mode, excludes Search Labs experiments, and can be segmented by page, country, device and date.
Two things the report doesn't give you, and they change how the dashboard should be built: there is no query-level detail, and there is only one metric, impressions. No separate clicks, no position. Anyone expecting an "AI Mode traffic" column will be disappointed, and anyone promising one in a client report is promising a number Google doesn't publish.
It is still useful, because it is the only place where you can see visibility growing while clicks stay flat. On the scale of the effect, the best reference is still the Pew Research Center study published in July 2025, which tracked the real browsing of 900 US adults in March 2025 across 68,879 Google searches, 12,593 of them with an AI summary: when the summary appeared, users clicked a traditional result on 8% of visits, against 15% when it didn't, and ended the session in 26% of cases against 16%. It is a US sample, not a UK one, and should be cited as such: anyone presenting it as a UK figure is translating, not measuring. We have gathered the figures we can stand behind in the articles on SEO statistics and on the impact of AI Overviews.
What to put in it, and in what order
The rule we use: every panel must be able to change a decision. If nobody does anything differently after looking at it, it is just taking up space.
- Organic clicks and organic conversions, side by side, with the same time window written above them. It is the only pair that tells you whether the work is producing anything.
- Impressions against clicks on the same chart. When impressions rise and clicks stay flat the cause needs finding straight away, and it is usually one of three: an unappealing title, a featured snippet or AI summary answering in your place, or new searches where you appear at the bottom.
- The top twenty pages by clicks, with the change. Not every page: the top twenty, because those are the ones you work on.
- Impressions in AI Overviews and AI Mode, from the dedicated report, as a series of its own.
- A minimal technical block: indexed pages, excluded pages with the reason, crawl errors. The rest belongs in a periodic audit, not in a dashboard.
Off the first screen, on the other hand, go three metrics we see almost every time and that almost never help. Average position for the whole site, which mixes thousands of different searches and moves for reasons you didn't cause. Bounce rate as a quality indicator, because a page that answers well and quickly produces the same bounce as a page that misses. The number of ranking keywords, which grows even when only the irrelevant long tail grows. On what to keep in their place, we have written a guide to SEO KPIs.
Building it in practice
The free tool that covers 90% of cases is once again called Data Studio. On 10 April 2026 Google announced on the Google Cloud blog the return to the original name after the Looker Studio interlude, specifying that existing reports, data sources and users move to the new experience without you having to do anything. Many guides you will find still use the old name: it is the same product.
The minimum setup is three native, free connectors: Search Console for clicks, impressions and positions, Analytics 4 for conversions and behaviour, Google Sheets for everything you keep by hand, such as the dates of changes made to the site. That last sheet is the part missing from 99% of the dashboards we open: without a column saying "on 12 March we rewrote the titles in category X", nobody knows why a rising chart is rising.
Paid platforms such as AgencyAnalytics, Databox or Klipfolio solve the assembly work and white labelling for agencies, not the data limits described above, which stay identical because the source is the same. On how much the suites feeding these dashboards cost, we compared the price lists in the article on the best SEO software.
A dashboard isn't a report
It is a distinction written everywhere in the literature and ignored in most day-to-day practice. The dashboard is always on and answers a single question: is something happening right now. The report is periodic, has an author, and answers a different question: what did we do, what changed and what will we do next. The commentary is the part that counts, and it can't be automated. Which source answers which question, with what delay and with what typical mistake, we have put in a table in the guide to SEO analytics.
Confusing the two produces the most common mistake we come across: the 24-page PDF generated automatically and sent every month, which nobody reads and which gets replaced in the meeting by one sentence said out loud. If the dashboard is done well, the monthly report fits on one page: three numbers, three lines of explanation, the decisions taken. On how to judge the work of whoever does it, we have written how to tell whether your SEO agency is working and what to expect in terms of timescales.
How many UK businesses have the data to do it
Some context that helps show how rare this still is. According to the UK Business Data Survey 2026 from the Department for Science, Innovation and Technology, 86% of UK businesses handle some form of digitised data, but only 25% of those say they analyse it to generate new insights. Among large businesses the figure is 69%, among small businesses 35% and among micro businesses 26%.
Put simply: three in four businesses that hold data don't analyse it in a structured way. A dashboard with six panels that actually get read is worth more than twenty that nobody opens, and for a small business it is also the quickest thing to set up.
If you would rather someone else read those numbers, our SEO services include ongoing monitoring and a monthly report, and in our projects you will find the numbers we use to judge ourselves.
Frequently asked questions
It is a screen that brings together a site's organic search numbers in one place: clicks, impressions, positions, conversions and the technical state of the pages. The data comes from external sources such as Google Search Console, Google Analytics 4 and a rank tracker; the dashboard assembles and refreshes it, it doesn't produce it.
In most cases for three reasons documented by Google. Search Console data becomes available 48 hours after collection, so the last few days are preliminary. The interface shows 1,000 rows while the API returns up to 25,000 and states that it doesn't guarantee all rows. Data by query, country, device and date is aggregated by property, while data by page is aggregated by page, so the totals differ by design.
Organic clicks and organic conversions side by side over the same time window, impressions and clicks on the same chart, the top twenty pages by clicks with the change, impressions in AI Overviews and AI Mode from the dedicated Search Console report, and a minimal technical block with indexed pages and crawl errors. Site-wide average position, bounce rate and number of ranking keywords are better kept off the first screen.
Sixteen months, as Google states in the documentation for the Analytics 4 integration. Beyond that limit the data can no longer be viewed, so year-on-year comparison only works for part of the year. To keep a longer time series the data has to be exported with the Search Analytics API to Google Sheets or BigQuery.
Yes, since 31 August 2026 the performance reports for generative AI have been available globally. They contain impressions earned in AI Overviews and AI Mode, which can be segmented by page, country, device and date, and they exclude Search Labs experiments. They have no query-level detail and impressions are the only metric available.
Not for a single site. Data Studio, the product called Looker Studio until April 2026, connects Search Console, Analytics 4 and Google Sheets for free and covers most cases. Paid platforms help agencies with white labelling and managing many clients at once, but they don't get round the data limits, because the source stays the same.
One last thing the industry rarely says. The most useful dashboard we have built in the last two years had no charts at the top: it had a text field, updated by hand every Monday, saying what had been done on the site the week before. The charts sat underneath. Anyone opening the page read the cause first and then the effect, and stopped asking why a number had moved. No connector does that work for you, and it is worth more than half the integrations people pay for.
