How to Build an SEO Dashboard That Answers the Right Questions
Most dashboards are built to display everything available rather than to answer anything, which is why nobody opens them twice. The useful discipline is deciding the questions first and letting those decide what appears.
Most Dashboards Answer Nothing
The usual dashboard is assembled from whatever the connected sources happened to offer. Everything available gets added, arranged neatly and scheduled. The result is a wall of charts. A wall of charts is not information.
How it comes about. Availability decides.
Somebody connects the data and adds what appears. Nothing in that process ever asks what anybody wanted to know, so nothing on the finished screen is there because it was needed.
What that produces. Comprehensive and useless.
Every figure is accurate. Nothing is missing. Nobody looking at it can say whether things are going well, which is the only reason it existed.
The tell. Nobody opens it.
A dashboard people stop looking at has already failed. That is the measurement of a dashboard and it is more reliable than any opinion about the design.
What this page covers. Not how to build one.
No configuration, no connections and no instructions. Which questions it should answer, which is the part that decides whether it works.
Decide The Questions First
Build it backwards. Write down three or four questions the business genuinely asks, then work out what would answer them. Anything that does not answer one of those questions comes off, no matter how interesting it is or how easily it was available.
Why backwards works. It excludes.
Starting from questions gives you a reason to leave things out. Starting from data gives you no reason at all, which is why the second approach produces everything.
How to find the questions. Ask the owner.
What do you want to know when you open this. The answers are usually shorter and plainer than anything already on the screen.
Why three or four. Anything more is a report.
A small number can be answered at a glance. Beyond that the reader has to search for what they wanted, which defeats the format.
The test for anything on it. Which question.
Point at any element and name the question it answers. If nobody can, it comes off. That test removes most of what is on a typical dashboard.
The Questions Worth Building For
Most businesses want the same four things from a dashboard. They map onto the three layers this series works through, with one addition. Yours may differ. Starting from these is faster than starting from nothing.
Are we visible for more of what matters. The first layer.
Not total appearances. Whether the range of commercially relevant searches you show up for is widening, which is the earliest genuine signal available.
Are the right people arriving. The second.
Visits, split by whether they searched your name and by which pages received them. An unsegmented total answers nothing.
Are they doing anything. The third.
Enquiries, calls or sales. The layer most dashboards omit entirely, because it needs the business to supply something.
Has anything broken. The addition.
A health check whose only job is to make a fault visible. It should be boring every month. The month it is not is why it exists.
One Screen
If it needs scrolling or tabs, it has stopped being a dashboard and become a report. That is not a stylistic preference. The two formats do different jobs and merging them produces something that does neither well.
What a dashboard is for. A glance.
Somebody opens it, spends a short moment and knows whether anything needs attention. Every design decision follows from that single requirement.
What a report is for. Explanation.
What changed, why and what happens next, with commentary. Our monthly report guide covers that document.
Why tabs are the warning sign. They hide the decision.
A second tab is what happens when somebody could not decide what to leave out. The constraint is doing useful work and the tab removes it.
What the limit forces. Genuine choices.
One screen means arguing about what earns its place. That argument is the most valuable part of building one. A scrolling dashboard never has it.
Sources And Why They Disagree
A dashboard combining search reporting with site analytics will show conflicting totals. That is not a fault in either of them and it is not something to fix. It needs labelling rather than reconciling.
Why they differ. Different events.
One counts a click leaving a results page and the other counts a page loading on your site. Between those two moments things get lost, so the totals were never going to match.
What the tools cover. Elsewhere.
Our Search Console and Google Analytics guides cover each platform. This page is about what belongs on the screen.
The mistake to avoid. Making them agree.
Adjusting figures until two sources match produces a number that describes neither. Hours disappear into it and nothing is gained.
What to do instead. Say where each came from.
A source label beside every figure. That single habit prevents somebody comparing two numbers that were never comparable.
Comparison Is The Whole Value
Every number on a dashboard needs something beside it. A previous period, a baseline or an expectation. A figure standing alone tells nobody anything. Most dashboards are full of figures standing alone.
What a bare number does. Nothing.
Visits this month, on its own, cannot be good or bad. The reader has no way to place it, so they either guess or stop looking.
What to compare against. Usually last year.
The same period a year earlier removes month length and seasonal position at once, which makes it the default worth building in.
The baseline version. Where you began.
For anything early in an engagement, comparison against the starting point matters more than against last month. Our benchmarking guide covers recording one.
The design consequence. Half the space.
Building comparison into every element roughly halves how many elements fit. That is a feature rather than a cost, since it forces the choices block four demands.
Who It Is For Changes What Goes On It
A dashboard for the person doing the work and one for the business owner paying for it are different documents. Attempting to serve both with one screen produces something that serves neither. It is the commonest reason a dashboard goes unread.
What the practitioner needs. Diagnostics.
Detail by page and by search, enough to find where a problem sits. That level of detail buries anybody who only wants to know how things are going.
What the owner needs. An answer.
Enquiries, direction and whether anything is wrong. Three things, in business language, without a search term appearing anywhere on the screen.
Why merging fails. Opposite requirements.
One needs granularity and the other needs summary. A compromise between those is a screen that is too detailed to scan and too shallow to diagnose.
The practical answer. Build both.
They draw on the same data and present it entirely differently, which costs far less than it sounds and produces two things people actually use.
What Not To Put On It
Four things appear on dashboards constantly and none of them earns a place. All four share a property, which is that they move reliably upward and therefore look like progress without describing any.
Third party authority scores. Not a Google measure.
An estimate produced by a company that sells software, describing your site rather than your business. Our guide covers why.
Ranking counts with no context. A number of positions.
How many searches you appear for somewhere, with nothing about whether any of them matter. It rises as coverage grows regardless of quality.
Anything cumulative. The one that only goes up.
Any running total will rise whatever happens, which makes it useless as a measure of anything and reassuring to look at.
Total traffic, unsegmented. The respectable one.
It looks like a real measure and hides everything, which our metrics guide covers in full.
Automation Does Not Make It True
An automated dashboard reproduces whatever was configured, faithfully, every month. Including the mistakes. Automation removes the effort of assembling something and it adds nothing whatsoever to its accuracy.
What goes wrong quietly. A filter or a date range.
Something set once, forgotten, then applied to every figure since. The output looks identical to a correct one, which is precisely the problem.
Why nobody catches it. It looks the same.
A wrong figure produced automatically has the same appearance as a right one. There is no visual signal that anything is amiss.
What compounds it. Nobody checks.
The point of automation was to stop doing it manually, so nobody returns to the sources. A misconfiguration can run for a year.
The habit that fixes it. Occasional verification.
Periodically check a figure against the source it came from. Not often. Often enough to catch something that has drifted.
Reporting Is Not Fixing
A dashboard is the most seductive version of this problem. It is visible, it updates itself and it demonstrates that somebody is paying attention. None of that changes anything about the website.
Why a dashboard is the worst offender. It runs itself.
A report at least requires somebody to write it. A dashboard continues producing evidence of diligence with no human involvement at all.
What it can hide. Months of nothing.
Figures updating daily while no page has been touched since spring. The screen gives no indication either way.
The question it cannot answer. What did you do.
No dashboard shows what work was carried out. That belongs in a report. Its absence is the clearest signal that measurement has replaced working.
Where the position sits. The pillar.
Our measurement guide holds it. Everything sits on the performance guide.
Point at anything
on it and name
the question it
answers.
If nobody can, it comes off. That test removes most of what sits on a typical dashboard, because starting from the available data gives you no reason to leave anything out. Starting from three or four questions the business actually asks gives you a reason to exclude everything else, which is the only way one screen ever stays useful.
How we build one:
No dashboard shows what work was carried out. Its absence is the clearest signal that measurement has replaced the work itself.
Every guide.
One question.
What each number actually means, how to read the data without being misled, what belongs in a report and what does not, then how to connect any of it to the enquiries the business is actually paying for.
