Here's the test that tells you whether your dashboard works: does anyone summarise it before it reaches leadership? If someone does, that person is the interface. The dashboard is an internal working document that happens to have an executive audience written on the label.
"Build me a dashboard" is rarely the request
Start by not building anything, because the stated request is almost never the actual need.
Underneath it there's usually one of three things: a specific decision someone has to make and doesn't feel equipped for, an anxiety about whether money is being wasted, or a meeting they have to walk into and defend something.
Those need different artefacts. A budget-allocation decision needs a comparison across channels with a recommendation. An anxiety needs a small number of stable indicators reviewed consistently. A meeting needs a page that can be read aloud.
The question to ask before building: "What decision will this change, and who makes it?" If nobody can answer, you're being asked for reassurance rather than information — which is a legitimate need, but it's served by something much simpler than a dashboard.
The diagnostic A dashboard nobody reads usually answers no question anyone was asking. It displays what was available, not what was wanted.
The audience problem
The most common structural failure: showing operational metrics to a strategic audience.
| Audience | Their question | What belongs |
|---|---|---|
| Leadership | Is this working? Where should money go? | Growth, efficiency, pipeline, risk — against target |
| Marketing lead | Which channels need attention? | Channel performance, cost per outcome, trends |
| Practitioners | What do I fix today? | CTR, bounce, deliverability, errors, spend pacing |
Click-through rate on a leadership dashboard is noise. It's a diagnostic for the person running the campaign, and putting it in front of an executive signals thoroughness while making the important numbers harder to find.
That trade is a net loss. Every additional metric reduces the attention available for the ones that matter, and leadership attention is the scarcest input in the whole exercise.
Which metrics belong at the top is a separate question with its own answer — the ground covered in setting goals and KPIs that drive growth. This article assumes you've settled that and are working out how to present it.
No number without a comparison
The single most improving design rule, and the most commonly broken.
Four thousand two hundred website visits is neither good nor bad. It's not information. It becomes information only beside something else: last month, the same period last year, the target, or a competitor.
So every metric gets at least one reference point. No exceptions. A dashboard of naked numbers asks the viewer to supply context from memory, which most won't do and some will do wrongly.
Which reference point you choose is itself an editorial decision that shapes interpretation:
- Versus target — answers "are we on track," best for accountability.
- Versus last period — answers "which way are we moving," but vulnerable to seasonality.
- Versus same period last year — handles seasonality, slow to reflect recent change.
- Versus a longer trend line — the honest one, since it shows whether a movement is unusual or normal variation.
That last option deserves more use than it gets. A single month-on-month arrow implies significance that a twelve-month line would frequently disprove — and showing the line prevents the conversation where a routine fluctuation becomes an emergency.
SurveyMonkey
Online survey platform for feedback, market research and forms with AI assistance and analytics.
Best for: Surveys, customer feedback & market research
Put the conclusion on the dashboard
The change that most directly addresses "will they actually read it."
Leadership doesn't want to perform analysis. If your dashboard requires interpretation, it will be delegated to someone who interprets it — and then the summary is the product and the dashboard is scaffolding.
The fix is to put the interpretation on the artefact itself. A short written line at the top: what happened, why, and what it means for the decision at hand. Two or three sentences.
This feels wrong to a lot of analysts, because it looks like editorialising over neutral data. But the alternative isn't neutrality — it's someone else supplying the interpretation, less well informed than you, out of your sight. Writing the conclusion is taking responsibility for it, not biasing it.
Match the cadence to the decision
Real-time is a feature, not a virtue.
If budget decisions happen quarterly and performance is reviewed monthly, a live-refreshing dashboard invites reaction to noise. And noise-chasing has a real cost — changes made in response to normal variation become the cause of the next problem, which is the same discipline that makes holdout testing worthwhile in any experimentation programme.
Set the refresh rate at roughly the rate decisions get made. Faster refresh mostly serves the operational team, which is a different audience with a different dashboard.
Be honest about attribution on the face of it
A specific credibility issue that quietly undermines otherwise good reporting.
Showing a single "marketing-sourced revenue" figure to two decimal places implies a precision that doesn't exist. Anyone senior who has looked closely at attribution knows this, and an overclaimed number damages trust in every other number beside it.
Better options: show influenced alongside sourced and explain the difference once. Show ranges where the underlying method warrants them. Or state the method's limitation in a single line beneath the figure.
None of that weakens your position. It's the difference between a number that survives scrutiny and one that collapses the first time someone asks how it was calculated — a distinction developed further in why attribution keeps getting harder and in the case for cross-channel measurement approaches.
The influence framing is also what makes hard-to-attribute activity reportable at all — events and webinars being the standard case, as set out in measuring webinar pipeline.
The uncomfortable question: should it be a dashboard?
Worth asking honestly, and no dashboard vendor will ask it for you.
Dashboards suit continuous monitoring by people who already hold the context — an operations team watching for anomalies, a media buyer pacing spend.
Periodic strategic review is a different activity. If the output is looked at monthly and needs interpretation, then a written page — conclusion first, two or three supporting charts, an explicit recommendation — will almost always outperform a live dashboard nobody logs into.
The honest position for many organisations: you need a monthly written summary, not a dashboard. The dashboard exists because it's easier to commission a tool than to commit to writing something every month.
If you build both, be clear about which is which. A live dashboard for the team, a written page for leadership, drawn from the same source. That's more work than one artefact and considerably more effective than one artefact serving neither audience.
Design decisions that matter
Assuming a dashboard is right, these carry disproportionate weight.
One screen, no scrolling. If it doesn't fit, you haven't decided what matters. The constraint is the feature.
Five to seven metrics maximum at the top level. Detail lives one click deeper for anyone who wants it, which is rarely who you think.
Consistent direction coding. Up should always mean good, or you should mark clearly where it doesn't. Cost per acquisition rising is bad; sessions rising is good. Mixing those on one screen without signalling it produces misreadings that get repeated in meetings.
Absolute numbers alongside percentages. A 300% increase on a base of two is three digits of drama about six events.
Show the data's age. A timestamp prevents the conversation where someone acts on a figure that stopped updating a fortnight ago.
Label anything estimated or modelled. Mixing measured and modelled figures without distinction is how credibility gets lost in one question.
The plumbing caveat
Briefly, because presentation can't rescue bad inputs.
A well-designed dashboard built on broken tracking is a faster route to a wrong decision than a spreadsheet would have been, because the polish confers unearned confidence. If conversions are double-counted, or a tag has been silently failing, the dashboard will present that clearly and convincingly.
So the sequence matters: get the measurement right first — the ground covered in setting up conversion tracking correctly — then present it. Presentation work on unreliable data is effort spent making an error more persuasive.
A build sequence
- Ask what decision this changes and who makes it. If there's no answer, stop and find the real need.
- Confirm the audience and build for exactly one. Separate artefacts for separate audiences beats one compromise.
- Choose five to seven metrics that answer that audience's question, and cut everything else however defensible it feels.
- Give every metric a reference point and decide deliberately which one.
- Write the three-line header — what changed, why, what should happen.
- Set cadence to the decision rate, not the data refresh rate.
- Label estimates, mark data age, code direction consistently.
- Ask the recipient after two cycles what they used it for. If the honest answer is nothing, change it or stop sending it.
Step eight is the one that gets skipped, and it's the only genuine feedback loop available. A dashboard sent for a year without anyone asking whether it's used is an organisational habit rather than a reporting practice.
For agencies, this conversation belongs at the start of the relationship rather than at the first reporting cycle — agreeing what will be reported and what decisions it should inform is part of onboarding a client properly, and it prevents the familiar pattern where reports get longer each month because nobody wants to remove anything.
If the constraint is that the data lives in six systems and assembling it manually consumes a day a month, that's an integration problem rather than a design one — and it's where a performance marketing partner already running consolidated reporting removes the work rather than adding a tool to it.
The short version
If someone summarises your dashboard before it reaches leadership, that person is the interface and the dashboard failed. Start by asking what decision it changes and who makes it — if nobody can answer, the real request is reassurance, which needs something simpler. Build for one audience, since operational metrics on a strategic dashboard crowd out the numbers that matter. Give every figure a reference point, because a bare number carries no meaning. Put the conclusion on the artefact in three lines: what changed, why, and what should happen. And ask honestly whether a written monthly page would serve better than a live dashboard, because for periodic strategic review it usually would — no tool vendor is going to tell you that.
Spending a day a month assembling numbers nobody acts on?
We run consolidated reporting that answers the questions your leadership is actually asking.
Explore Performance Marketing →Frequently asked questions
Why doesn't leadership read the marketing dashboard?
Usually because it does not answer a question they are actually asking, and because reading it requires analytical work they will delegate. A dashboard built from whatever metrics happened to be available shows activity rather than consequence, which leaves the viewer to work out what any of it means. When that happens, someone summarises it for them, and that person becomes the real interface while the dashboard becomes an internal working document that nobody senior opens.
What metrics should be on an executive marketing dashboard?
Only those that answer a question leadership genuinely has — typically whether the business is growing, whether marketing investment is working, where the next increment of money should go, and what is at risk. Click-through rates, impressions and bounce rates are operational diagnostics for the team running the work, not decision inputs at that level. Including them signals thoroughness while making the important numbers harder to find, which is a net loss.
Should a number ever appear on a dashboard without a comparison?
No, because a bare figure carries no meaning. Four thousand website visits is neither good nor bad until it sits beside last month, the same period last year, or the target. Every metric needs at least one reference point, and the choice of reference point is itself an editorial decision that shapes interpretation. Dashboards full of naked numbers force the viewer to supply context from memory, which most will not do and some will do incorrectly.
How often should a marketing dashboard update?
At roughly the rate decisions get made, not as fast as the data allows. Real-time refresh on a metric reviewed monthly invites reaction to noise, and noise-chasing is a genuine cost rather than a harmless quirk. If budget decisions happen quarterly and performance reviews monthly, a monthly cadence with quarterly trend context serves better than a live feed. Faster refresh mostly benefits the operational team, which is a different audience with different needs.
Is a dashboard always the right format for leadership reporting?
Frequently not. If the output is reviewed monthly and needs interpretation, a written page carrying the conclusion first, with two or three supporting charts, usually outperforms a live dashboard nobody logs into. Dashboards suit continuous monitoring by people who already hold the context; periodic strategic review is a writing problem rather than a visualisation problem. Tool vendors rarely make this point, because the answer does not involve buying a tool.
