To build a marketing dashboard that gets used, start from the decisions you make every week and work backwards to the metrics that inform them. Five questions, one screen, no scrolling. Dashboards fail because they are built from available data instead of from needed decisions, which produces forty charts and zero actions.
A good dashboard is boring and short. You open it, you know whether things are fine, and if they are not fine you know where to look next. That is the entire specification.
Start With the Five Questions
Every useful marketing dashboard answers these, in this order:
- Are we getting more or fewer leads than last month?
- Where are those leads coming from?
- How many turned into customers?
- What did each customer cost to acquire?
- Is anything broken right now?
That is it. If a chart on your dashboard does not help answer one of those five, it belongs in an ad-hoc analysis, not on the screen you check weekly.
Write the five questions down before you open any dashboard tool. The temptation to add a metric because it is easy to pull is the single reason most dashboards become unreadable within a month.
The Metrics That Earn a Place
Map one or two metrics to each question and nothing more.
| Question | Primary metric | Supporting metric |
|---|---|---|
| Lead volume | Qualified leads this period | Trend vs previous period |
| Source | Leads by channel | Share of total per channel |
| Conversion | Customers won | Lead-to-customer rate |
| Efficiency | Cost per customer | Revenue per dollar spent |
| Health | Site conversion rate | Form or checkout error count |
Notice what is missing. No impressions, no reach, no follower counts, no email open rates, no bounce rate, no average session duration. Those are diagnostic metrics — genuinely useful when you are investigating why a number moved, and pure noise on a weekly dashboard.
Define Every Metric in Writing
Beside the dashboard, keep a one-page definitions sheet. What counts as a qualified lead. Whether cost per customer includes your retainer or only ad spend. Whether revenue is booked or collected. Which date a lead is counted against — created or closed.
Undefined metrics are the reason two people look at the same dashboard and reach opposite conclusions. The definitions sheet takes twenty minutes to write and saves the same argument every month.
Layout Rules That Make It Readable
Big numbers at the top. Four tiles across: leads, customers, cost per customer, revenue. Each tile shows the current period figure, the previous period figure beneath it, and the direction of change. A person should be able to read the top row in five seconds.
Trend beneath the numbers. Two line charts maximum — leads over time and cost per customer over time. Twelve weeks or twelve months of history, never both on one screen.
Breakdown last. One table of channels with spend, leads, customers and cost per customer. Sorted by spend descending so the biggest line of budget is the first thing you read.
One screen, no scrolling. If it does not fit, you have too much on it. Build a second page for diagnostics rather than extending the first.
Same place every time. Metrics do not move position between refreshes. Muscle memory is a feature.
Use Comparison, Not Absolutes
A number alone says nothing. Forty leads is good or bad only relative to last month, the same month last year, or your target. Every figure on the dashboard carries a comparison, and you pick one comparison basis and keep it.
Period-over-period works for most small businesses. If your business is seasonal, use year-over-year for the same period instead, because comparing December to November will mislead you every single year.
Build a Marketing Dashboard in Four Steps
Fix the sources first. A dashboard is a window onto your data, not a repair for it. If your conversion tracking double-counts or your CRM has three records per person, the dashboard will present that confidently and wrongly. Clean the underlying data before you visualize it.
Choose the simplest tool that connects to your sources. A spreadsheet pulling from two exports beats a sophisticated platform you never finish configuring. Upgrade when the manual refresh genuinely becomes the bottleneck, not before.
Build the top row only, then stop for a week. Live with four numbers for one week. You will discover that two of them answer most of your questions, and that will tell you what the next layer should be.
Add the breakdown and one trend. Then freeze it. Resist additions for a full quarter. If a question comes up repeatedly that the dashboard cannot answer, that is real evidence for a change. One-off curiosity is not.
The Review Habit Matters More Than the Build
A dashboard nobody opens on a schedule is a decoration. Put a fixed thirty minutes in the calendar weekly, same day, same time.
In that half hour, do three things. Read the top row and note anything outside its normal range. Pick the one number that moved most and spend ten minutes finding out why. Write a single sentence naming what you are changing this week as a result.
That last sentence is the output of the dashboard. Not the charts. The decision. Keep those sentences in a running log and after six months you will have a record of what you changed and what happened next, which is more valuable than any report.
Add Alerts for the Broken Things
Some problems do not wait for a weekly review. A form that stops submitting, a checkout that errors, tracking that goes silent — these cost money every hour they persist.
Set threshold alerts on the small number of things that signal a break: zero conversions in twenty-four hours, a conversion rate drop beyond a set percentage, a spike in form errors. Send them to email or a chat channel somebody reads. Keep the list short enough that an alert still gets attention, because alert fatigue makes the whole system worthless.
Mistakes That Kill Dashboards
- Building for everyone. A dashboard serving the owner, the marketer and the sales lead at once serves none of them. Build one per audience and keep each to its five questions.
- Vanity metrics on the main screen. Impressions and followers feel like progress and inform no decision.
- No target lines. A trend without a target is a shape. Add the number you are aiming at so the chart means something.
- Stale data with no timestamp. Every dashboard shows when it last refreshed. A number from nine days ago being treated as current produces bad calls.
- Charts where a number would do. A single figure with its comparison is clearer than a pie chart of the same figure. Use charts only for trends over time and for comparing items in a list.
- Adding instead of removing. Dashboards only get longer unless someone deliberately prunes them. Review for removal once a quarter.
Short Beats Comprehensive
The dashboard you will still be using in a year is the one that fits on a screen, answers five questions, uses metrics you have defined in writing, and ends in a weekly decision. Everything else is reporting theatre.
Build the top row, live with it, add one breakdown and one trend, then protect it from additions.
If you want an outside view on how your market and your category are shifting before you set this quarter’s targets, start with the Market Trends Report page on this site.