Building a Marketing Dashboard: Which Numbers Actually Belong on It
A practical breakdown of what belongs on the CEO's dashboard versus the media buyer's, and why most dashboards lie long before the tool gets picked.
Someone on your team spent three weeks pulling everything into a beautiful dashboard. First exec meeting, the CEO scrolls through 40 tiles and asks one question: is the paid channel actually working or not? Nobody can answer in under two minutes. That dashboard didn't fail because the charts were ugly - it failed because nobody decided who it was built for.
This is the mistake almost everyone makes the first time they set up a marketing dashboard: trying to build one view that answers every question for every person. It doesn't work. A founder deciding whether to double next quarter's ad budget and a media buyer deciding whether to kill an underperforming ad set need completely different windows into the same business.
Start by asking: who is actually looking at this
Before you touch a single chart, figure out which of two dashboards you're building. There's the exec version - looked at maybe once a week, used to make go/no-go calls. And there's the operator version - open all day, used to adjust bids, pause creatives, and reallocate budget in real time. Trying to serve both with one screen is how you end up with 40 tiles and no answer.
The exec dashboard: only numbers that change a decision
An exec dashboard should never have more than 8 metrics on it. If you're past that, you're not informing the decision - you're hiding from it. Every number on this view should pass one test: if this number moved 20% tomorrow, would we actually do something differently? If the answer is no, it's decoration.
- Customer acquisition cost (CAC) by month, trended, not just a single snapshot
- LTV : CAC ratio - the single number that tells you if the business math works
- Marketing efficiency ratio (MER) - total revenue over total marketing spend
- Pipeline or booked revenue value, not just leads generated
- Payback period - how many months before a customer's spend covers their acquisition cost
- Revenue by channel, as a trend line, not a pie chart for one month
If a number on the exec dashboard doesn't change what you'd do next week, it's decoration, not data.
The operator dashboard: built for the person running ads today
This is where granularity should be roughly ten times finer than the exec view. Your media buyer doesn't need "revenue by channel" - they need to know which ad set inside which campaign is bleeding money right now, and at which stage of the funnel prospects are dropping off.
- CPL and CPA broken down by ad set, not just by campaign
- Cost per marketing-qualified lead (MQL), and MQL-to-SQL conversion rate
- SQL-to-closed-won rate, so sales handoff quality is visible, not assumed
- Spend pacing against budget, daily, not just at month-end
- Creative fatigue signals - frequency climbing while CTR decays
If your team is building this internally and it's the first time, expect a real learning curve - most of the pain isn't the charts, it's getting funnel-stage data to sit next to ad spend data cleanly. That's normal. It's also exactly the gap a proper marketing dashboard setup service is meant to close, rather than just making the existing mess look nicer.
Where dashboards actually break - and it's rarely the tool
Here's the classic failure mode: Meta counts a lead as a conversion the moment the form is submitted. Your CRM only counts a customer once payment clears. Finance only recognizes revenue after the refund window closes. Three systems, three "truths," three different numbers on three different screens - and every team insists their number is the correct one.
That's not a tooling problem. You can rebuild the same broken dashboard in Notion, Looker Studio, or Power BI and it'll still lie to you, because the underlying definitions were never reconciled. Fixing this means sitting every data owner in one room and writing down, in plain language, what counts as a "new customer," a "conversion," and a "qualified lead" - then locking that definition into every report that touches it.
This is also the real question to ask anyone offering a marketing dashboard setup service: how do you handle conflicting definitions across data sources before you even open the dashboard tool? If the answer jumps straight to "we'll build it in Looker Studio," that's a vendor who's going to hand you a prettier version of the same lie.
Whatever platform you land on, what actually determines whether the dashboard holds up over time is boring: a clean, documented data structure; a fixed refresh cadence everyone can rely on; and version history so you can trace back when and why a number's definition changed.
A dashboard is a decision-making tool, not a report. If it's not making anyone's next decision faster or clearer, it's just a very well-designed spreadsheet nobody opens past week two.
Frequently asked questions
What numbers should actually be on a marketing dashboard?
It depends entirely on who's looking at it. An executive dashboard should hold no more than 8 numbers that would actually change a decision - CAC, LTV:CAC, MER, pipeline value, payback period, revenue by channel trend. An operator dashboard needs roughly ten times that granularity: CPL and CPA by ad set, stage-by-stage funnel conversion, spend pacing, and creative fatigue signals. Building one dashboard to serve both audiences is the most common reason dashboards feel useless in meetings.
Why does my marketing dashboard show different numbers than my CRM?
Because your data sources define the same word differently, not because a chart is broken. A common example: your ad platform counts a lead the moment a form is submitted, your CRM counts a customer only after payment clears, and finance recognizes revenue only after the refund window closes. Three legitimate definitions, three different totals. The fix is writing down one shared definition for terms like "new customer" and "conversion" and locking it into every report before you touch any dashboard tool.
Should I build my marketing dashboard in Notion or Looker Studio?
The platform matters far less than most people assume. Notion, Looker Studio, and Power BI can all produce a solid dashboard or a lying one - what actually determines reliability is a clean underlying data structure, a fixed refresh cadence, and version history you can trace back. Pick whichever tool your team will actually keep updated; consistency of use matters more than feature comparisons between platforms.
How much does a marketing dashboard setup service cost, and is it worth outsourcing?
A proper build - auditing data sources, reconciling definitions across platforms, and standing up both an exec and an operator view - typically runs into the low thousands of US dollars as a one-off project, or as a monthly retainer if you want someone maintaining it. It's worth outsourcing if nobody in-house owns data hygiene long-term, since an unmaintained dashboard tends to go stale within a quarter. Be wary of any vendor quoting a flat price without first asking how your systems define a conversion.