What a marketing dashboard should actually show you

Most marketing dashboards are built backwards. Someone connects a licensed BI tool to Google Ads, Meta and GA4, drags in the metrics each platform makes easy to pull, and calls it reporting. What comes out looks professional and answers almost nothing, because the question it's built to answer is "what did each platform do," not "what actually happened."

The wrong starting point

Building from the APIs outward means the dashboard inherits every platform's blind spots. Spend, impressions, clicks and platform-attributed conversions are easy to pull and instantly available — so that's what gets shown. Whether those conversions qualified, whether they closed, and which channel actually gets credit when a sale is influenced by three of them, none of that lives in an ad platform's API. A dashboard built from what's easy to pull will always be missing the part that matters.

Start from the warehouse, not the report

The fix is sequencing, not software. Every source — ad platforms, tracking, CRM — lands in a warehouse first, gets reconciled into one model, and the dashboard reads from that model rather than querying each platform separately. Built this way, the dashboard is downstream of the truth instead of a repackaging of five separate claims.

This is also why a custom-built dashboard tends to hold up better than a licensed BI tool with your logo on it: a licensed tool is built to connect to platform APIs directly, which means it inherits the same blind spots. A dashboard built on your own warehouse can show whatever the warehouse actually knows, including things no single platform can see — and without the per-seat licence fees that make it expensive to give your whole team access.

What a useful one actually shows

Take an illustrative example: ad spend, leads reported by the ad platforms, leads qualified in the CRM, closed revenue, and blended ROAS across every channel — five numbers, but two of them (qualified, closed) can only come from the CRM, not the ad accounts. That gap between "leads reported" and "leads qualified" is usually the most useful number on the page, because it's the one no platform will ever show you unprompted.

The same principle applies to channel comparisons. Ranking channels by ad-platform leads and ranking them by closed revenue routinely produces two different orders. A dashboard that only shows the first ranking will have you scaling the wrong channel with confidence.

Build vs buy, properly framed

"Build vs buy" usually gets asked as a tooling question — which platform, what it costs per seat. The more useful version of the question is where the numbers come from. A bought tool that reads directly from ad platform APIs will always be constrained by what those APIs report. A dashboard built on a warehouse you own can show what actually happened, because it isn't limited to what each platform is willing to admit.