Why this order, not another
The order matters because the second number always reframes the first, and the third always reframes the second. If you put pacing first, you lose the denominator. If you put conversions first, you lose the context. We tried six orderings. This is the one that survived.
1. Active accounts — the denominator
Every other number on this page is meaningless without knowing how big the book is. Forty active accounts spending $284k is a healthy month. Twelve active accounts spending $284k is a different story and a different conversation. The denominator goes first because every interpretation flows from it.
The number also tells you something the other numbers cannot: did the book change since last week? A new account onboarded. An account churned. An account paused because the client's brick-and-mortar is closed for renovation. That single number prompts the right first question — "anything new in the book?" — before you start arguing about CPA.
2. MTD spend — the volume
Month-to-date spend tells you, in one number, whether the engine is running. Not whether it is efficient, not whether it is converting. Whether it is running. If MTD spend is wildly off plan on day 18, you have a problem before you have anything else. Pacing math, learning periods, budget pauses, billing issues — all of them surface here first.
We resisted putting MTD spend in dollars instead of percent of plan for two reasons. First, dollars are concrete; a percentage averages too much. Second, when MTD spend is $284k and the plan is $290k, the strategist's brain can compute the gap faster than reading a "98%" label. Numbers feel like reality. Percentages feel like dashboards.
3. Pacing average — the trajectory
Pacing is the first derivative of spend. MTD answers "where are we?", pacing answers "where will we be if nothing changes?". The average across the portfolio is a single number, but the pill underneath it ("2 accounts off") is where the action is. If 37 accounts are pacing 96–104% and 2 are pacing 60%, the average is fine. The 2 outliers are not.
This is the only KPI in the top row that we annotate with a count. Counts matter when averages hide the tail.
4. Conversions 30d — the result
Last in the row, on purpose. Conversions are the outcome, not the lever. A strategist looking at the top row should already know the engine is running and the trajectory is right before they ask "and did it convert?". Putting conversions first invites a kind of victory-or-defeat thinking that is bad for portfolio decisions. Putting it last invites a healthier read: "engine is running on plan, here's what came out."
30 days, not MTD, because conversion windows and learning periods make MTD conversions a misleading number until at least day 14. A rolling 30-day window smooths the lag without smearing the signal.
What we kicked out
Removing things from the top row was harder than choosing what to keep. Here is what almost made the cut, and why it lost.
| Number | Where it went | Why not in the top row |
|---|---|---|
| Average CPA | Per-account table | A portfolio average across verticals is meaningless. Auto repair, schools, and SaaS do not share a CPA. |
| CTR | Per-campaign drill-down | Useful for ad iteration, not for portfolio health. |
| Quality Score | Hidden by default | Lagging indicator. Nothing you can do today changes the number on screen. |
| Impressions | Per-account table | Reach without intent is noise at the portfolio level. |
| Total cost | Replaced by MTD | "Total" across mixed billing periods is a number that confuses every reader of every dashboard. |
| ROAS | E-commerce dashboards only | Mixing lead-gen and e-commerce ROAS in one number is malpractice. |
Every one of these has a defender. Every one of them has its place in the dashboard, just not in the top row. The job of the top row is not to show every number — it is to be the first thing a senior strategist looks at, with the highest possible information density per glance.
The test we use
Whenever someone on the team wants to add a number to the top row, we run a small thought experiment we call the five-second rule: can a senior strategist, looking at the row for five seconds, decide which client to call first? If yes, the row earns its place. If the new number does not change the strategist's first call, it does not belong at the top.
The general lesson
A dashboard is a forcing function, not a museum. Every number that earns a top-row seat must change the next decision. If it does not, move it down.
One number we still argue about
Honest internal disclosure: about once a quarter, someone on the team pushes to replace Conversions 30d with Conversion value 30d for the e-commerce side of our portfolio. We have not. The reason is that the same dashboard runs across lead-gen and e-commerce accounts, and "value" without currency context and margin context creates a worse number than "count". For agencies that run only e-commerce, swap it. For mixed portfolios, count wins.
You can read the rest of the layout — state cards, health donut, tCPA buckets, briefcase, weekly actions — in our free portfolio dashboard post. The top row is the same as the one in this article. It is also the same row that ships at the top of every actcenter account.