Pulsar
field guide · reading ads manager
Present: ← → Space · Esc back to Doc

Pulsar · Field guide · No. 19 · Why “Results” isn’t sales, why the attribution window changes everything, and the column set that tells you the truth.

The Columns That Matter (and Lie)

Ads Manager shows dozens of columns by default and most of them mislead. Here’s the handful that tell the truth — and how to build a view you can trust.

01

“Results” is not your sales

The default Results column counts Meta-attributed conversions — windowed, and partly modeled where tracking is blocked. It is not the count of Purchase events your pixel actually fired, and it’s not your store’s confirmed orders.

Meta’s “Results” answers “what can Meta claim?” — not “what did the business actually make?”

Both numbers have a job (see No. 11). Just never mistake the optimization number for the truth number.

02

Always show the attribution window

A conversion count with no window attached is meaningless. Add the Attribution setting column so every row says how it was counted — 7-day click / 1-day view by default now.

i
Comparing a 7-day-click number to an old 28-day one is comparing two different questions. Lock the window before you compare anything.
03

Never read a cost column alone

A cheap CPM can mean fresh reach (good) or high frequency on a burned-out audience (a trap). Read the trio together: CPM + CTR + frequency over the same window (this is No. 01).

ColumnReads as truth only when…
Cost per resultyou know the window and compare it to your real CPA
CPMyou read it next to frequency (cheap + rising frequency = saturation)
CTRpaired with frequency — falling CTR + rising frequency = fatigue
04

Build a column set you trust

Ditch the default view. Save a custom column set with only what earns its place:

Spend · Impressions · Frequency · CTR · CPM · Results · Cost/Result · Attribution setting · Purchases (pixel) · Purchase value

Put Attribution setting right next to Results, and your pixel Purchases next to Meta’s Results — the gap between them is the most useful thing on the screen.

05

The breakdown traps

Breakdowns (age, placement, region) are useful — until they aren’t. Two rules keep them honest.

  • ×
    Over-slicing. Cut the data enough ways and every cell is a tiny, noisy number that “proves” whatever you hoped.
  • ×
    Modeled conversions don’t always break down. Some conversion rows can’t be split by breakdown — the totals won’t reconcile, and that’s expected, not a bug.
06

The one-screen weekly read — and reconcile monthly

You don’t need to stare at it daily. Once a week, on your trusted view, scan four things:

Frequency — creeping past ~2.5? Fatigue is coming.
CTR trend — drifting down? Diagnose fatigue vs saturation.
Cost/Result vs your target CPA — and vs your real CPA.
Results vs pixel Purchases — note the gap.

Then once a month, reconcile Ads Manager against your store’s confirmed orders. A stable gap is fine; a widening gap is a flag. Optimize with Meta’s columns; decide with your own numbers.

1 / 8