Start free
Features › Google Ads Reporting & Insights › Google Ads budget pacing report

A Google Ads budget pacing report that projects the month and stages the fix — with a statistical gate on the optimizer.

Pacing is the report that keeps a client from a surprise invoice. It needs one number from you — the monthly target — and gives back where the month lands at the current rate, what can be adjusted, and by how much. This page covers the report, the two ways to stage a correction, and the statistics behind the optimizer.

Last verified: September 10, 2026 App version: 50.1 Canonical page for this topic: Google Ads Budget Alerts

Reading the pacing view

FigureHow it is computedWhat to do with it
Month-to-date spendSum of daily cost from the 1st to the last completed dayThe fact
Run rateMTD spend ÷ days elapsedThe pace
ProjectionMTD + run rate × days remainingCompare with target; the gap is the correction needed
TargetSet per account on the reportThe plan; also feeds AI Full Mode's pacing module — an account with no target reports "not checked"
Adjustable itemsCampaign budgets and manual CPC bids with current valuesWhat can move, and from where

Two ways to correct

Percentage adjustments are the blunt tool: the month is projecting 15% over, so cut selected budgets 15%. Tick the items, enter the percentage, stage. Each item becomes a Checkout row with its old and new value, and the confirmation sums the daily delta. The budget optimizer is the sharp tool: rather than cutting everything evenly, it moves budget from campaigns that convert worse than the account's CPA predicts to those that convert better, without changing the total.

The optimizer's statistics

For each eligible campaign — enabled, a base campaign rather than an experiment, on a standard budget — the optimizer computes expected conversions over the window as the campaign's cost divided by the account's CPA. It compares that with the campaign's actual conversions using a one-tailed Poisson test: for an under-performer, the probability of seeing that few or fewer; for an over-performer, that many or more. The p-value sets both the direction and the size of the move.

p-valueCap on decreaseCap on increaseReading
≤ 0.01to 0.20× currentto 5.00× currentStrongly significant — large moves allowed
≤ 0.03to 0.50×to 2.00×Significant
≤ 0.06to 0.75×to 1.50×Marginal
> 0.06to 0.80×to 1.20×Not significant — small moves, further shrunk toward no change
  • Non-significant results are additionally shrunk by a piecewise factor that reaches near zero above p = 0.5, so noise does not move budgets.
  • After individual changes are computed, the set is rebalanced so total daily budget is exactly unchanged — the optimizer redistributes, never inflates.
  • It needs at least five conversions of data in the window; below that it reports not checked. AI Full Mode's allocation module reuses these rules rather than a second copy.

A worked correction

Target $12,000. On the 18th, MTD is $8,100 and the run rate is $450 a day; projection $13,500, 12.5% over. The optimizer proposes: Campaign A, 61 conversions against 38 expected at the account's $52 CPA, p = 0.004 — budget up from $150 to $290; Campaign B, 9 against 31 expected, p < 0.001 — down from $200 to $60; Campaign C, 22 against 24, p = 0.41 — a shrunk move from $100 to $97 that rebalancing rounds away. Net daily change $0. That fixes allocation, not the over-pace, so a 12% percentage cut across the three follows as a second staged change. Checkout shows both sections and a −$54 daily delta; the reviewer pushes both, or the reallocation alone.

Shared budgets and Performance Max

Two campaigns on one shared budget appear in the adjustable-items table as one budget pool, labelled, and a change to it stages once for the pool rather than once per campaign — the reviewer should not see two rows that would double-apply. The optimizer's eligibility rules exclude experiment campaigns and non-standard budget types for the same reason: a change that Google would reject or apply differently should not look like a normal row. Performance Max campaign budgets are editable in place on the PMax overview and stage through the same path, so a pacing correction can be split between Search and PMax in one push.

Pacing and anomaly detection, together

Pacing asks whether the month is on plan; anomaly detection asks whether a day was unusual. A month can pace to 130% of target with every day looking normal — the run rate is simply too high — and a single spike can land in a month that is otherwise under-delivering. Both are needed, and they meet in two places: the budget history that anomaly detection's AI can request as evidence, so a spend alert can say the daily budget was raised on the 5th; and the change history, which shows the budget push from Checkout beside any budget edit made in Google Ads.

Setting a target that means something

The monthly target is a number you set per account, and everything on the page is relative to it. It should be the client's actual monthly commitment, not the sum of daily budgets times days — Google's own pacing can overspend a daily budget by up to twice on a given day and under-spend on others, so the daily-budget sum is a ceiling, not a plan. An account with no target renders the projection against nothing and reports not checked in AI Full Mode's pacing module rather than inventing a plan.