Budget Pacing in Google Ads: The Month-End Save

Google Ads runs on two clocks that do not agree. Google spends to an average daily budget, surging and easing day by day; you answer to a monthly number — a client's approved figure, a finance line, a plan. Budget pacing is the discipline of reconciling the clocks before month-end does it for you — and it is arithmetic a connected AI runs in seconds from one prompt: spend to date, run rate, projected close, and the reallocation that fixes the trajectory while there is still month left to steer.

Daily budgets versus monthly promises: why pacing goes wrong

The daily budget is not a daily limit — it is an average Google optimises around. The monthly promise, meanwhile, lives in an email or a contract and is exactly what it says. Between them sit the standard failure modes:

• The multiplication nobody did. Daily budgets across the account, summed and multiplied by the month's days, exceed the monthly plan on day one. Not drift — arithmetic that was never run.
• Mid-month raises with tails. A daily raise on the 18th feels small and compounds across every remaining day. Serial small raises are how accounts arrive at month-end genuinely surprised.
• The asymmetric miss. Overspend gets noticed because someone is charged; underspend passes silently — yet an account landing well under plan quietly missed the volume the budget was approved to buy. Pacing is about both edges.

The 2x daily overdelivery rule, explained honestly

Google may spend up to twice your average daily budget on any single day — deliberate overdelivery, designed to buy the good days hard. Two protections cap it: the month is limited to your average daily budget × 30.4 (the average number of days in a month), and spend beyond that limit is credited back. The mechanics are documented in Google's overdelivery article.

Honestly told, the rule is fine — and it carries two operational stings. Day-level spikes are normal, so a doubled Tuesday is not an emergency and reacting to single days is how pacing turns into whiplash. And the 30.4 cap tracks the budget as currently set: every mid-month raise raises the ceiling, so the cap protects you from Google's surges, never from your own edits. Budgets changed mid-month make the naive "daily × days" mental model wrong in both directions — which is exactly why projection beats intuition. The model keeps all of it straight without effort — caps, credits, mid-month edits — which is the quiet reason pacing prompts beat pacing spreadsheets.

The pacing prompt: spend-to-date, run rate, projected close

The arithmetic, defined once:

projected close = spend to date + (recent daily run rate x days left)
pace index      = spend to date / (monthly plan x fraction of month elapsed)

A pace index near 1.0 is on plan; materially above is trajectory to overshoot; below, an underspend that will surrender volume. Concretely: against a monthly plan of 60,000, the account has spent 31,200 by the morning of the 13th, with a seven-day run rate of 2,750 a day. Projected close: 31,200 plus 2,750 across the 18 remaining days — 80,700, roughly a third over plan. Pace index: 31,200 over (60,000 times 0.4) — about 1.3, the same verdict from the other direction. Caught on the 13th, the correction is a modest trim across the worst over-pacers; caught on the 27th, it is an apology.

The prompt that runs it across the account:

For account 123-456-7890 against a monthly plan of [amount] split
[per campaign or total]: month-to-date spend, run rate over the last
7 days, projected month-end close, and pace index — per campaign and
account-wide. Flag anything pacing materially over or under, note
which campaigns are limited by budget, and check whether any daily
budgets changed this month in change history.

The last two clauses earn their keep: "limited by budget" marks the campaigns that would use a surplus productively, and mid-month budget edits are the usual culprit when the projection disagrees with someone's memory of the plan. The AI returns the table in seconds; what used to be the spreadsheet is now the reading of it.

Reallocating between campaigns without whiplash

Projection without action is commentary. The reallocation discipline, in three rules:

1. Fund performance with the overshoot's surplus. Trim the over-pacer that converts worst; feed the budget-limited campaign that converts best. The pacing prompt already named both.
2. Move in small steps. Budgets are inputs to bidding systems; violent swings destabilise what Smart Bidding has learned. Two moderate moves a week apart beat one heroic correction — and leave room to be wrong cheaply.
3. Recheck on a date, not a feeling. Every reallocation gets a scheduled follow-up projection. The loop is: project, move, verify, and each pass tightens the estimate as the month shortens.

Underspend deserves its own protocol, because the late-month temptation is to dump surplus into whatever can absorb it — which buys the month's worst traffic at the month's highest urgency. Better options, in order: scale what is budget-limited and converting, pull a planned next-month test forward, or land under plan and say so with the numbers attached. An honest under is cheaper than a panicked even.

Applied through a connector, each budget change surfaces for one-click approval — the campaign and the exact new figure visible before anything moves, attributed to you in change history afterwards.

Shared budgets and when they help

A shared budget pools several campaigns under one figure and lets Google distribute it — which is pacing outsourced, inside the pool. It helps when campaigns genuinely share one monthly envelope and you care more about the envelope than the split: Google will shift daily headroom toward whatever is absorbing traffic, and the pool self-balances.

The costs are control and legibility: you can no longer guarantee any single campaign's spend floor, and per-campaign pacing signals blur into the pool. Keep separate budgets when campaigns answer to different stakeholders, serve different margins, or need clean per-campaign accountability. A workable hybrid: pool the flexible middle of the account, keep hero campaigns and fixed client commitments on their own budgets — and pace the pools as units in the prompt above. Review pool membership quarterly; campaigns change roles, and a hero campaign trapped in a pool is how a launch quietly underfunds itself.

Setting a mid-month checkpoint that actually happens

Pacing fails as a habit before it fails as arithmetic. Two anchors hold it: around the 10th, when a bad trajectory is cheapest to fix, and around the 20th, when the projection is solid and there is still month to act in. Put them in the calendar, and make the artefact the saved prompt — same definitions, same thresholds, every month, so the checkpoint costs five minutes. Tie each checkpoint to an owner and an escalation line — who trims, who approves, what size of correction needs a client's nod — so the projection always has somewhere to go. A pacing number that reaches nobody is trivia.

For a roster, the same discipline scales across clients in one pass — pacing flags are a standing section of the multi-account morning triage. Month-end saves are made in the middle of the month; the end is just where they get noticed.