Short answer: Compare each Ad Set's performance in the days before a budget change to the days after it, judge both against your account's own goals, and label the change right or wrong. Leave out changes that are too recent to judge or that were overridden by another change before they could land. Review the batch on a regular schedule, and you will see clearly whether your raises and cuts are helping.
Most media buyers make dozens of budget changes a month and never check whether any of them worked. The day after a cut, results look a little better, and it feels right. A week later nobody remembers. The account drifts, and no one can say which calls helped and which hurt.
We built a simple review to answer one standing question, the kind Jake asks himself all the time: "I raised it, had a few rough days, pulled it back. Was I right?" The answer should come from the actual days, not from a feeling.
What is a budget decision review?
A budget decision review is a scheduled check that compares each budget change's before and after performance against your goals, and scores whether the call was right.
It looks at every budget change in a period, usually the last week or the month so far. For each one, it asks: what was happening before, what happened after, and given that, was this the right move?
We run it on creative testing Ad Sets by default. Changes to a Scaling Campaign's overall budget behave differently and are better judged at the campaign level, on their own.
How do you find every budget change that was made?
Pull them from Meta's activity log, which records each change with the date, the old and new value, and who made it.
The activity log, also called the account history, is the honest record. Do not try to reconstruct changes from budget snapshots, because the current budget in a report is just today's number, not the history. If the activity log is not available for an account, say so and skip the review rather than guessing.
Pull a few extra days before the start of your window, so the first change has a "before" period to compare against.
How do you compare before and after a budget change?
Look at a short, equal window of full days before the change and the same number of full days after it. Leave out the day of the change itself.
A few rules keep this fair:
- Skip the change day. On the day a budget changes, part of the spend ran at the old budget and part at the new one. It is a mix and tells you nothing clean.
- Stop at the next change. If the same Ad Set was changed again, the "after" window ends there.
- Add up the totals, do not average the daily ratios. Sum the spend, sales and revenue across the window, then calculate cost per sale and return on ad spend from the totals. Averaging daily ratios gives quiet days the same weight as big ones.
- Match the window to your buying cycle. For impulse products, a few days is enough. For considered purchases, where people think about it for a week or more, use a longer window. A short window on a slow product will make every working ad look dead.
What counts as a right or wrong budget raise?
A raise is right if the Ad Set held your goal at the higher spend. It is wrong if it missed goal and got clearly worse than before.
There is also a middle case worth naming.
- At goal. What happened after: It held your goal at the higher spend. Verdict: Right
- Missed goal, but did not get much worse. What happened after: The raise did no harm, but it scaled something that was never at goal. Verdict: Held, but not at goal
- Missed goal and got clearly worse. What happened after: The raise broke it. Verdict: Wrong
"At goal" means judged on cost per sale and return on ad spend together, not cost alone. An Ad Set with a slightly high cost per sale but a strong return can still be at goal.
A raise made with no history, where the outcome was poor, is also wrong. It was raised before it had earned anything.
What counts as a right or wrong budget cut?
A cut is right if the Ad Set was below goal before the cut. It is premature if it was at goal before, and the cut reacted to a blip.
The question with a cut is simple: were the rough days real?
There is one important twist. If a cut comes shortly after a raise, do not judge it on the days before the cut. Judge the raise instead. Look at the days between the raise and the cut.
- If the raise had failed, the cut is right. It reverted a failed raise. Note whether it went all the way back to the budget that worked, or only partway.
- If the raise was holding, the cut is premature.
Without this rule, a perfect revert can look like a mistake, because the days before the cut are really the days before the raise, when everything was fine.
Which budget changes should not be scored?
Changes that were overridden too quickly, changes that are too recent, and budgets set on an Ad Set's first day.
Overridden. If the same Ad Set was changed again within a day or two, the first change never landed. You cannot score it. Count these separately, because the count is a finding on its own. A lot of overrides usually means too many hands on the same Ad Sets, or too much reacting.
Too recent. A change needs a few full days after it before it can be judged. Give the count, and list only the big ones, so the reasoning is visible before the verdict is. Name the date they become judgeable.
Launch day resizes. A budget changed on an Ad Set's first day, before it spent, is just the budget being set. It is a setup choice, not a performance call.
Should you judge the decision or the outcome?
Both, separately. A good outcome does not mean it was a good call.
Picture an Ad Set running at a cost per sale far above goal. Someone raises its budget anyway. A few days later, by luck or by a new ad finding its feet, it comes good. The outcome was right. The call was not. It was a bet on something that had not earned it.
Say both in one line. Over time, this is how a team learns the difference between being right and being lucky.
How do you present the score?
As a short block: the overall score with counts, then one line per media buyer, then the changes grouped by verdict.
Here is the shape we use:
- Overall: how many judgeable changes were right, how many were wrong, as both counts and percentages, plus how many were too recent and how many were overridden.
- One line per buyer: the same numbers, so each person can see their own calls.
- Buckets: Wrong, then Held but not at goal, then Right, then Overridden, then Too recent, then Launch day resizes. One change per line: date, who, which Ad Set, old budget to new budget, before numbers to after numbers, and a one line reason.
Always print the counts next to every percentage. Two right out of two is one hundred percent, and it means almost nothing. The counts make a small sample obvious.
Is scoring budget calls a competition between buyers?
No. Never rank buyers or build a leaderboard. The point is to learn how the account responds, not who is best.
Attaching names to changes is useful because people learn from their own calls. But the moment it turns into a ranking, people start avoiding changes, or making only safe ones. That is worse for the account than any single wrong call.
What patterns does a budget review usually reveal?
In our experience, cuts tend to be right more often than raises, and most damage comes from raises that were too large or came too soon after another change.
That has led to a few standing rules we follow:
- Do not change the same Ad Set again before the first change can be judged. Otherwise you never learn whether the first call worked.
- Raise winners in modest steps. Big jumps are where at goal Ad Sets tend to break. If you want more, do it in two steps. See promoting a winner to Scaling for the other way to scale.
- When a raise fails, cut back to the last budget that worked, not partway.
What are the limits of this kind of review?
Recent sales are still attributing, the numbers are platform only unless you check a second source, and small budgets give thin samples.
State these every time:
- Attribution lag. Sales from the last several days are still being credited. Verdicts on recent changes will often improve. Mark them provisional and name the date to look again.
- Platform only. If the review uses Meta's numbers alone, some "wrong" verdicts may be Meta under reporting. Check before acting on a wrong verdict as a pause.
- Thin samples. Short windows on small budgets may only hold a handful of sales. The direction is useful. The precision is not. The big Ad Sets give the most reliable reads.
Any change that is still wrong and still in place should go straight onto the daily to do list. The review explains. The daily list acts. You can see how that fits in how to run a daily Meta optimization.
Key takeaways
- Score every budget change by comparing equal windows of full days before and after it, skipping the change day.
- Judge against the account's own goals, on cost per sale and return on ad spend together.
- A raise is right if it held goal at higher spend. A cut is right if the days before it were truly below goal.
- When a cut follows soon after a raise, judge the raise.
- Leave overridden, too recent and launch day changes out of the score, but always print their counts.
- Never turn the score into a ranking. It is for learning, not competing.
FAQ
How long should I wait before judging a Meta budget change?Long enough for sales to attribute, at least a few full days for impulse products and longer for considered purchases. Changing the same Ad Set again before then means you never learn whether the first change worked.
Why not just average the daily cost per sale?Averaging daily ratios gives a quiet day the same weight as a big one. Add up spend, sales and revenue across the whole window and calculate from the totals instead.
What if my account's activity log is not available?Say so and skip the review. Rebuilding changes from budget snapshots is unreliable, because reports usually show today's budget on every row, not the budget on each day.
Should I score changes to my Scaling Campaign the same way?Campaign level budget changes behave differently, because Meta decides how to split the budget inside. Judge those at the campaign level, on their own.
What is the most common mistake this review catches?Changing the same Ad Set again before the first change landed. It feels like active management, but it means none of the calls can be learned from.
Want to know which of your budget calls are working?
If your account changes budgets often but nobody checks the results, there is usually a pattern hiding in the history. Our free Meta account audit includes a full account audit plus net new reach and cumulative reach reports. Request your free audit here.
Keep reading
More on the CTSG system:
What Is the CTSG System? Creative Testing, Scaling and Graveyard for Meta Ads
What Is a Graveyard Campaign in Meta Ads, and Why Keep Losing Ads?
Does Meta Under Report Conversions? When Not to Pause an Ad That Looks Bad
How to Stop Holding Losing Meta Ads Too Long (The Prisoner of Hope Problem)
See every article on the CTSG system
Ready for your free Meta audit?
We will review your account and run net new reach and cumulative reach reports, free.
Check what you learned
Tap an answer to see if you got it right.
1. How do you score a budget change?
2. When is a raise the right call?
3. Should the score become a ranking?
Ready for the full test? Take the Media Buying exam and get certified. Want this run on your account? Get my free audit.
Module 1: Media Buying · article 14 of 19
Next: What Is Meta's Amplified Budget Strategy? Inflated Budgets Plus Bid Caps Explained
Previous: The Graveyard Report: What Your Dead Ads Can Teach You
About the author
Jake Mehani, founder of Creative Analytics
Jake has spent 15 years running Meta and paid social ads for consumer brands. Meta published his RBX Active Advantage+ case study, he taught the DMAnc Meta and TikTok ads masterclass, and he spoke at Motion's 2024 Creative Strategy Summit. His senior team runs media buying, creative and measurement like an in house team.