Short answer: A daily Meta ads optimization is a short, ordered review of your account that ends in a list of tasks: what to pause, what to hold, what to promote and what budget to move. Start by checking what already changed since your last review, read every campaign on more than one data source, judge on a short recent window rather than one bad day, and write every call as an action you can do in Ads Manager.
Most people open Ads Manager, look at yesterday's numbers, panic about one red cell, and change something. That is not optimization. It is reacting. A daily review should be calm, repeatable and written down.
What is a daily Meta ads optimization?
A daily Meta ads optimization is a fixed routine that reviews every campaign in an account and turns what it finds into specific tasks, each with a reason and a next check date.
We run ours on the CTSG structure (Creative Testing, Scaling and Graveyard), so the review moves through those three campaigns in order. If you are not familiar with the structure, start with the CTSG system. The routine works on any account, though. The ideas are the same.
The output is a to do list, not a report. Every line should be something a media buyer can open Ads Manager and do: pause this ad, decrease this Ad Set, promote this ad, hold this one until a set date. If a line does not lead to an action, it does not belong.
Why do daily reviews go wrong?
Daily reviews go wrong for three reasons: people react to single days, they trust one data source, and they forget what they already changed.
Single days lie. Meta reporting is noisy. Sales show up late, attribution shifts, and one day can look terrible for reasons that have nothing to do with the ad. Pausing on one bad day kills good ads.
One source lies too. Meta's own reporting can count too many sales or too few. A third party attribution tool (we use one that tracks new customer orders by ad) can disagree, sometimes a lot. If you only look at one, you will make confident calls on bad numbers.
Memory lies most. In a busy account, it is easy to recommend a budget cut that someone already made two days ago. Then the cut happens twice, and you have no idea which one worked.
The routine below is built to fix all three.
What order should you check things in?
Check in a fixed order, every time. Ours looks like this:
- What changed since the last review. Pull the account's activity log (Meta's change history) before you recommend anything.
- Every campaign in the account. List them all, with spend. Do not trust your notes about which campaigns exist.
- Both data sources. Pull Meta and your third party attribution for the same dates, at the Ad Set and ad level, broken out by day.
- The big picture. Is the account as a whole at goal, under it, or drifting? This decides whether you are pulling back or pushing.
- Creative Testing. Every ad with spend gets a call.
- Scaling. Budgets, fading ads and whether promoted ads are getting any spend.
- Graveyard. Is it delivering, and what should move into it?
- Status check. Before any pause line, confirm the thing is actually still active.
Let's go through why each one matters.
Why start with what already changed?
Because you should never recommend a change that has already been made. The activity log tells you every budget move, pause and edit since your last review, and who made it.
This step does a second job too: it lets you grade past decisions. For each change, look at the numbers before and after. Did the cut help? Was the pause right? Was it too early to tell? Over time this becomes a score for your own judgment, which is the fastest way to get better. We wrote about this in scoring your own budget calls.
One pattern to watch: when a pause was the right call, people often cut the budget instead, because a cut feels safer. Track those. If the Ad Set is still failing after the cut, the next call is the pause, with a note that the cut was already tried.
Why list every campaign, not just the ones you know?
Because campaigns get renamed, duplicated and quietly launched. If you only check the campaigns on your list, you will miss the new one spending money, or you will declare a campaign dead when it was just renamed.
Any campaign spending money that is not on your list is a finding. Any campaign on your list with zero spend is also a finding. Both deserve a line.
Which data window should you judge on?
Judge on a short recent window, a few days long, not on one day and not on the last month. The short window shows what is happening now. The longer view is context, not a verdict.
Here is the reasoning. One day is too noisy. Thirty days hides change: an ad that was great three weeks ago and dead this week still looks fine on a monthly average. A short rolling window of recent days sits in the middle. It is recent enough to catch decline and long enough to smooth out a single bad day.
A few refinements:
- New ads are read from their first day of spend. If an ad started three days ago, a fixed window could cut off its first sales. Read all of its days.
- Lifetime numbers never save a failing ad. If an ad is failing now, a great first month does not change that. Print the longer read as context, then decide on the recent one.
- Older Ad Sets get a trend line too. For Ad Sets that have run for weeks, add a slightly longer read to show direction. It informs the call, but it does not overrule a failing recent window.
- Watch for one big order. If an ad's return rests on a single large order, it is not a real pass. Judge it on cost per customer and on the return without that order.
How do you use two data sources without getting confused?
Decide in advance which source wins, and name every case where they disagree.
We treat new customer numbers from our attribution tool as the deciding read for most calls, with Meta alongside. When the two agree, the call is easy. When they disagree, it gets a label. For example, Meta might show no sales while the other source shows real new customers. That is a sign Meta is under reporting, and the ad should not be paused on Meta's numbers alone. We cover that case in Meta under reporting.
The rule that keeps this honest: say which source set the verdict on every line. If you cannot reach one source, say so and mark the call as provisional.
How do you make calls in Creative Testing?
Every ad with spend gets a call, and the call depends on whether it has sold.
- No sales yet: judge on spend. Below a set amount, hold. Past it, pause.
- Sales: spend no longer decides. Judge on cost per new customer and return together, against your floor (the worst numbers you can accept).
Two habits make Creative Testing calls much better:
- Never pause an Ad Set on its average alone. If the Ad Set is failing but one ad inside is working, pause the failing ads and keep the winner. If the Ad Set is passing but one ad is failing, keep the Ad Set and pause that ad.
- Check before you promote. Before any promote call, look for the ad in every Scaling campaign and the Graveyard by its exact name. If it is already there, the question is whether it gets spend, not whether to promote it again.
The full set of judging rules is in how a creative testing ad gets judged.
How do you make calls in Scaling?
Scaling calls are mostly budget calls, made more slowly. At goal on both sources over the recent window: increase. Below floor: decrease. A bad day or two is not a reason to pause.
Three things to check in Scaling every day:
- Campaign and Ad Set health. Spend, purchases, cost per sale, return and average order value on both sources, against your goal.
- Fatigue in your biggest spenders. When a top ad slides over several days and its daily frequency climbs, it is wearing out. Reduce, recheck, reduce again.
- Are promoted ads getting spend? Compare each newly promoted ad's spend to the average ad in the campaign. If a winner from testing gets almost nothing, it is starving. That is a structural problem with a known fix (a runoff), covered in inside the Scaling Campaign.
Newly promoted ads are protected from daily pauses during their first couple of weeks. They only get cut early if they spend a clear multiple of the campaign's cost per sale with nothing to show for it.
How big should budget changes be?
Small and steady. Change budgets in modest steps, and never make several big moves in a row on the same Ad Set without waiting to see the result.
A few principles we follow:
- Move in steps, not leaps. Large jumps can reset how Meta delivers and make the result hard to read.
- Wait a few full days after a change before you judge it. Judge it too early and you will undo good moves.
- Make changes at a consistent time, ideally right at the start of the account's day, so each day's numbers reflect one budget.
- When you both raise and cut in the same review, note whether total account spend goes up, down or stays flat.
How should each task be written?
Every task should start with a clear verb, name the exact entity, and carry a short why and next.
Good task lines look like this in spirit:
- The verb names the level. "Pause ad," "pause Ad Set," "decrease Ad Set to a stated budget," "promote," "move to Graveyard," "hold." Never a bare "pause," because nobody knows if you meant the ad or the Ad Set.
- Full names, exactly as they appear in Ads Manager. Short names cause mistakes.
- Why: one plain sentence with the numbers against the floor and the one fact that settles it.
- Next: what happens at the next check, with a date. "If still below floor on that date, pause the Ad Set."
Number the lists so whoever acts on them can say "do 1, 3 and 4."
What about soft calls?
Flag them, every time. Holding a weak Ad Set "one more day" feels kind, but repeated holds are how accounts lose money quietly.
Our rule: whenever a call is a hold on something already well below goal, add one line that states the case for pausing, with the numbers. The decision maker sees both sides and chooses. We call this the prisoner of hope flag.
Key takeaways
- A daily optimization ends in a numbered task list, not a report.
- Start with the activity log so you never repeat a change that was already made.
- Judge on a short recent window, with longer reads as context only.
- Read two data sources and say which one set each verdict.
- Never pause an Ad Set on its average; look inside it first.
- Flag every soft call with the case for cutting, so hope does not run the account.
FAQ
How long should a daily Meta review take?Once the routine is set, most accounts take well under an hour. The time goes into reading, not clicking. If the review produces more than a handful of urgent actions every day, the structure probably needs work.
Should I make changes every day?Only when the data says so. Many days the right answer is "hold, check again on a set date." A daily review is about looking every day, not about changing things every day.
Why not just look at yesterday?Because one day is too noisy. Late sales, attribution delays and random swings make single days unreliable. A short window of recent days shows the real trend.
What if Meta and my attribution tool disagree?Decide ahead of time which one wins, and note every disagreement. If Meta shows no sales but your other source shows real new customers, do not pause on Meta's numbers alone.
Does this work for accounts with long buying cycles?Yes, with a slower cadence. Accounts where people take weeks to decide usually work better on a weekly review with longer windows. The order of checks and the way tasks are written stay the same.
If you want a second set of eyes on your account, we offer a free Meta account audit. It includes a full account audit plus net new reach and cumulative reach reports. Request one at our contact page.
Keep reading
More on the CTSG system:
What Is the CTSG System? Creative Testing, Scaling and Graveyard for Meta Ads
How Early Can You Judge a New Meta Ad? The First Verdict Explained
Why Do Some Meta Ads Start Selling Late? Late Bloomers Explained
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. What should a daily optimization end with?
2. What should you check first?
3. An Ad Set's average is below goal. What do you do before pausing it?
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 2 of 19
Next: Should Meta Creative Testing Run to Cold Audiences Only?
Previous: What Is the CTSG System? Creative Testing, Scaling and Graveyard for Meta Ads
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.