Meta ads audit checklist for small businesses

Before you start

• Pick a fair date range. The last 30 days is a good default; compare with the 30 days before.
• Write down the goal. Orders, leads or messages. Every check below is judged against it.
• Have enough history. In Karobar, import 12 months when you connect an ad account, so seasonal patterns are visible.

Check 1: Structure and the learning phase

Meta explains that a new or significantly edited ad set goes through a learning phase. An ad set is marked Learning limited when it is unlikely to get about 50 optimisation events in the week after the last significant edit.

• Count the ad sets that are Learning limited. Many small ad sets splitting a small budget is a common cause.
• Meta's suggestions include combining ad sets, widening the audience, raising the budget, or optimising for a more frequent event.
• Check how often ad sets are edited. Pausing, or changing the audience, creative or optimisation event, counts as a significant edit and restarts learning.

Check 2: Objective and results match the goal

Look at each campaign's objective next to what you actually want. A Traffic campaign reports clicks, so it can look cheap while bringing no orders. In Ad Insights, the Objective and Results columns sit side by side for every campaign, so mismatches stand out.

Check 3: Where the money goes

• Sort campaigns, then ad sets, then ads by amount spent. The top few usually take most of the budget.
• Compare cost per result (spend divided by results) across them. A big spender with a high cost per result is the first place to act.
• Look for spend on things that are off but were running in the range, and on ads with issues or in review.

Check 4: Frequency and tired creatives

Meta defines frequency as the average number of times each person saw your ad: impressions divided by reach. When frequency keeps rising while results fall, the same people are seeing the same ad too often.

• Check frequency for each ad set over the last 7 and 30 days.
• Look at click-through rate over time for the ads with the most spend; a steady drop is a sign of fatigue.
• Plan fresh creatives before the current ones wear out, not after.

Check 5: Which creatives actually win

The Creative Library in Karobar shows every image and video ad with its results and a label: Winner, Good, Declining or Underperformer. Each creative is judged against creatives serving the same campaign objective, on that objective's measure: return for sales, cost per lead for leads, CPM for awareness, and click-through rate when the objective is unknown.

• List the Winners and look for what they share: the first frame, the product shot, the text, the offer.
• Use Compare to put two creatives side by side, and Karobar AI Vision to explain why one beats the other.
• Move budget from Underperformers to Winners, and brief new creatives on what the Winners have in common.

More on this is on the creative performance page.

Check 6: Tracking you can trust

Meta calculates purchase ROAS as purchase value divided by amount spent, and notes that some values may be estimated. If purchases are not being recorded, ROAS is meaningless.

• Check that purchase value appears for sales campaigns at all.
• If you sell by cash on delivery through chat, Meta never sees the final sale. Compare ad spend with orders actually delivered, not just with Meta's numbers.
• Karobar's Social selling funnel report follows the path from ad spend and chats to leads, orders and delivered sales.

The checklist on one page

For a written version, build an Ad account audit report: it covers health, wasted spend and budget allocation, and can be shared with a link or emailed every month. See Build and schedule a business report.

What to do after the audit

1. Write down the three biggest problems, with the money each one costs per month.
2. Fix the largest first: usually budget on underperforming ads or ad sets stuck in learning.
3. Plan new creatives based on what your Winners share.
4. Leave the changes for at least a week before judging them; every significant edit sends an ad set back into learning.
5. Run the same checks next month and compare.

Mistakes to avoid

• Changing everything at once. Fix the biggest leak first and give it a week.
• Judging on a few days. Small budgets need time to show a pattern.
• Turning off winners to save money. Cut the losers first.
• Trusting clicks. Clicks are not orders; follow the money to delivered sales.