Which ad creative actually drives sales? Reading a Meta ads report by sales, not cost per lead

“Which ad creative actually drives sales?” Ads Manager answers with cost per lead, and a lead is not yet a sale. Below, how to analyse Meta (Facebook and Instagram) creatives by sales: what to read in the table, what to pause, what to scale and how long to wait. The numbers come from a real report.

Why can’t Ads Manager show which ad sells?

In a lead-form campaign the last thing Meta sees is a submitted form. The call, the meeting and the payment happen later, in the CRM. In the driving school example (2–8 October 2026) Meta counted 863 leads in a week; that week’s 36 sales were closed in the CRM, and Meta learns of a sale only if the CRM tells it.

So sorting in Ads Manager answers one question: which ad collects forms most cheaply. Which one sells shows only when the CRM outcome sits next to it.

What happens if you judge a creative by cost per lead? 79 leads, 0 sales

One driving school, one week (2–8 October 2026), two creatives. The first brought leads at $0.87, below the account’s average cost per lead ($1.03). The second, at $1.28: 47% more. Lead volume was nearly equal: 79 and 77.

Product screen · capip.uz

On cost per lead the first creative wins. On cost per qualified lead too: $3.29 against $3.80. Judged by CPL or CPQL, the first would get more budget and the second would be paused as “expensive”. Only the sales row showed the difference: 0 and 5.

A cheap lead is not a bad lead; cost per lead simply says nothing about sales. The reasons are in the article on cheap leads with no sales.

How do you read a per-creative table?

Each creative is one row. Some columns come from the ad account, some from the CRM, and each answers a different question.

Read the table from the end: sales first, then qualified leads, cost per lead last. The usual habit is the reverse: sort by cost per lead and decide on the spot.

How much budget goes to creatives with no sales?

The first creative was not alone that week.

Read this number with care. Such a creative spent about $17 on average, less than one customer cost that week ($24.58): too early to call each of them bad.

But the sum is large: 44% of the budget brought no sale, and Ads Manager on its own does not show it: it counts leads, not sales. So sort by spend: check the creatives that took the most money first, and let the small ones collect leads.

How long should you wait before judging a creative?

A sale is a rare event, so small numbers mislead. That week one sale took about 24 leads on average (863 ÷ 36). A creative with 10 leads and no sale tells you nothing: at that rate it was due less than half a sale. One with 79 leads was due about 3 (79 ÷ 24). Zero where three were expected is rarely chance.

The first creative (79 leads, 0 sales) falls in the third row. The second’s 5 sales are a small number too: next week it could be 3 or 7. So the decision is “move the budget and check again next week”, not “delete”. With few leads, judge on 30 days, not one week.

What should you pause and what should you scale each week?

Once a week, in the same order. By hand that means exports from Ads Manager and the CRM joined in a spreadsheet; the CAPI Platform report has it in one table.

1. 1

   Open last week and the last 30 days

   Group by creative name: ads with the same name merge into one row. The week shows what changed; 30 days give numbers large enough to judge.

2. 2

   Sort by spend, not by cost per lead

   Start where the most money goes: a wrong call costs most there.

3. 3

   Compare each row with N

   Fewer leads than N: leave it alone. N to 3N: look at qualified leads and CPQL. Above 3N with no sale: a candidate to pause.

4. 4

   Move the budget

   Pause what does not sell, add to what does. In CAPI Platform an ad, ad set or campaign is paused and re-enabled from the table itself (owner or admin only; the Meta token needs ads_management).

5. 5

   Write the decision down and check it next week

   One line: what changed, when and why. A week later, check in the cohort view whether the paused creative’s earlier leads turned into sales; if they did, switch it back on.

What does a creative to pause look like?

Two kinds. One has enough leads and no sales, like the card below: 79 leads, 0 sales. The other spends money and brings no qualified lead at all: no need to wait, since a qualified lead is known long before a sale. The CAPI Platform report’s “Worth a look” block lists creatives that spent 5% or more of the budget without a single qualified lead.

Product screen · capip.uz

What does a creative to scale look like?

It has sales, and its cost per sale is below the account average: $19.74 for the second creative against $24.58. Raise the budget in steps, not all at once, and check again every week: as the budget grows, lead price and quality can both change. An “expensive” lead is no reason to pause a creative.

Product screen · capip.uz

What has to be in place before you can trust the table?

A per-creative table tells the truth only when three things hold.

• Stages are kept in the CRM. The sales team has agreed which stage means “qualified lead” and which means “sale”, and managers move leads through them on time.
• Each lead is tied to its ad. A lead from a Meta lead form is tied exactly. A website request needs a UTM tag with the ad ID in the ad’s link. Leads from Instagram Direct, WhatsApp, Telegram or phone calls are not tied to an ad automatically: the report shows them by channel.
• People are counted, not rows. One person who submits twice is not two leads. The CAPI Platform report merges repeat leads by phone.

Once you know which creative sells, tell Meta. Qualified leads and sales go back from the CRM as signals through the Conversions API; the article on optimising for qualified leads covers the rest.