A profitable account can contain many losing ads. In the large beauty account discussed below, the median new prospecting ad lost money while a small group earned further scaling. That makes hit rate and total account return different questions, both worth tracking.

The useful review is how much the unsuccessful tests cost, what they established and whether the successful ads produced enough contribution to cover that testing.

What the beauty account showed

The largest account I have creative-level data on is a large beauty account, and several people write ads for it, so this is not one person's taste. On that account, the median brand-new prospecting ad lost money, and only a small minority of everything tested was ever promoted into the scaling campaigns.

The account performed well overall despite that result for the median new ad. Similar patterns appeared in a laptop-sleeve account and a tanning pilot. These observations do not establish a universal hit rate or mean losses are what caused the profitable result.

Some of my own ads on the beauty account received substantial spend and still returned below breakeven. Delivery alone did not establish that the opening or any other part of the ad had worked.

Define what counts as a hit

A hit rate needs a denominator and a rule. Are you counting every uploaded ad, only eligible ads, or only those that received enough delivery? Does a hit mean a first profitable purchase, a sustained return or promotion into a scaling campaign?

Use the same definition across batches and record the spend required to produce each finding. Otherwise changes in the reported hit rate may come from the counting rule.

Separate untested ads from measured losses

CBO allocates budget across ad sets, and delivery within them can also be uneven. An ad with zero spend has not produced conversion evidence. It may be reasonable to replace it, but that is different from concluding it cannot convert.

Read concentration alongside campaign output and each ad's evidence. A few ads receiving most of the budget can be compatible with good results, without settling the unanswered questions about the others.

Check delivery eligibility, the reporting window and conversion lag before judging a new ad. A fixed 24- or 48-hour wait does not guarantee that it has enough data.

Decide whether an unresolved concept deserves a separate test or whether a better candidate should take its place. The CBO delivery guide explains those choices.

Review the batch by evidence and cost

Keep the review on a regular schedule, but allow an inconclusive result to remain inconclusive. The CBO batch-size guide connects the number of candidates to budget capacity.

When a losing batch is actually telling you something

An ad with substantial spend and poor return gives you a loss to investigate. Read the funnel and creative together before deciding which part needs to change.

In my beauty-account example, the ads received budget but fell below breakeven. That warranted checking the opening, body, offer and landing experience; spend alone could not isolate the failed part.

A food-brand audit raised a different question: many ads repeated a similar claim, and engagement metrics did not usefully separate returns. Reviewing the arguments supported testing different reasons to buy, rather than simply producing more executions of the same claim.

Keep findings scoped to the account and test. A losing ad can reflect creative, audience, offer, delivery or site conditions. A winner from a laptop-sleeve account provides a hypothesis for another category, not a guaranteed template.

A previously strong ad still needs to meet the current economics. Investigate a decline and apply the agreed pause criteria, even if the replacement is not ready.

Keep a record the next brief can use

For each ad, record its purpose, delivery, purchases, cost and the decision taken. The creative testing framework has the log I use for this, with a real entry filled in.

Use those notes when briefing the next batch. Preserve unresolved questions so a later team member does not mistake an underfunded idea for a disproven one.

My creative strategy work for DTC brands includes research, briefs, review, testing analysis, and iteration. See the creative work and current service details.