The same reach can bring a very different cashier result

Two placements can reach the same number of accounts and finish with very different FTDs. Even equal CPM leaves that outcome open. Read past the audience columns to the regs, the players who deposited and the money they added. The ranking can change at each step: a placement that leads on registrations may fall behind once those players reach the cashier.

The audience number stops before the brand

Reach tells you how many distinct accounts were reached under the platform’s counting rules. Impressions tell you how many times something was shown. One account can contribute several impressions. A follower count answers another question again: how large the creator’s audience is. Calling all of these “reach” makes a tidy report that compares different things.

Check one more thing in the reach figure: how much of that audience belongs to the test geo? A large audience in the wrong geo does little for a brand whose product serves a different market. Audience size doesn’t say whether a person registered or completed a payment. Those events happen after the audience figure has already been counted.

You can have strong reach, plenty of interest and a weak cashier result. You can also have a smaller audience that produces more FTDs. The reach number does not choose between those stories. Regs begin to show what happened on the brand, and the first deposit takes the comparison further.

Keep the original name of the audience metric and its reporting period. That is enough for the background column. There is no reason to stretch it into a deposit forecast before the test has supplied a deposit result.

The tie ends at the cashier

Here is a hypothetical comparison of two placements in the same geo. Both use the same attribution rule, and each group of regs is followed for the same 30-day window. Every number is invented for the calculation. These are neither agency results nor targets for another test.

The audience columns give a tie. Regs put A ahead. FTDs and deposit total put B ahead. That is the whole problem with choosing from reach: you can announce a winner before reading the rows that describe the test’s goal.

“We reached the audience” leaves the cashier question unanswered.

CPM has no deposit in its formula

CPM is spend divided by impressions, multiplied by a thousand. It is a way to describe the exposure figure. Regs, a first deposit and repeat deposits are all outside that calculation. Two placements can therefore match on CPM while their players take very different paths after arriving at the brand.

That is why a lower CPM cannot settle a deposit comparison either. It may come with more impressions while the cashier result stays weak. Conversely, a higher CPM can sit beside a larger deposit total. Neither direction turns CPM into a shortcut for the brand’s deposit goal.

A CPM comparison and a deposit comparison follow separate parts of the journey. For the latter, find the players brought in by each placement and follow the same reporting window. The exposure metric gives no basis for assuming that equal shares of those people registered, opened the cashier or paid. Those steps have to come from the brand’s records.

When someone offers reach as the denominator for CPM, the calculation has already changed. It needs impressions. If the report only contains reach, leave CPM out until the impression count is available. Leave the denominator open instead of giving a guess the weight of a calculation.

Follow the result from regs to payments

Read these rows together: regs, FTDs, reg-to-FTD rate and deposit total. First-deposit and repeat-deposit sums show how the total was built. These rows tell different parts of the same story, so dropping one can change the apparent result.

In A, more people registered, but a smaller share reached FTD. In B, fewer registered and more reached FTD. The higher deposit total then adds another reason to examine B closely. Keep the player counts and the composition of the deposit total visible when describing the better result.

The example also gives an average first deposit of €2,000 ÷ 50 = €40 for A and €4,500 ÷ 90 = €50 for B. The glossary sets out the exact meaning of that average. It belongs beside the counts rather than taking their place.

The ranking has changed even though reach stayed equal. Use the deposit report structure to keep that result tied to the players and payments behind it.

A better result still leaves a question

B leads in this example. Why? The table does not say. It could reflect audience fit, the path through the brand or a problem in the reported events. A result and an explanation are different jobs. The table gives you the result; the next question should help explain it.

“Reach was good” tells us little while we still don’t know where the regs stopped. If the brand has many regs and few FTDs, the agency and brand need to check that gap together. If the reported deposit figures disagree, start with the tracking checks. A bigger audience number won’t settle either issue.

For that check, the brand compares geo and creator type while the agency handles individual creators.

A reply should identify the loss after registration. Did players reach the cashier, attempt a payment, or complete one that disappeared from the export? A finding about one of those steps explains something the reach figure leaves open. It also gives the teams a check to repeat when they think the problem is resolved.

The example supports B within this test and these windows. A new group may behave differently, even with the same creator type, so carry the result into the next test as a reason to investigate further.

Keep the comparison narrow enough to read

A test combines a geo with a creator type. Pooling several geos into one reach figure can hide which market brought the FTDs. Pooling the creator types can hide which part of the test needs another look. Read the separate rows first. A combined total may help later, once the geo and creator-type results are clear.

Show the dates and windows in the comparison header. The worked example uses equal windows. Apply the cohort window rules on both sides. The comparison header should tell you when each group entered and how long its players had to deposit.

Keep the deposit result attached to the geo and creator type that produced it. If either changes, say so in the next test’s question. The reader should be able to tell what was tested without working backwards from a large audience total.

The final choice must survive without the reach column

Cover the reach column and read the conclusion again. The deposit goal, the result and any unresolved problem should carry the argument. If the case for the next test disappears along with the audience number, return to the rows that show what players did on the brand.

A comparison can show which result is stronger without showing that either met the goal. That last check belongs in the decision too. Winning the table is an incomplete reason to scale.

In the example, B brought more FTDs, a higher reg-to-FTD rate and a larger deposit total from the same reach. Compare that result with the goal agreed for the test. Then check the cause of A’s weaker rate: the table shows the gap but leaves the player journey unexplained. That distinction matters when choosing what to repeat. Copying B’s audience size alone would copy the one row where both placements were equal.

The advertiser starting point is a test budget and a report in deposits. Reach gives that report context. The players who paid give the brand its result.