There is a separate funnel between registration and FTD
When registrations are enough and the first deposit sags, the team often blames the traffic entirely. But at this stage, several new factors are already appearing: the availability of the user script, the clarity of the interface, technical errors, and the local context.
Where to start?
Compare registration → FTD by date, GEO, device, and source. If the drop appeared simultaneously in all sources, a common factor is more likely. If only in one, study this segment first.
| Slice | Possible cause |
|---|---|
| Just one GEO. | Local availability or mismatch of expectations |
| Only mobile. | interface or technical problem |
| Only a new source. | audience quality |
| All segments from the same date | alteration |
Why you can’t make a conclusion on one day
Some fluctuations are natural. If the volume is small, several users noticeably move the percentage. Look at absolute numbers and percentages.
Comparison with an advertising promise
Overly aggressive creativity can lead to people who easily register out of curiosity but have no real intention of moving on. A high click → registration in this case is not a victory.
Do not collect sensitive data on your own.
The partner does not need to try to track the payment details of users. To analyze enough aggregated events and technical statuses, which are provided by the system.
How to Discuss a Problem with a Manager
Prepare a specific period and segment: "KZ mobile, campaign-04, after August 12 registration → FTD decreased from X to Y." Such a request can be verified. The general phrase “deposits have stopped going” does not give an entry point.
For the full picture, use the approach from the article about Click → Registration → FTD.
Separate product from advertising
If registration → FTD is deteriorating across all campaigns that lead to the same user path, there’s no point in taking turns stopping each source. First, check the overall factor. Conversely, if only one SubID shows an anomaly, the cause is more likely to be in its audience.
Time to FTD as a diagnostic metric
See the time distribution between the registration and the first deposit. A sharp change in the median may suggest that the path has become longer or that the audience has changed. There is no need to track sensitive data – aggregated event time is enough.
Compare the same days of life
You can't judge yesterday's sign-ups as if they already had as much time on the FTD as the weekly cohort. Use the window: For example, the proportion of users who have made an FTD within 24 or 72 hours of signing up.
After correction
Check to see if not only the conversion but also the quality of the follow-up has recovered. Sometimes simplification of the step increases the amount of FTD, but worsens the confirmation. This is another example of why you need to look at the whole funnel.
See the distribution of refusals by time
If the user massively stops the path immediately after registration, this is one scenario. If many reach almost the end and then stop, friction on a particular step is more likely. Even without sensitive data, a sequence of aggregated events helps distinguish between causes.
Do not change the advertising message ahead of time.
If CTR and click → registration remain the same and the problem appears later, the new creative can only mask the original cause. First, confirm the part of the funnel, then change the corresponding element.
Separate inspection of the new GEO
When launching a country, do not transfer historical registration → FTD from another market to a financial plan as a fact. Use a conservative range before accumulating your own sample.
