GTM Sauce

Sago Mini · B2C · preschool app studio

Measure ads with Apple's anonymous data only and model lifetime value yourself

Works because 80% of purchases happen in 2 days; still a slow manual process

Mixedanalytics-attributionpaid-social

What they did

Kids apps can't show the ATT prompt, so Sago Mini runs a SKAdNetwork-only setup and only buys on networks that fully support it (Unity Ads, IronSource, AppLovin). Because ~80% of purchases happen within 48 hours of install, SKAN conversion events give usable early conversion data. The growth team cohorts by geo and platform, builds frequently updated LTV models in Tableau/Google Sheets, computes projected LTV over CPA, and adjusts campaigns on each network by hand. It cannot split paid vs organic retention.

What happened

Solid first-48-hour conversion data, but highly manual and coarse; guest says it's more 'yes' than 'no' that they're back to not knowing which ads work.
Stage
Scale (scaling)
27:19
“we're- we're going with a SKAdNetwork implementation only. So we don't have that ATT prompt, we don't have, you know, the full attribution, which means that the ad networks that we work with have to support a full SKAdNetwork implementation for- for iOS.”
Brennan Clark
30:56
“So we cohort those out, build LTV models that are that are updated pretty frequently, and then feed that back into what we're doing on the paid UA side of things.”
Brennan Clark
“be looking through a Tableau or a Google Sheets and figure out, "Okay, this is our c- LTV over CPA, projected LTV over CPA," and then make those changes on the platforms.”
Brennan Clark

Related topics

Get tactics like this every Monday

New tactics from the week's founder interviews, each linked to where it was said.

More from this episode