GTM Sauce

Sizzle AI · B2C · AI learning app

Buy test traffic with your best ad and run 3-4-way paywall tests in days

Cost per purchase fell from ~$500 to ~$300 within days

Workedexperimentationpaywallpaid-social

What they did

Take a winning (top-10%) creative and split paid traffic by goal. Most goes to 'crazy testing': send tens of thousands of installs in a few days into 3- and 4-way experiments on the paywall and other in-app steps, accepting up to ~$200 cost per purchase (vs $50-$100 profitability targets) as long as it isn't $1,000-$2,000. ~20% of traffic is held to profitability targets using churn (3-7 months) and LTV ($30-$70+ per purchase) by country and device. Keep the two data sets separate so a $300 CPA on test traffic doesn't lead to turning everything off. Also tried new audiences such as teachers and tutors.

What happened

Cost per purchase dropped from ~$500 to ~$300 within days of rapid in-app experiments (depending on targeting).
Stage
growth
17:37
“anytime I'm split testing part of the app usage, I kind of like don't care about the cost per purchase, because I'm sending in tens of thousands of installs to optimize that data within a few days.”
Yev Marusenko
18:19
“It's like within days, I do three-way experiment, four-way experiment. I'm testing like crazy things like in the paywall, other parts of the app, and right away that cost per purchase goes from $500 to $300 per, depending on what I'm targeting.”
Yev Marusenko
19:19
“I don't blend this data with the rest, because you might make a decision where like, "Uh, it's $300 cost per purchase, turn it all off."”
Yev Marusenko

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