Make Ugly Ads to Grow Your App – Yuliya Lennox (Solid Starts)

Treat steady CPMs/CPAs and a modest positive margin as a warning sign, not a reason to scale. Early creatives shouldn't be good; keep testing concepts until one creative produces a step change (CPA drops sharply, purchases triple in a day). Presented as what made BetterMe special.

Unpolished ads that explain the product and the pain in the first second rather than pretty brand spots; e.g. BetterMe's illustrated 'belly fat' ad showing fat hanging off the front. Avoid 15-second polished brand creatives (and AI assistants' advice to make them). Lead with the product/pain as in launching a new product category.

Ran an ad using the trending AI 'fruit babies' format (babies made of fruit) for an infant-feeding brand backed by doctors.

EWA's founder attended every marketing meeting and came up with creatives himself; after Yuliya left he took marketing back over. Advice: don't look for a CMO 'silver bullet'; founders must feel marketing hands-on.

EWA launched a funny, whimsical app in the founders' native language/home market, then grew into other languages. Similar pattern cited for Babbel and Blinkist in Germany. Advice: don't default to the US if you know another market better.

Test ~500 ads per week; run the tests in countries with similar conversion rates but low CPMs (e.g. Indonesia) at ~$10/day instead of ~$1,000/day at US CPMs, then scale winners in the US.

Made a big bet on Japan with a Japanese-speaking team member doing full localization and culturalization of the app.

Apps with huge organic (Replika, Solid Starts) eventually saturate their niche; organic stagnates and adding paid later largely cannibalizes it. Start performance marketing early: every ad also feeds organic (EWA's Instagram grew to ~7-10M partly from ad-driven followers).

Palta's apps regularly shared tests and results so others could steal and localize them. Recommends marketers share wins publicly, form a peer group with weekly calls, and even a 'marketer camp' to solve one problem together.

Someone added a blinking animation to the continue button.

Ladder signed actress/singer Hilary Duff for brand ads.

Solid Starts' team members each messaged the founder out of passion for baby-led weaning; Replika hired ML people before ML was popular.

Founders and core team worked ~17-hour days, officially six-day weeks, sometimes sleeping in the office.

With AI, translate the product into other languages in hours and test non-US markets that are culturally close (e.g. Turkey converting well); avoid markets you can't understand culturally unless it's a cheap add-on.

~60-question web quiz builds sunk cost, then offers 'your plan for a dollar' via Apple Pay; follow-up screens offer add-ons (PDFs etc.) that users click through with 'continue', ending in a ~$350-400 charge disclosed only in fine print; 30-day wait before refunds.

Team she joined was running three apps simultaneously, launching new ideas because the developer was already on salary.