Sub Club by RevenueCat · 2026-07-22

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

App marketer Yuliya Lennox (BetterMe, EWA, Palta, Replika, Higgsfield, Solid Starts) shares lessons from a decade of app growth: distrust stable ad performance and hunt for breakout creatives; make ugly, product-first ads (BetterMe's belly-fat ad) unless the brand truly matters (Solid Starts fruit-baby backlash); validate demand before building; founders must do marketing hands-on; localize and test creatives in cheap geos; pair organic with paid because organic plateaus; share wins across teams; hire obsessed people; and black-hat web quiz funnels hurt the industry.

Play the episode

16 tactics

BetterMe · B2C · fitness and weight-loss app

Keep testing new ads until one sharply cuts costs, even if current ads work

“Has there ever been a moment when you thought, wow, this is so much better than everything that I've done before where your CPAs went down, doubled down, when your amount of purchases tripled in one day.”

Unknownpaid-socialexperimentation

Sub Club by RevenueCat · 2026-07-22

BetterMe · B2C · fitness and weight-loss app

Run ugly ads that show the product and the pain in the first second

One ran for years; called 'the best ad ever'; no numbers

“There is this famous saying like make ugly ads. I think we invented that and I believe in that hundred percent because people don't want the TV shots at any point.”

Workedpaid-social

Sub Club by RevenueCat · 2026-07-22

Solid Starts · B2C · baby feeding app

Run a trendy AI 'fruit babies' ad for a doctor-backed baby feeding brand

Backlash: parents questioned whether to trust the brand

“Fruit babies are super popular on Instagram. I should definitely do an ad like that. And then we saw some feedback of, should we trust a company that invented this to feed our babies?”

Failedpaid-social

Sub Club by RevenueCat · 2026-07-22

EWA: Learn English · B2C · English learning app

Have the founder join every marketing meeting and write the ads himself

App reached #2 after Duolingo; link to this is loose

“And the founder, he actually was super into marketing. He was on every marketing meeting. He was doing it all himself. He was thinking of creatives himself.”

Workedteam-processpaid-social

Sub Club by RevenueCat · 2026-07-22

EWA: Learn English · B2C · English learning app

Win your home market first, then add languages one by one

Liked locally, then expanded; later #2 after Duolingo

“So if you have a benefit of knowing some local market better, definitely use it because this is what Eva did.”

Workedmarket-selectionlocalization

Sub Club by RevenueCat · 2026-07-22

A company testing 500 ads weekly · guest's last employer

Test new ads at $10/day in Indonesia, not $1,000/day in the US

Same test signal at ~1% of the cost

“You need to test 500 ads per week. And if you do that on US CPMs, you will definitely go broke. But if you do that with countries that have the similar conversion rates like, I don't know, Indonesia, and you will see the same tests for $10 a day instead of $1,000 per day”

Workedpaid-socialexperimentation

Sub Club by RevenueCat · 2026-07-22

Replika · B2C · AI companion chatbot

Fully translate and adapt an app for Japan with a Japanese speaker

Didn't work at all; harder than expected

“I can say that in Replika we did make a big bet on Japan. We even have person that knew Japanese in the team and we did the actual localization, the culturalization of the app didn't work in any way because it is harder than we think.”

Failedlocalizationmarket-selection

Sub Club by RevenueCat · 2026-07-22

EWA: Learn English · B2C · English learning app

Run paid ads from the start alongside organic growth

Organic-only apps plateau; paid added late often ate into organic

“And you will see that your organic stagnates.”

Mixedpaid-socialorganic-social

Sub Club by RevenueCat · 2026-07-22

Palta · app portfolio company

Share test results across apps and with a weekly peer group

“So you would have a lot of companies coming together and just sharing like today we did this test, this test performed this way. You can steal this test and localize it and put it in your app and see how it works.”

Unknownexperimentationteam-process

Sub Club by RevenueCat · 2026-07-22

Unnamed company

Make the paywall's continue button blink

Raised conversion, secondhand; no numbers

“someone somewhere added a blink on the continue button. You know the... And turned out it increases your conversion”

Workedpaywall

Sub Club by RevenueCat · 2026-07-22

Ladder · strength training app

Sign a celebrity like Hilary Duff for brand ads

Big influx almost at once, secondhand; ROI hard to measure

“they took a huge bet signing Hillary Duff as a influencer for the app, and it has already paid off. And he shared in the panel that almost immediately they saw a pretty big influx.”

Workedinfluencer

Sub Club by RevenueCat · 2026-07-22

Solid Starts · B2C · baby feeding app

Hire people obsessed with the problem

Credited for product quality; no numbers

“it's crazy to think that each one of them messaged the founder being like, "I'm so passionate about that, that I want to be on your team no matter what."”

Workedteam-process

Sub Club by RevenueCat · 2026-07-22

Higgsfield · AI video generator

Have the core team work ~17-hour days, six days a week

Said to lead the field now; no numbers

“They worked 17 hours per day. They made all the people work 17 hours per day”

Workedteam-process

Sub Club by RevenueCat · 2026-07-22

A mental health app company · guest's current employer

Translate the app into many languages with AI and test nearby markets

Her biggest win of the year; Turkey converts well; no numbers

“And with current AI, when you can translate any service in three hours into any languages, why don't you try different languages?”

Workedlocalization

Sub Club by RevenueCat · 2026-07-22

A web quiz diet app · host's wife's diet app

Sell a $1 plan after a 60-question quiz, then stack add-ons to ~$350

Buyer got half back after demanding a refund; invites regulation

“And then it takes her through a whole nother flow where it was like, "Do you want to add on this PDF? Do you want to add on?"”

Unknownweb-to-appupsell-expansion

Sub Club by RevenueCat · 2026-07-22

A mental health app company · guest's current employer

Run three apps at once with one team

Biggest fail of the year: attention spread too thin

“So I recently joined a team where they had three apps in the same time and they were trying to do it all in the same time.”

Failedportfolio

Sub Club by RevenueCat · 2026-07-22

Numbers mentioned

Subject	Metric	Value	Quote
EWA: Learn English	ad spend Ambiguous: '35K per month in traffic with a lot of spending' - likely paid traffic spend, could be revenue	USD 35K/month approximate	“when I joined it, it was doing 35K per month in traffic with a lot of spending”
EWA: Learn English	rank Category/app store ranking, second only to Duolingo	2 rank approximate	“at some point it was number two after a Duolingo”
EWA: Learn English	followers Instagram account, grown partly by paid ads	7M–10M followers estimate	“with EWA where we have a 10 million Instagram, probably seven million Instagram”
Yuliya's last company (unnamed)	activity volume Ad creatives tested per week (stated as a requirement: 'You need to test 500 ads per week')	500 ads/week approximate	“You need to test 500 ads per week.”
Yuliya's last company (unnamed)	ad spend Creative-test budget in a similar-converting cheap geo (e.g. Indonesia)	USD 10/day approximate	“you will see the same tests for $10 a day instead of $1,000 per day”
Yuliya's last company (unnamed)	ad spend Equivalent creative-test budget at US CPMs	USD 1,000/day approximate	“you will see the same tests for $10 a day instead of $1,000 per day”
Replika	count User messages claiming a real human was responding, pre-LLM ML chatbot	millions approximate	“we literally got millions of messages saying that you have a person in Ohio that is responding to me”
Solid Starts	team size Team members passionate about baby-led weaning (phrased loosely: 'if you have 20 people')	20 people estimate	“But if you have 20 people that are just about that, this is why Solid Starts is actually changing the whole baby-led weaning”
Higgsfield	count Working hours per day for founders and core team; six-day weeks	17 hours/day approximate	“They worked 17 hours per day.”
web quiz-funnel diet app that charged ~$350 after a $1 plan	price Advertised plan price at end of a ~60-question web quiz	USD 1/one-time	“"Hey, you can get your plan for a dollar."”
web quiz-funnel diet app that charged ~$350 after a $1 plan	price Actual total charged after pre-selected add-on upsells; about half refunded after complaint	USD 350–400/one-time approximate	“it was like 350 or $400 that it ended up charging”