For my entire professional career, I've been fighting with my founder-bosses over creatives.

Nobody liked my creatives.

too ugly

too "off-brand"

too much about things the product doesn't do.

Creatives that show features we don't have. Using neon colors that have nothing to do with the brand.

Everyone wanted polished banners and videos that sell the product head-on. Even better if it's some arthouse story that doesn't sell anything at all.

Unfortunately, that doesn't work.

Non-marketers hate misleads. Because they "ruin the brand", attract "wrong users", and overall feel "bad".

Marketers love misleads. Because metrics are better, CAC is lower, and ROAS is higher. And otherwise, you usually just can't market effectively.

In this newsletter, I'll talk about the concept of misleads, how general AI wrappers really grow, why users actually don't need nano banana, and what 4X strategy games have to do with any of this.

I've spent my entire life working on marketing for consumer B2C mobile products:

What all these products have in common: they're mass-market. And there are hundreds of other identical products.

Audience – everyone. Competitors – thousands.

These products (tbh any products) can't be sold head-on.

I had to use misleading creatives. What even are those?

Misleading creatives are ads that put users under the wrong impression about:

The more general the app, the more you have to rely on misleads. Simply because otherwise it's impossible to advertise such a product.

Let me explain with specific examples.

From the market I'm in:

All of them let you do anything: search for information, answer simple questions, write SEO texts, emails, college essays, create images, etc.

Then you have:

Niche apps are very easy to differentiate. But that doesn't always mean they're easier to market.

General apps – you can't really differentiate them at all. But that's where the biggest money is.

But you can add complementary viral features to your product that don't strengthen your core product at all, yet still help attract your target audience into the core loop.

GPT vs Gemini war through AI image models is the best example.

After Nano Banana launched, Gemini got close to GPT in downloads.

Six months earlier, GPT itself spiked after launching text-to-image and image-to-image models.

A year ago, I was promoting OpenChat using text-to-image creatives. I spent over $200K on that one creative concept.

But here's the real point:

Users who come through a mislead-ish specialized feature, within a short time end up using 99% classic text-to-text and search. Because you won't generate images every day, but you will use AI instead of Google every day.

Same with GPT and Gemini – they invest in image models not because that's their core value, but because those features bring in cheap users. It's marketing.

You can't impress people with text generation or simple search anymore.

But you can impress them with image generation.

Even though they don't actually need it in daily life.

That's the paradox.

You acquire with one feature, you monetize with another.

Gaming is famous for misleading creatives.

Hyper-casual games get the cheapest traffic. Fast, addictive mechanics give extremely low CPI.

The problem: hyper-casual games barely make money on in-app purchases. And overall, there aren't many hyper-casual games on the market with any significant revenue.

Mid-core games, on the other hand, make the real money.

Take 4X strategy games.

Huge revenue, high ARPPU.

But payback is 1–2 years, and the cost of acquiring a payer can be hundreds of dollars.

Misleads saved this category.

Stage 1: use hyper-casual mechanics in creatives – even if you don't have them in the game

Stage 2: add those mechanics as mini-games inside

Stage 3: integrate them deeper into the core loop

So initially, these mid-core games used fake gameplay in creatives to lower CPI. As a result, thanks to the large funnel of cheap incoming users, they reduced CAC significantly.

At some point, Meta and Google started fighting against such creatives, and they also began performing worse.

As a result, the fastest and "smartest" projects began integrating fake mechanics as mini-games.

Usually, it didn't look that great. But it solved its purpose.

And few projects actually focused on implementing fake gameplay into the main core loop.

Those who managed to do it most successfully are making a lot of money.

For example, Last War – making $150-200 million per month.

Everyone's happy. Well, except maybe some users.

A couple more cases…

A local food aggregator that had a bonus-points system: new users got N points, could spend them only in some restaurants, on some dishes, with a minimum order amount – a lot of conditions.

Which creatives worked best?

Order food for free!

Food for points!

Get a burger for free!

Founders were unhappy. Support was unhappy. Some users were unhappy.

But our financial model was very happy: marketing paid back like never before, conversion to first order was great, retention unchanged, and growth accelerated.

I used the same strategy later for an e-commerce aggregator. Works flawlessly.

The product was an app with a big library of full video workouts across different categories. At that time, most apps on the market were low-quality GIF-based exercise demos.

The problem: in creatives, the GIF-style animated visuals performed much better, meaning the fake, low-quality visuals converted better than our real product.

Everyone hated me, but I couldn’t do it any other way, because those creatives were printing money.

And yes – I copied that creative entirely from a competitor. Thanks to them.

This newsletter was really tough for me to write.

If you're not too lazy, click if you liked it or not.

And if you have time, maybe write back with feedback? Even if it's negative. Trying to figure out what actually works here and what doesn't. Any feedback helps.