Marketing Without Losing Authenticity

A lot of engineers treat marketing like something slightly embarrassing. I understand that. But I've come to think it's just this: helping the right people notice something that's actually useful to them.

Done well, that isn't manipulation. It's communication.

What this means

Marketing is getting attention, then earning the next few seconds, then earning trust.

Authentic marketing does that by showing people something true. The product doing what it does. A real result. A real person's reaction. The dishonest version fakes one of those. It usually works for a while and then costs you trust.

What I learned

HairOver's product is visual. That makes marketing easier in one way: the result is the pitch. One face with five different haircuts explains the app better than any description.

So the marketing rules I wrote for HairOver's content are mostly about getting to that result quickly and honestly:

• Show the result early. Problem → visual → curiosity → product.
• Don't open a video with the logo. Nobody cares about the logo in the first second.
• Ask one simple question. "Which one would you actually get?"
• No engagement bait. Replies should add a real thought.
• Judge by installs, not likes.

Hooks, simply

A hook is the first second or two that decides whether someone keeps watching.

The best hooks don't trick people. They show the right person, immediately, why this is for them. For HairOver, that's the moment of "wait, that's the same person?" when a face changes haircut. For a different product, it might be the problem, stated so precisely that the right person thinks "that's me."

A hook that attracts the wrong people isn't a good hook, even if the numbers look good.

Real example: experimenting with AI UGC

UGC (user-generated content) means videos that feel like they come from real users: someone holding their phone, trying the app, reacting. That style works well on TikTok, Reels, and Shorts because it looks like a normal post, not an ad.

Real creators take time to find, though. So I started experimenting with AI-generated UGC-style content for HairOver. From my own files, the experiments so far include:

• An AI-generated creator-style person filming a casual selfie video in a bedroom
• The same person holding a phone with a green screen, so real app footage can be placed onto the screen
• Several vertical point-of-view videos, including a "stylist POV" angle, and an ad video

The green-screen idea is the part I think is right: the app on screen should be the real app. The product demo stays true even if the person holding the phone is generated.

I don't have results from these yet. I don't know if they'll perform, or if they'll look fake to people. That's still being tested.

The honesty line with AI UGC

Here's the question AI UGC raises, and I'd rather deal with it openly.

Using AI to show how a product works is fine. It's a demo, like a screen recording.

Using an AI-generated person as a fake customer is not. An AI person saying "this app changed my life" is a made-up testimonial. It misleads people, and it's exactly the kind of thing that destroys trust once it's noticed.

The line I think is right:

• The product footage must be real.
• An AI person can demonstrate. It can't testify.
• Reviews and testimonials come only from real users.
• Social proof numbers ("join X people") are only used when they're true.

Social proof, reviews, and testimonials

Real social proof is one of the strongest things you can have, and one of the most tempting to fake.

Ask real users for reviews. Share their results with permission. The DM template I wrote for this is simple: "Your Hairover result looks great. Would you be okay with us featuring it? We'll tag you."

Then only use what you actually have. If you only have a few reviews, show a few reviews.

Ads and testing creatives

Ads let you pay for attention directly. They're most useful for testing: showing two different hooks to similar people and seeing which one gets real action.

Test one thing at a time: the hook, the format, or the message. Measure installs or signups, not views. And don't scale spending until you know people who arrive actually stay and pay.

What to do

1. Find the moment your product proves itself and put it in the first two seconds.
2. Write your honesty rules down before you make any content, especially AI content.
3. Keep product footage real, always.
4. Collect real reviews and results with permission.
5. Test hooks against each other, and measure actions instead of views.

Mistakes to avoid

Fake testimonials, AI-generated or otherwise.

Starting with the logo instead of the result.

Engagement bait that brings the wrong audience.

Inflated social proof. "Thousands of users" when you have dozens.

Scaling ads before the product converts.

My current thinking

AI makes content cheap to produce, which means honest content will stand out more, not less. I'm still experimenting with AI UGC for HairOver, and I don't know yet if it works. But I'm fairly sure about where the line is, and I'd rather stay on the right side of it than win a few extra clicks.

Key takeaway

Marketing is helping the right people notice something true. Show the real result fast, keep product footage real, let AI demonstrate but never testify, and only use social proof you actually have.

Next: Metrics That Matter.