Most founders underestimate the importance of ad creatives in marketing, focusing too much on fine-tuning the product to perfection. But sometimes, testing creatives is the only way to validate whether people actually need your solution.
Ad performance
We currently run approximately 500 creative experiments every month and plan to double that figure. And there's no need for skyrocketing budgets. Just a strategy and understanding of patterns that actually convert.
Let me show you exactly how we built a factory of creatives, starting from zero.
What a Factory Actually Means (And What It Doesn't)
Perhaps right now you are thinking, “If I am an early-stage founder with a limited marketing budget, does it make sense to read about some creative experiment factory?” Absolutely. Otherwise, you will waste your budget on pointless experiments and bring in irrelevant users.
Let me clarify right away: the factory of experiments is not about massive budgets, 20 marketers, and thousands of creatives out of the gate. It's about identifying the channels and formats that work for your product, pain points that trigger action, and scaling winning patterns.
Here is a step-by-step guide on how to build this system.
Step 1: Research Before You Create Anything
There is no point in starting an advertising campaign if you do not understand how the platform works. What goes viral on TikTok dies on Meta. What converts on Instagram gets ignored on X.
Conduct preliminary trend-watching research. My go-to tools for this are AdSpy, Meta's ad library, and TikTok's trends dashboard.
Start by observing which channels and content perform well for your competitors and screenshot successful organic posts. Save Reels that keep you watching past 3 seconds, and note which CTAs encourage you to click.
Join your audience's communities on Reddit, follow relevant hashtags on TikTok and Instagram, and read the comments — that's where users share their pain points, which you can use in your creatives.
Step 2: Build Your Creative Library
Once you've gathered enough references for your creatives, jump right in. I mean immediately, as trends don't last long, and you don't have time to waste.
Take your customers' pain points (not fictional ones, but real ones from user interviews, support, comments), and for each, develop different ways to articulate it, visual styles, hooks, CTAs.
Important: if you just copy a creative concept from another app, it rarely works. Once you understand why it performs and adapt it to your product, that's when you succeed. This means you need to combine an effective market format with your own unique angle. Then you will have a winning creative.
Step 3: Launch Tests, Identify Patterns, Scale Smart
Running 20 options at once, testing them for a week, and then analyzing all the results together will get you nowhere.
Proceed gradually.
- Launch 10 options with different hooks, and after a week, select the 3 that worked best.
- Develop new creatives with these winning hooks, testing different pain points, and again select the top 3 for scaling.
- Generate new combinations of winning hook + pain point, testing visual styles.
This way, you will progressively arrive at a winning formula and avoid spending resources on things that fail to deliver. With this strategy, you are narrowing down what works through elimination, rather than attempting to analyze everything at once.
If you have the budget, you can test more concepts in parallel. If you're bootstrapped like we were, you actually need to test more aggressively in a different way: don’t spend much on creatives that don’t show early traction, cut them fast, and quickly reallocate budget to stronger performers. In both cases, speed of decision-making matters more than the size of the budget.
Don't bet everything on a couple of creatives that perform best. Analyze in detail what exactly made them work, and then produce many iterations based on those successful ads: new variations, angles, visuals, and formats. Increase the quantity wisely and don't discard failed tests — this is important data that you will need in the future.
The Real ROI of Factory Thinking
My app now generates $470k in MRR and reached 1 million users. But the numbers aren't what matter most.
What matters: we can now identify what will work faster. That's the real ROI of factory thinking — you build an unfair advantage in speed of learning.
Don't compare your progress in building a system with one piece of viral creative from your competitors. While they were working on it, you had already tested fifty options and identified three models that actually drive conversions. By the time they launch their perfect ad, you've already scaled your winners and moved on to the next hypothesis.
Speed of learning beats quality of guessing.