Hey legends! 👋
SURI's playbook caught my attention because they mine customer reviews for creative, stress-test hooks before scaling spend, and let the data pick winners. That's the real AI leverage, not flashy video generation, but turning your existing customer voice into a conversion machine.
Enjoy the strategy deep dive. Hope you get something out of it!
-Dev
THE PLAY
Using AI to power creative testing and scale
For decades, the toothbrush market has been a "clinical duopoly" where two major brands hold 75% of the market.
SURI saw a gap in an antiquated category. They noticed customers wanted fewer gimmicks and a product that wasn’t perfect but an improvement on the options they had. Instead of spending millions on traditional commercials (prominent in this category), SURI took an AI-first approach.
They used AI to test hundreds of creative variations to find out exactly what makes a modern shopper stop scrolling. Using an AI-driven creative engine and optimized their website for “agentic discovery” across LLMs like ChatGPT and Gemini, they were able to drive up their customer acquisition.
They ended up selling over 5,000 units in their first week and now have over 500,000 active subscribers, proving that any boring hardware product can be scaled when performance marketing is powered by machine learning.
THE DEEP DIVE
Review AI → Scrappy Hooks → Meta Scale
1. High velocity ‘scrappy’ ads using AI generated better results
SURI uses a custom GPT on 15,000+ Trustpilot reviews to extract "love" phrases and pain points (gunk, batteries), fueling dozens of 3-second hook variations weekly.
This level of speed in their testing showed them that scrappy UGC beats polished ads, moving them away from big budgets and adopting a leaner approach that let’s them produce 10 ads in 4 hours from existing assets.
This shift was huge for them because it allowed them to constantly iterate on winning angles (e.g. us vs them angle, waste narrative angles proved the most effective).
Here’s their 3-step lean AI ad creation framework:
Download existing UGC footage, product b-roll and customer testimonial clips.
Use simple, easy-to-use editing software to create new hooks
Test 3-second hook variations (different creator faces, opening pain lines like “gunk disgust” vs “battery fail”)
The learning is simple: Whip new hooks quickly to see what lands without investing too much ad budget or creative build resource. Use AI & simple software to generate 10+ variations of an ads first 3-seconds (the hook) to find the one emotional trigger that stops the scroll. Then double down.
2. Combining customer voice with AI to uncover angles
SURI built a custom gpt and fed it over 15,000+ verified customer reviews to fgure out the specific words that drove conversion.
This helped them find the most frequently used words, one of them being “love” an insight allowing them to pivot their marketing from clinical specs to “brand advocacy”. By knowing users “love” their product and share a strong sentiment, they were likely to naturally gift it to a family or friend.
Another way they used their customGPT was to analyse initial surveys that showed a universal hatred for “sink gunk”. Naturally, this led them to create different variations of ads that promoted their magnetic mirror mount. A visual cue that showed how SURI tackles this pain point.
3. Meta scaling and LTV optimisation
While media buyers are complicating the Meta ‘setup’, SURI knows what works for them. They opt for a lean and consolidated structure that allows them to optimise for creative signal and lifetime value.
This is how they do this:
Run 1-2 broad prospecting campaigns with 1 retargeting ABO campaign. They avoid CBO campaigns or micro-audiences to force the pixel to learn purely from the best creative signal interactions.
Optimise exclusively for purchasers and subscribers, forcing Meta to value high LTV signals over cheap traffic.
They only launch new campaigns when something major changes, like a new offer, price or expanding into new countries. This keeps Meta’s algorithm learning from high-quality data across the entire ad account/ongoing campaigns.
They don’t rotate creatives to “stay fresh” and only tweak with text overlays and hooks when performance (MER + ROAS ) drops. One of their hero video ads has been running for 266 days due to evergreen, authentic scroll-stopping power.
Thanks for tuning in this week. Let me know how you’re finding this new, more concentrated format? Catch you on the next one!
-Dev
Poll: What did you think of today's edition?
When it makes sense for you, here are 3 ways I can help you:
1:1 Growth Consulting - if you need help with dialling in your ads, performance, creative and conversion strategy. Jump on the waitlist here.
Free Resources and community - I'm slowly building this out so it becomes a hub of useful tools and a way to connect with other ecom entrepreneurs and marketers.
Partnership - if you want to advertise your brand in the newsletter. You can express your interest here.