Multivariate grids test pixels. CO8 tests arguments.
Permuting headlines against backgrounds finds a local maximum. Testing genuinely different angles finds new winners.
| CO8 | Marpipe | |
|---|---|---|
| Testing unit | Distinct arguments — hooks, claims, motivations | Element permutations of one concept |
| What you learn | Which story your market believes | Which button color nudged CTR |
| Creative source | Angles mined from your reviews and offers | Assets you upload and permute |
| Ceiling | New concepts every cycle — no local maximum | Bounded by the concept you started with |
| Brand system | Identity encoded and enforced | Grid-level consistency |
| Monthly price | XXX | XXX |
Comparison reflects publicly available information; corrections welcome.
Proven in the wild
5K–20K
Ads shipped per month inside IM8
$200M+
IM8 year-two revenue, fueled by CO8 creative
$1B+
Comfrt projected 2026 revenue — the operator playbook behind CO8
5+
Channels covered — Meta, Instagram, TikTok, Pinterest, Google Display
Common questions
Isn't multivariate testing more scientific? +
It's precise about small questions. Permuting elements of one concept optimizes that concept — it never discovers that a different argument entirely would outperform it 3-to-1. Concept-level testing finds bigger wins; element testing polishes them.
Can CO8 do variation testing too? +
Yes — once an angle wins, CO8 produces disciplined variations of it: new hooks on the same claim, new formats of the same argument. Explore concepts first, exploit winners second.
What does a CO8 test cohort look like? +
Twenty distinct arguments, not twenty shades of one. Testimonial, comparison, offer-led, feature spotlight, social proof — each a different reason to buy, every one on-brand.
Keep exploring
See the difference on your own products.
Request access — no demo gate, no hidden pricing.
Request access