Experimentation intelligence for fashion and retail.
See which experiments brands run, keep and roll back. Then simulate your next experiment on AI shoppers before it uses real traffic.
See it on the brands you compete with · or start with the free biweekly digest
A/B See
Every live experiment, in one place.
280+ brands, tracked daily. Filter by brand, page or goal. Every card is a real experiment: what changed, the hypothesis behind it, and whether it shipped, was reverted or is still live.
Product page separate new colors row
Homepage chat widget addition
Mini-cart inventory warning addition
Navigation menu category rename
Category page discount display removal
Cart page promo code accordion addition
See the brands you care about, in your inbox.
An experiment, opened
Click an experiment. See the full story.
Every card opens into the full read: the real control next to the variant, what changed, the problem it solves and the takeaway.
How we call outcomes
We watch what survives.
A/B See tracks which variant a brand keeps serving after an experiment ends. Shipped, reverted or ended: the call comes from rollout patterns, audience scope and how long the variant has stuck.
Colours in one long list
New colours in their own row
The variant became the live page
A/B Simulation
Run the experiment on simulated shoppers first.
Agents built from your own shopper behaviour, plus everything A/B See knows the market has tried, shop version A and version B. You get a likely winner before the experiment uses real traffic.
Each dot is one simulated shopper. Bright dots added to bag. Illustrative run.
Proved on your past experiments
The agents predict experiments you have already run, without seeing the result. You see the hit rate before you trust a forecast.
Forecast the next one
Share an idea. Thousands of agents shop both versions and you get the likely winner, and which shoppers it moves.
Sharper with every experiment
Each real result is compared with its forecast, and your model recalibrates to your customers.
See A/B Simulation run on your own experiments.
What you get
One workstation for every experiment decision.
- 0101Evidence
Fill the Because in every hypothesis.
Search similar experiments across 280+ brands. See which changes stuck and which were pulled, and bring that evidence into the brief alongside your own data.
colour swatches on product pagesticky size selectorfit quizApple Pay above emailbundle drawer - 0202Patterns
See where the market is heading, page by page.
Three brands moving the delivery date next to the add-to-bag button this month? You see it that week, not next quarter in a deck.
PDP CTA above fold+3 brandsMarAprMay - 0303Following
Follow the brands you compete with.
Pick the brands that matter to you. We tell you when a new experiment goes live, when one ships, and when one is quietly pulled.
Your watchlist3 alertsTory BurchNew experiment liveNew BalanceShippedLevi'sPulled · reverted - 0404A/B Simulation
Spend real traffic on your best bets.
Experiment slots are scarce. Forecast an idea on simulated shoppers before it takes one, then run the strong ideas and rework the rest.
1New colours in their own row on product pages2Low-stock warning in the mini-cart3Promo code tucked into an accordion in cart
Pick the brands you want us to watch.
Know what the market keeps.
Forecast what you ship.
See A/B See and A/B Simulation on your own market, or start with the free biweekly digest.