One keyword scrolled past 83% of the market.
- The actual product listing and its title
- The single diet-category keyword under scrutiny
- The platform's shopper mix across three cohorts
- The internal hypotheses (needs marketing, reviews, lower price)
- Each shopper's scroll-past vs engage response to the title
- The share of audience the keyword excluded
- How an alternate, tribe-neutral title performed
- Which cohort the keyword actually served
The company was sure the fix was more marketing, more reviews, a lower price.
A functional-food brand had built a genuinely good product and listed it on one of India's largest e-commerce platforms. Sales were underwhelming, and every internal theory pointed outward — spend more, discount more, gather more reviews. The simulation was asked a narrower question: before you spend, is the problem actually the listing itself?
The problem was one word in the title.
A keyword chosen for search volume was read as a tribe membership test.
The brand had put a diet-category keyword in the title to maximise search reach. To the shoppers who belonged to that dietary tribe, it was a signal to click. But to the 83% who didn't, the same word was an identity mismatch — an instant "not for me" that fired before they read the product, the ingredients or the price. The listing wasn't under-marketed; it was pre-filtering its own market at the title. No amount of review volume or discount reaches a shopper who has already scrolled past on word three.
Don't spend into a listing that rejects most of its audience. Rewrite the title around the product's functional benefit rather than a tribal keyword, keep the diet term lower in the listing where it helps the tribe without gatekeeping everyone else, and A/B the neutral title before committing media.
These are modelled outcomes, not measured ones. Every score and probability here is a synthetic-population prediction, produced before any real-market test. Thirty synthetic agents across three cohorts — sized to isolate the title effect, not to forecast category sales. Treat the numbers as directional — the direction and size of the effect is the finding, not a forecast. A design partner validates the top interventions against real behaviour before a decision commits to them.
“One word. 83% of your audience. Gone — before they ever saw the product.”SIMULATION SYNTHESIS · SEARCH-TO-LISTING ROUNDS
Run your version of this decision.
This is the real engine. Bring the thing you're weighing and see how the population responds — before you commit to it.