How We Work
How Palette reads the market before backing a brand
Every category Palette enters has already passed an AI-driven demand read. We do not commit to a direction based on intuition or trend reports.
Real purchase sequences before category decisions
Palette processes anonymised purchase and reorder event streams from Indian e-commerce. We compute per-category reorder rates, repeat-buy intervals, and related purchase patterns across shopper cohorts. The output shows which categories have measurable consumer pull and which are mainly driven by trial spend. That read drives every category decision we make.
A demand signal should survive the second purchase
A real demand signal is a reorder rate above the median for comparable Indian SKUs, with a repeat-buy interval short enough to build around. Social virality and first-purchase spikes do not qualify. We look for categories where the first buyer comes back, because most category analyses optimise for the first purchase and ignore the second.
Three tests every category must pass before we act.
Reorder Rate Threshold
A category must beat our internal benchmark across independent purchase histories, not just one successful SKU. The benchmark is calibrated to Indian e-commerce category baselines so the comparison reflects actual shopper behaviour.
Geographic Spread
Concentrated demand in one city is not a brand. We need reorder behaviour across at least two distinct geographies, because a durable brand has relevance beyond one local shopper pocket.
Price Sensitivity
Demand must hold when the product is not discounted. In purchase sequence data, discount-driven reorder shows up as a drop in repeat buying when promotional support falls. That is not the same as brand pull.
Two models we use once a category passes the read.
If a category clears all three tests, Palette either builds from scratch or backs a founder who already has proven pull in that space.
Working in a category with reorder potential?
If you are building a brand in a category you believe has genuine reorder potential, tell us the read. We want to understand the purchase sequence behind the conviction.
Talk to Us