The problem
Digital activity leaves a footprint everywhere, and eCommerce is no exception — every review, every stock change, every livestream, every ranking report matters. The hard part isn't collecting that data, it's spotting the real signal inside all that noise, across markets, without chasing every data point by hand.
What I did
Working with the Analytics team, I helped design a market-share analysis product built on trend models. We combined six candidate signal metrics into one algorithm and checked how each one actually correlated with sales:
We used models that analyze the behavior of each individual data point to surface the most accurate, key changing endpoints — the moments that actually signal a shift, not just noise.
What happened
The result is a market trend report that's roughly 90% accurate. One output: tracking market share shifts between platforms over time.
This is now used across Nutrition, F&B, Personal Care, Home Care, and Pet Care categories — one of the most-used products by senior management chasing an edge over competitors.