DataHawkDataHawk
AppWebsite
WelcomeHelp CenterTroubleshootingChangelog

EMS v6 introduces a daily units-sold methodology

Four experimental EMS v6 datasets in data sharing, with a new daily ratio between observed sales and units bought per product.

Four experimental datasets are now (or soon) available in data sharing, updating EMS with a new daily units-sold methodology.

PRODUCT schema:

DATAHAWK_SHARE_V2.PRODUCT.product_sales_estimates_v6_child
DATAHAWK_SHARE_V2.PRODUCT.product_sales_estimates_v6_raw
DATAHAWK_SHARE_V2.PRODUCT.product_sales_estimates_v6

MARKET schema:

DATAHAWK_SHARE_V2.MARKET.product_sales_estimates_v6_raw

Each mirrors its _v5 counterpart but with new logic, so the content differs.

TL;DR

We've introduced a daily ratio comparing observed sales to units bought for each product. It's an early step toward a seasonality-aware EMS. The ratio isn't perfect yet; coverage depends on available sales data and will expand as we gather more. We plan to improve it by clamping extreme values, testing alternate calculations (sum/avg/median), incorporating Amazon calendar events (Black Week, Prime Day), and boosting unit-bought coverage.