Retail Analytics - Forecasting
Table Name:
retail_analytics_forecastingThe forecast tables are Amazon’s demand forecast on how much product they anticipate customers will order during a given time period (not how much they anticipate ordering from a vendor). It is a sell-out forecast, not a sell-in forecast.
Amazon provides 4 different forecasts, each broken out by 26 weeks: mean, P-70, P-80, & P-90. The P stands for "probability" and the number stands for a percentage of the statistical confidence interval. As an example at P-90, Amazon is 90% sure that customers will order up to the listed amount of product, while at P-70, Amazon is only 70% sure. As such P-90 represents Amazon's most aggressive forecast.
Generally, Amazon typically orders from vendors based on their P-70 level forecast.
Schema & update details
UI Report path | |
Amazon update frequency | Daily, with 48-72 hour latency. Learn more |
Reason update frequency | Every 8 hours |
Time Zone | Country local time zone |
Granularity [?] | 1 row per:
- ASIN
- Date
- Period
- Program
- Distributor View |
Historical data & change management
Time Period | |
History available from Amazon | Snapshot |
Reason lookback period | N/A |
