Forecasting
Forecasting is a service under MixShift Intelligence, on the same catalog, access gate, and metering Attribution already uses. Intelligence is open to every account by default. Forecasting has three requests today.
- The readiness check asks whether a merchant has enough sales and ad-spend history to support a forecast, how much of the required floor it already covers, and what would close the gap. It fits no model itself.
- The forecast baseline returns the forecast as the MixShift forecasting app published it, with its users' corrections applied: the model, its monthly expectations and 95% ranges, the measured backtest, and what the forecast was trained on. It covers a merchant or one of its sub-brands, in revenue or units, and says how old the published copy is. Where nothing is published for a merchant with enough sales history, MixShift computes a revenue baseline from your warehouse data with the same engine the forecasting app uses and labels it as computed. A published forecast always takes priority.
- The report-month tracker returns the forecast figures for one report month in the shape the monthly report reads: the year to date, what the model forecast before the month closed, what the current model projects for it, and the months still to come. It says whether the forecast is current, out of date, or not published, so a report shows a forecast only when it is current.
When a published forecast was withdrawn, when it is below the history floor, or when nothing is published and the warehouse history is too short to compute one, the answer says so, with the reason, rather than guessing.
Enter your own corrections
With the corrections:write scope you can enter your own forecast corrections, such as a month's sales, ad spend or budget. They are saved under your name in the same place the forecasting app reads, so a correction made in one place shows up in the other. Interactive sign-ins have the scope by default, and a tenant admin can grant it to a service credential. The endpoints are in the Builder Platform reference.
What it is
A top-down revenue forecast built from the brand's own historical performance. The forecast establishes an expectation before the period begins. As actual results arrive, Attribution measures the beat or miss, identifies where it occurred, and explains the metrics and products carrying the difference. Expected performance, actual performance, and what explains the difference.
The objective is not false precision. It is a defensible performance baseline, grounded in the brand's own history, that a team can plan against.
It is a top-down forecast at the level of a merchant or a sub-brand, not a bottom-up build from individual products. From the plugin, Forecast Max can forecast one brand inside an account on its own; see Get a forecast with Forecast Max.
Alongside it
Two more things give you a forward-looking read, besides asking the plugin for a forecast on request (see Get a forecast with Forecast Max):
- Brand memory, the brand context and change log every Intelligence service reads. A promotion or a launch recorded there is something your reports can cite when they explain a period, so recording what is coming pays off in every report. See Work with brand memory.
- Attribution, specifically the monthly read, which gives you month-over-month and year-over-year in one bundle today. See What Intelligence answers.
Related
Shifty is the MixShift docs assistant. Pick a question or open the chat for anything else.