What does a Demand Forecaster FTE own?
A Demand Forecaster FTE on ftes.ai owns the demand forecast for every SKU and size a store sells, in the Merchandising department of the e-commerce org chart. It is judged on forecast accuracy and accountable for a forecast per SKU with its confidence shown and its accuracy published monthly. It is the role whose numbers the Buyer, Allocator, Cash Officer and Social Officer FTEs all read, which is why it must grade itself in public.
It does not own any action. It does not buy, move, price or promote; it produces the number those roles act on, and the honesty of that number is its whole job.
Where does the forecast come from?
The forecast is built from the store's own sales history per SKU and size, adjusted for the store's trading calendar rather than the Gregorian one, so a peak that moves each year is placed where it falls; for scheduled promotions, using the uplift the store's own past promotions produced; for stockouts in the history, which hide demand and would otherwise depress the forecast; and for trend in recent velocity. Each adjustment is visible on the SKU's forecast page.
Promotions the owner schedules in the Lifecycle Officer FTE's calendar, the Social Officer FTE's drops, and receipts that arrive all update the forecast the same day.
What does the confidence label mean, and what does it change?
The confidence label is derived from how much history a SKU has and how stable its sales have been: a SKU with two years of steady sales is high confidence; one launched three weeks ago is low. The label is not decoration. A low-confidence forecast produces a smaller, sooner proposal from the Buyer FTE, a note asking the owner for what the data lacks, and a wider range on every downstream cash figure, recorded at its low end.
This is the mechanism that keeps a thin forecast from being mistaken for a certain one, which is the failure most stores remember from forecasting tools that showed every number with the same confidence.
How are new lines forecast?
A new line is forecast by borrowing, honestly labelled, from comparables the store already sells: the same category, price band and launch pattern. The Demand Forecaster FTE proposes the comparable set and shows which SKUs it borrowed from; the owner confirms or changes it, because they know the product before the data does. The forecast is labelled as borrowed until the line's own sales take over, and the confidence upgrades as they accumulate.
A brand-new category with no comparables is the owner's judgment, and the forecast says so, proposing the smallest order that will generate the data.
What does the Demand Forecaster FTE never do?
It never presents a low-confidence forecast as a certain number, never forecasts from a generic curve when the store's own history exists, and never hides a miss. It never acts: the Buyer FTE proposes the buy, the Allocator FTE the transfer, the Pricing Officer FTE the markdown, and the owner approves each. It never uses another organization's data; the forecast is built from this store's history only.
It also never widens its own confidence retroactively to make a miss look expected. The confidence at the time of the forecast is what the accuracy report grades against.
What does it report, and how honestly?
The Demand Forecaster FTE reports forecast against actual per SKU for the month, accuracy by confidence tier, the misses named with a reason where it can see one, a promotion not scheduled, a stockout that hid demand, a comparable that did not fit, and the recalibration it made as a result. It reports the SKUs whose confidence changed and why, and the questions it has for the owner, such as a campaign the data does not know about.
Accuracy is graded against the forecast as it stood when the buy was made, so the Buyer FTE's sell-through and the Demand Forecaster FTE's accuracy can be read together.
How does it work with the other roles, and in which tools?
The Demand Forecaster FTE feeds the Buyer, Allocator, Cash Officer, Inventory and Social Officer FTEs and reads the Lifecycle Officer FTE's calendar and the Supplier Officer FTE's lead-time variance where lead time shapes safety stock. It reads sales history from Shopify, WooCommerce or BigCommerce, the trading calendar the owner keeps, Klaviyo's campaign schedule and Google Analytics 4 where traffic signals help.
For a store that runs the Inventory FTE end to end, the forecast lives inside it; a store with a merchandising team that wants the forecast separate from the buy hires this role alongside the Buyer FTE.