Forecasting demand for an online brand is its own sport. Demand is spikier than retail, a single creator can sell a month of stock in a day, and you are usually selling the same SKU across Shopify, Amazon, and TikTok Shop at the same time. The fundamentals of demand forecasting still apply, but a few things matter more for DTC.
Ecommerce demand forecasting predicts online sales per SKU and per channel, factoring in promos, ads, and viral spikes. The output drives your reorder points on each channel.
Why ecommerce demand is harder to forecast#
A few things make online demand jumpier than a physical store:
- Promotions and paid ads can multiply demand on a schedule you control, so they belong in the forecast, not as a surprise.
- Virality. One TikTok can 10x a SKU overnight. Your forecast has to react in days, not months.
- Multi-channel. The same SKU sells across channels that each move at a different rate and replenish on a different clock.
- Seasonality and launches stack on top of all of it.
The data to forecast with#
Pull these per SKU, per channel:
- Order history, with stockout days removed so missed demand does not read as low demand.
- Marketing calendar: upcoming promos, ad budget changes, and launches.
- Traffic and conversion signals when you have them, as an early read on demand.
- Seasonality from prior years where it exists.
Forecast each channel, plan each channel#
This is the move that matters most for DTC. Do not forecast a blended total. Your warehouse (Shopify DTC), Amazon FBA, and TikTok Shop each sell at different speeds and replenish on different timelines, so:
- Forecast demand for each channel on its own.
- Set a reorder point per channel using that channel's lead time.
- Hold safety stock sized to each channel's variability, which is usually highest on the channel most exposed to virality.
If Amazon is accelerating while TikTok cools, a company-wide average looks fine right up until FBA runs dry. Per-channel forecasting is the only way to catch it early.
New products with no history#
For a launch, start with qualitative input: comparable SKUs, pre-orders, and the ad spend you plan to put behind it. Switch to a data-based forecast after a few weeks of clean sales, and revise often, since new products move fast and early signals shift. The full method breakdown is in demand forecasting methods.
Forecast Shopify, Amazon, and TikTok Shop demand in one place
The bottom line#
Ecommerce demand forecasting is demand forecasting tuned for spiky, multi-channel, promo-driven online sales. Build promos and launches into the forecast, react fast to viral demand, and forecast and plan each channel separately. Enough Stock watches live demand across Shopify, Amazon, and TikTok Shop and turns it into per-channel reorder timing, so a viral week does not become a stockout.