Common demand forecasting mistakes, and how to avoid them

Most forecasting mistakes aren't about the maths. They come from bad inputs, missing context and forecasts that nobody updates. The result is the same each time: buys and commitments set too high or too low, then stockouts on the lines that sell and markdowns on the ones that don't. Here are the mistakes we see most, how they inflate commitments, and what to check before you buy a forecasting tool.

Seven common forecasting mistakes

1. Forecasting from sales that were capped by stockouts

If a line sold out in week four, its sales history shows what you had, not what customers wanted. Forecast from that and next season's buy is cut on exactly the lines that were short. Clean the history for stockouts before you forecast from it.

2. Ignoring promotions and events

A promotion, a holiday or a weather spike lifts sales for a few weeks. Leave it in the baseline and the forecast expects that lift every year. Take it out and plan it back in only when it is happening again.

3. Over-trusting last year

Last year is a starting point, not an answer. New stores, closed stores, a price change or a shift to online all change demand. Adjust the history for what has changed before using it.

4. One forecast for every store

A total forecast split evenly across stores sends too much to weak stores and too little to strong ones. Forecast by store or store cluster, and by size where size matters.

5. Ignoring lead times and stock on hand

A forecast that doesn't know the supplier needs ten weeks, or that there are already six weeks of cover in the warehouse, will order at the wrong time and in the wrong quantity. Forecasting and replenishment need to see the same stock and the same lead times.

6. Not updating the forecast in season

A forecast set before the season and left alone is out of date by week three. Re-forecast weekly against actual sales, so re-orders and markdowns follow what is really selling.

7. Trusting a forecast nobody can explain

If planners can't see why the forecast says what it says, they either follow it blindly or ignore it. Both are expensive. Use forecasts the planner can see into and override, with the override recorded.

How forecasting mistakes inflate commitments

Each of these mistakes feeds straight into the buy. An inflated baseline from a one-off promotion raises the order for every week after it. A forecast that ignores stock on hand orders units that are already in the building. Because buys are committed weeks or months ahead, the error is locked in before the first sale proves it wrong, and it comes out later as overstock and markdown.

Why demand planning projects fail

• The data isn't ready. Stock on hand, lead times and sales history are inaccurate, and nobody owns fixing them.
• The tool sits apart from planning. The forecast lives in one system while the buy and the MFP (WSSI) live in another, so it is never acted on.
• Nobody trusts the numbers. Planners can't see the reasoning, so they keep their spreadsheets.
• It takes too long. A long implementation loses momentum before anyone sees value.

What to check before you buy

• Will the vendor run it on your own data before you sign?
• Does it clean sales history for stockouts and promotions?
• Can it forecast by store, cluster and size?
• Does the forecast feed the buy, open-to-buy and replenishment directly, without exports?
• Can planners see why it forecasts what it does, and override it?
• How long until it is live, and what does that depend on from your team?
• Does it keep your ERP as the system of record?

merchmix's demand forecasting runs inside the same system as the MFP (WSSI), open-to-buy and replenishment, so the forecast feeds the buy directly. You can test it on your own data in a free six-week proof of concept, with no contract. Typical go-live is 12 to 15 weeks.

Start a free 6-week POC  |  Book a demo

FAQ

What are the most common inventory forecasting mistakes to avoid?
Forecasting from sales history that was capped by stockouts, leaving one-off promotions in the baseline, using one forecast for every store, and ignoring lead times and stock on hand. Not updating the forecast once the season starts is close behind.

What forecasting mistakes inflate commitments?
Promotions left in the baseline, forecasts that ignore stock already on hand, and a single total forecast split evenly across stores. Each raises the buy before the season starts, and the excess comes out later as markdown.

Why do demand planning implementations fail, and how do you reduce the risk before buying?
Usually because the data isn't ready, the forecast sits apart from planning, or planners don't trust it. Reduce the risk by testing the tool on your own data before signing, checking that it feeds the buy directly, and making sure planners can see and override the forecast.