Consumer goods and FMCG
Your forecast is wrong, and you don’t know by how much. A model measures the error and plans stock and production around it.
- Free 30-minute call
- No commitment
- You own the code
Typical problems in consumer goods and FMCG.
The promotion you paid for starts on Monday, and by Wednesday the shelves are empty. Meanwhile, last month’s new flavour still fills the warehouse.
Forecast is wrong
Every forecast misses. Nobody can tell you by how much, so nobody knows what the miss costs.
Stockouts and excess
One flavour sells out in a week while another fills pallets in the warehouse. Both come from the same plan.
Promotions misfire
The two-for-one runs in every store from Monday. The plan didn’t include the extra demand, so the shelves are bare by Wednesday.
You don’t trust the forecast
The system proposes an order. You double it, because the forecast never explains itself.
In short: An FMCG demand plan forecasts what each fast-moving consumer good (FMCG) will sell, promotions included, and turns that forecast into stock and production. It measures how far off the forecast usually is. The bigger this error for a product, the more reserve it plans for that product. This reserve is called safety stock.
What a wrong forecast costs you
Monday morning, and the two-for-one promotion goes live in every store. By Wednesday the shelves are empty, and the extra stock is still on a truck. In the warehouse, pallets of last month’s new flavour wait for orders that won’t come. Forecast error is daily life in fast-moving consumer goods (FMCG), and every error ends up in your stock and production plans.
If the forecast is too high, you pay for excess stock and wasted production. If it’s too low, shelves go empty during the promotion you paid for. You don’t trust the forecast, so you overrule it on gut feeling. Nobody can say how wrong it is, so nobody can say what the error costs you.
Plans built around your forecast error
Today, the system proposes a number, and you change it, because nothing tells you how far to trust it. With our model, you see how far off your forecast usually is for each product, in promotion weeks and in normal weeks. You also see what that error costs in excess stock and empty shelves. So you adjust only where you know something the forecast doesn’t.
Our model sets safety stock and production from the measured size of the error. Safety stock is the reserve you keep in case you sell more than forecast. Every buffer comes with its reason, so you can see why it is that size. You stop paying for buffers you don’t need, and shelves stay full during the promotions you paid for.
Measuring forecast error in FMCG demand planning
FMCG demand planning starts with your statistical forecast. It reads three things from your sales history: how much you normally sell (level), whether sales rise or fall (trend) and how they change with the season. Promotional forecasting estimates the uplift, the extra sales a promotion brings. Safety stock optimisation then sets the buffer for each product from the measured error. Replenishment planning turns that buffer into orders.
Do your overrides help or hurt?
Almost every demand planning team overrides the system. Few measure whether it pays. The method is called forecast value added (FVA). You compare the error of three forecasts at the same horizon: a naive one, the statistical one and the final one after overrides. Each step should beat the one before. Split the result into promotion weeks and normal weeks, because overrides often help in one and hurt in the other.
Use a weighted error, so a big seller counts more than a slow one. In one practitioner survey, over half the respondents used WMAPE, the volume-weighted mean absolute percentage error (Forecasting Blog benchmark). And keep an open mind about the result. Forecasting researchers Fildes and Goodwin found that overrides “rarely add value but always take time” (as cited by Logistics Viewpoints).
How a project runs.
Each of the three phases ends with something you can check before we go any further.
-
Understand
We start with your operation. What do you plan, where does it go wrong, and what would a better week look like? Then we tell you straight whether optimisation can help, and sometimes the answer is no.
- A written assessment
- A clear go or no-go
-
Model
We write your decisions, goals and rules into a mathematical model, which is just a precise description of how your operation works. Then we run it on your own data. You see side by side what each plan costs and what it delivers, and pick the one that fits your priorities.
- A documented model
- Results on your own data
-
Implement
We test with your data and go through the results with your team. Once they trust it, we connect the model to your systems, so it runs every day without anyone stepping in by hand.
- Source code and documentation
- Training for your team
Common questions.
Something else? Ask us directly.
Mostly because it doesn’t explain itself. A number with no reason behind it gets overruled, however accurate it is. You start using a forecast once you see two things: what moves it up or down, such as a promotion or the weather, and how far off it usually is. Then measure your overrides, so you learn where the system forecast is right and where you are.
Report two numbers. The first is WMAPE, short for weighted mean absolute percentage error. It tells you by what percentage the forecast is off across all products. Products that sell a lot count more than products that sell little. The second is the bias. It shows whether your forecast is too high or too low on average. Measure both for the moment when you order. A forecast for next week doesn’t help if you have to order six weeks ahead.
Start from past promotions: the discount, the display, the season and what sold before, during and after. Your forecast should separate two things: the extra sales during the promotion and the dip afterwards. The dip often comes because customers have stocked up. It should also flag products that steal sales from each other. Then you stock the promotion for the uplift you expect.
It depends on the problem, so we agree the timeline with you up front. Every project starts with a diagnostic, a short look at how you plan today. It shows which data, how much time and what budget a model for daily use needs. Only then do you decide.
No. We connect to your ERP, your warehouse management system or your planning spreadsheet. We build the model only for the planning that gives you trouble today. Your systems keep running as before.
Almost nobody’s data is clean. The diagnostic shows what the model needs and what’s missing. Then we build the model to cope with messy data.
Off-the-shelf tools are built for many companies at once, so they don’t know your special rules. We build a model that follows exactly your rules. Then we hand it over, and your team owns it.
Similar pages.
Terms on this page
Every abbreviation explained in plain words, in our article “Metrics and methods”.
- WMAPEWeighted mean absolute percentage error
- Forecast bias
- FVAForecast value added
- FMCGFast-moving consumer goods
Tell us what you’re planning.
In a free 30-minute call with a specialist, you find out whether optimisation can help.
- Free, no commitment
- You talk to a specialist directly
- A clear go or no-go