Supply Chain Blog

Making a bid for improved FMCG Forecast Accuracy Levels

Posted by Dave Jordan on Tue, Jun 04, 2013

When a company asks me to take a look at their S&OP process yet boasts a Forecast Accuracy (FA) of 98%, alarm bells are immediately sound. If you are at a genuine level of 98% then of course there is a little more incremental improvement possible but it indicates your S&OP process is in very good health. Or does it?

After posing a few questions to junior staff I discovered that the company was in fact cheating itself. The 98% FA boast was based on volume alone. Oh dear, the sales team were patting themselves on the back when in fact accuracy was far from the case. They simply did not understand the chaos multi-sku volume FA measurements cause within the company.

Ok, so the revised level was 75%. Not the worst figure I have ever seen but at least it presented a clear opportunity for improvement in market performance. The newly generated data was flashed onto the screen and just as debate kicked off I noticed that the FA values were in a column headed by brand names. Oh dear, forecast accuracy by brand may well be a decent measure for sales and marketing category management evaluation but it does not lend itself to top class Customer Service achievement.

So the number reduced again to just below 50%. However, by this time the junior members could smell blood and went in for the kill. The latest figure quoted did not include imported products. You can get away with this approach if you have very few and relatively unimportant imported materials but not when the ratio is 75/25. I wonder what the logic was for the decision to omit imports.

Quiet descended around the room as the newly calculated number popped up as 28%, a far cry from the initial boast in the high nineties. At least we now knew the baseline for potential top and bottom line improvement. Like a classic Christmas pantomime, “oh no we don’t” came out of a few mouths who I guess had known these defects all along but were persuaded or out-graded not to reveal the ugly truth. “Shouldn’t we be including promotions in the calculation?”

Again, if you carry out little or no promotional activity then fair enough leave these out of the calculation if you wish, but when you are in a D&E market and promotions form a high proportion of sales you need to include them.

S&OP Planning Forecast AccuracyLike a negative live auction the number dropped out of the 90’s, passed the 70’s, quickly entered danger territory below 50% before almost settling in the 20’s before finally hitting near rock bottom at 16% FA. Sold! Or not sold, to be more accurate. The finally accurate and very low FA bluntly indicated that a huge opportunity existed to sell more products.

Yes, they had an S&OP process but in name only as they failed to measure FA accurately and continually underperformed. When measurement of a key performance indicator is flawed there is no way you can achieve your objectives as the target is moving and you are blindfolded.

With this calculation now fixed I will see how they get on but one thing is certain, they can only get better!

Image courtesy of Keerati at


Tagss: FMCG, Dave Jordan, Performance Improvement, S&OP, Forecasting & Demand Planning