Scroll

Insights and News /

23 May 2019 3 min read

The thrill of improving forecast accuracy

Sofia Gkiafi

Sofia Gkiafi
Senior Manager | Supply chain and procurement

It’s the small things in life that really do it for me – a day by the beach, homemade pistachio ice-cream, and an upward trend in forecast accuracy to name but a few.

This time last year, we embarked on what was to be a challenging yet rewarding journey with our client, a leading European Fast Moving Consumer Goods (FMCG) business. The exam question was simple: how do we improve our forecast accuracy? But there was a catch: no new technology. Fair enough, improving forecast accuracy is after all rarely about technology and most solutions in the market offer adequate functionality. Improving forecast accuracy is all about people and process – and yes, I know it sounds old school but at its core it really is – and we’ve proven it.

At this point I need to clarify that there was nothing wrong with either the people or the process in theory. We had seen this operating model working across the FMCG sector, however in this instance it just wasn’t producing the goods. The FMCG market in the UK is predominantly driven by promotional activity across the majority of channels. The incumbent design meant many hands were manipulating the forecast of each and every stock keeping unit (SKU), adding, moving and removing promotions, each trying to get the forecast just right. But as the saying goes, “too many cooks…”

We worked with the business to really understand the root cause of the problem and to identify the right solution, showcasing different operating models that ‘best in class’ FMCG supply chains have successfully implemented.

Our chosen design was one that gave control of a portfolio of products to a single demand planner, maintaining the strong link between Demand Planning and the Commercial business units. The product portfolio was segmented so planners would have a ‘compass’ for where to focus their efforts. Further, we made a couple of quick, simple changes to demand planning and commercial systems integration and we did some nerdy modelling to produce proposals for statistical forecast models.

Forecast accuracy doesn’t improve overnight and as expected, it didn’t. Demand planners were dealing with an enormous amount of change and there was a steep learning curve to navigate, and of course at the same time the business wasn’t standing still. However, for the past three months forecast accuracy has been improving week after week, following an incredible trend upwards and improving by 5-10%. The results are directly attributable to the brilliant and talented team of planners who made it all happen.

I wish I could put a pause to it now that we are on a high – but obviously that’s not an option here. There will inevitably be blips in forecast accuracy but I will not agonise about it as I know we have laid the foundations that will allow results to bounce right back.

There are exciting opportunities ahead with the use of new technology and analytics capabilities – the UK market will be ready for them.