Do’s and Don’ts of Demand Forecasting

'The best advice: start small and take the plunge'

October 6, 2015, 12:12 pm By John Karolefski,

Grocery retailers engage in demand forecasting by deploying sophisticated technology, using a merchant’s savvy, or a blend of both. For smaller operations, a lot of forecasting and demand prediction is based on gut-feel and judgement at the store level. The business of large chains is more complicated, so staying with manual ordering, intuition and what worked in the past become roadblocks to success and growth.     

When doing research for a feature story on this topic, I recently spoke with several experts in demand forecasting. What do they say grocery retailers should do and not do? Much of their advice, and some of my own, is below:

First, three Don’ts:

Don’t rely only on historical patterns: Understanding historical patterns is only one dimension to demand prediction. Forecasts that are actionable and developed in near real-time are better than relying only on historical patterns or even sophisticated post-facto analysis.

Don’t staff teams with junior analysts: Demand forecasting involves supply chain, store and DC replenishment, merchants, marketing, and planning departments. Senior executives and an internal champion must be on the team for demand forecasting to succeed and become a true asset to the organization.  

Don’t start and stop: Setting up a demand forecasting team is only the beginning. As the team members become familiar with the technology and benefits, it will be time to build on the success of first steps to expand to the rest of the business. Don’t stop with the prediction of promotion demand.

Here are three Do’s:

Focus on two fronts: Addressing out of stocks with a single forecast mentality is insufficient. Focus on two fronts: The distribution center to support service levels with better demand forecasting, and at store level with better forecasts.  

Account for seasonal variations: Seasonal variations are a key factor in the success or failure of a forecast. There is always an estimated 10 percent accounting for waste and undersold products – all the more in seasonal items.

Measure demand forecast effectiveness: Be as specific as possible. Share success with top executives, who need to be supportive of the initiative. Identify a corporate champion for demand forecasting – the more senior, the better.

To forecast demand, smaller grocers often rely on the previous year’s shipment withdrawal data or transaction data in Excel. Useful technology is available, but the investment always gives retailer pause – especially those with limited resources.

The best advice: Start small and take the plunge.