Nancy is a typical Clear Demand client. She works at a leading convenience store retailer. She is planning a promotion using our Promotion Planning Solution. The tool identifies the best products and offers for driving traffic to the store, building basket sales, and utilizing vendor deals. In this case, the tool recommends an offer of 2 for $3 on a large-size candy bar. The recommended offer requires that the shopper purchase two candy bars to receive the promotional price. Otherwise, the shopper pays the regular price of $1.89 for one candy bar. Nancy is intrigued that the Clear Demand tool provides a forecast that shows how many units will be sold by each store and that each store’s sales are broken out by regular sales and promotional sales. In this case, the tool shows a forecast of 102 units at regular price and 231 units at the 2 for $3 price. The tool also shows that without the promotion, the store would sell 147 units at regular price. The tool also shows forecast profit, revenue, and margin.
Nancy noticed that the forecast is broken out into components including baseline sales, ad lift, and offer lift. Further, she sees the impact of seasonality and even a projection of lost sales due to stock outs. Finally, she sees the impact of cannibalization and affinity that results in promoting the candy bar.
She called Clear Demand’s support desk to ask about using the forecast to drive store orders. Nathan, one of our support staff took Nancy’s call, and explained that yes, the Clear Demand forecast is extremely useful for driving store orders. He described how the confidence interval can be used along with targets for safety stock and display stock to achieve target service levels.
Nancy asked, how accurate is the forecast. Nathan explained that accuracy is measured in terms of forecast bias, forecast error, and confidence interval accuracy. Nancy was familiar with other forecasting solutions and observed that Clear Demand was much more accurate. Nathan explained that the accuracy is based on some very advanced mathematical algorithms developed and patented by Clear Demand. In particular, the forecast is generated from store-level demand. Other solutions require aggregating stores into zones to deal with data sparsity issues. In particular, other solutions depend on guess-work about store-level availability and their algorithms are very sensitive to bad guesses in the availability. Clear Demand uses an innovative algorithm that models store-level demand without this problem.
Nathan communicates that the improved store-level accuracy is worth real money by reducing lost sales from stock outs and reducing excess inventory. A recent case study showed a $2.2M savings for a $1B retailer.