Many retailers are seeking to expand their business through new formats. Smaller format and convenience stores provide an excellent opportunity to reach new customers in grocery retail for example. New formats, business models and customers introduce uncertainty and costs. It is essential to both understand and predict this new demand in order to achieve profitable growth.
Blue Yonder works with many retail industry leaders who are constantly challenged by their customers and competitors to rapidly adapt their business to match market expectations. In grocery, for example, shopping habits are undergoing a fundamental change with the big weekly family supermarket shop in rapid decline leading to smaller more frequent shops.
While new formats and stores offer high potential to address changing customer expectations, it is essential to look at both the demand and supply challenge uniquely in the context of these expansion efforts, for example, replenishment. Incorporating new cost models, new supply constraints and demand profiles into the replenishment process is mission critical for fresh item demand planning in convenience stores. The very essence of a convenience store is availability; you win or lose on shelf availability.
New store and item forecasting has always been challenging. Typically the process is managed through a series of semi-automatic steps or interventions that effectively put these new demand profiles on ‘hyper care’ for a category or replenishment manager to deal with. Slow and manual processes can lead to sales and inventory problems, with new store forecasts having longer than desired periods of high forecast error until a stable model is established.
Tacking this ‘new demand’ on to existing store forecasting capabilities is not the answer. While consistent with existing processes, it totally neglects the need for agility and accuracy. Convenience store models require operational excellence from the business. Retailers also need precision from its data and forecasting capabilities.
There are three key replenishment and forecasting capabilities to consider for profitable new store growth:
- SKU level – Demand predictions need to be the best available at SKU level. This way, we identify products that have strong seasonality, are highly influenced by weather, price changes, promotions or other external and internal events, including new store openings.
- Store level – Prediction models also need to examine demand on a per-store basis (and learn what that means). This way we can identify regional and local patterns, but also understand how competition, demographics, store size, assortment and price level affect demand.
- Automation – Models need to rapidly learn how to combine per-product forecasts with a cluster-analysis of stores similar to the one that was just opened. That way, findings from stores with similar sizes, locations, assortments, price levels or demographics can be taken into account for every new store opening.
This three-step approach gives us very accurate demand forecasts from day one, limiting your risk of running out of stock and leaving a bad first impression with your new customers and brand strategy.
Blue Yonder forecasts delve into an extremely high level of detail. We don’t just make demand predictions for every single product in every single store for every single day, but we go way beyond that. Behind the scenes, every prediction is made 100 times, to cover every possible scenario and assign a probability to it. The extreme scenarios (almost nobody shows up or there are lines around the block) are equally unlikely, but they help us in making trade-offs depending on the high-level goals you have set for the store or the product category. This way we can reduce write-offs and waste for some products, while reducing the risk of out-of-stock situations for others.