Every retailer strives to maximize its return on the money and effort it places into various investments. The question is which parts of these investments will drive the biggest and most certain returns? Increased forecast accuracy is the single biggest driver of direct savings, revenue, and total return on investment.
Spend precious open-to-buy dollars on the product most likely to succeed and improve planner productivity! Retailers of all maturities are looking to automate forecasting and replenishment processes to improve planner productivity. With increasing pressure on margins, retailers can benefit from using automated demand forecasting systems to do the “heavy lifting”, decreasing the need for hiring more planners and allowing planners to add value and business acumen.
AI-based forecasting systems are a major upgrade to existing systems, in which merchandisers attempt to manually cluster and define seasonal patterns. Unsurprisingly, results are often disappointing and lead to out-of-stocks, lost sales and markdowns.
Store-level forecasting is a difficult endeavour with traditional systems, especially when retailers are dealing with millions of possible product and store combinations.
By alleviating the need for so much manual intervention and by accounting for so much information at any given time, artificial intelligence based forecasting can deliver far more accurate forecasts, particularly in replenishment.
Improving investment in inventory and supply chain. As tempting as it is to hold inventory buffers to ensure product availability, this increased investment into inventory eats into retailers’ balance sheets. In the face of ever more competition, retailers are all looking to hold less inventory to maximize their ROI, while still ensuring that they have sufficient inventory available to meet customer demand. Replenishment has evolved.
More accurate forecasts can help retailers optimize their inventory levels by minimising required inventory holdings (to free up working capital and reduce the need for markdowns to clear product later). This is while ensuring that retailers do not miss out on sales by not catering to customer demand. It is important that demand is accurate at the store level, as having excess inventory in one store but a shortage in another leads to a double jeopardy, whereby customer demand is not being served, leading to missed sales in one store, while a separate store is selling the same product at a discount.
Inventory imbalances will often occur right after seasonal peaks. One of the biggest reasons retailers continue to fall into this trap is because they are often forced to act retroactively with top up orders, transfers or priced drops. Providers such as Insite.ai that provide highly accurate demand forecasts at store level (accounting for demand seasonality) ahead of time can enable retailers to avoid these costs.
Advanced analytic forecasting combines historic sales with seasonality, trends, product lifecycle effects and statistically tested assumptions to improve accuracy. Increased forecast accuracy is the single biggest driver of direct savings, revenue and total return on investment. Improvements to cash flow will be realised in the following ways:
The DuPont Financial Performance Model is a framework for viewing how changes in sales, capital, and operating expenses impact return on shareholder value.
Improving forecast accuracy has a direct impact on all three of these factors: an accuracy increase of 10% can increase sales revenue by the same percentage, while decreasing inventory costs and freeing up working capital trapped in inventory.
The DuPont Model below demonstrates how improvements to revenue and savings driven by better forecasting convert to higher profits and decreased invested capital, which, in turn,create greater shareholder value.
Financial managers can perform more precise net revenue forecasting by calculating the expected cost of ordering products, price cuts, promotions and advertising and the gross revenue from the expected increasing sales. In addition to performing individual forecasts, those in merchandising can utilise what-if analysis on different forecasts with tweaked parameters, retailers can make more effective pricing, allocation and promotion decisions to maximize net revenue.
With respect to pricing, promotions and clearance, inaccurate forecasts mean retailers are often running reactive promotions without necessarily understanding the impact of each promotion on demand, or the true incremental profitability and cost of a promotion. Traditionally, retailers have also struggled to accurately forecast promotional efficacy. Having a highly accurate forecast that accounts for more influencing factors such as price, type of promotional offer, uptake etc will give greater certainty to any analysis.
The ability of artificial intelligence to identify trends and extrapolate from patterns in data can help retailers understand the relationships between product and price. For example, retailers can better identify like promotions and automate product affinity groupings to identify over- and under-performing products.
AI can also be used to provide more nuanced insights into each product. For example, artificial intelligence algorithms can look more holistically at the performance of a product / group of products and apply learnings across categories. This significantly cuts costs by reducing errors in inventory management and avoiding situations of excess or insufficient stock. The resulting boost in efficiency is accompanied by an increase in gross margins and top-line growth.
A typical retailer has a labyrinth of systems and departments: IT, Marketing, Customer Experience, Planning, Buyers, Suppliers, Warehouse and Distribution. Historically, the information known to one department such as merchandising is not well leveraged by other departments in supply chain. The opportunity for collaboration across a typical retail support office is huge. Engaging all departments in the AI / Forecasting process so that the business can make smarter, faster decisions, that deliver the necessary profitable revenue growth will immediately pay dividends! With machine learning’s ability to synthesise and understand vast amounts of disparate data, this will have a huge effect on forecasting accuracy and in turn, ROI.
An organizational structure that supports an integrated demand and supply management implies that decisions made by executives not only affect their own departments, but also the overall performance of the company. Each department empowered to make decisions that has a wider impact which serve to maximize the overall value created for the company and its customers.
For retailers, improving their overall forecasting capabilities is a critical need in meeting consumer demand. Staying in-stock with accurate and timely forecasts while minimizing supply chain costs and excess inventories has never been more important.