Retailers must embrace online analytics to understand their big data, improve their forecasting and grow sales, claims Michael Feindt, founder of Blue Yonder, a leading provider of predictive analytics
As consumer spending confidence falls, retail margins continue to decline, making today’s environment increasingly difficult for retailers.
The KPMG/Ipsos Retail Think Tank (RTT) – a panel of retail industry commentators that provide authoritative and independent views on issues affecting the sector – recently reported an even bleaker outlook for retailers, with the RTT Retail Health Index expected to fall to an all-time low of 77 this summer.
As a result, attracting new customers and maintaining existing ones is even more critical. Many retailers opt for excessive promotions on lower-priced goods to entice customers and sell off surplus stock. However, while this approach can shift sales in the short term, it isn’t sustainable over the longer term. Every retailer ultimately needs to maximise their revenue by optimising their product forecasting process to ensure stores are well-stocked with the right amount and right category of products.
Effective forecasting avoids retailers being left with surplus goods and having to sell them off cheaply or shelves being left with gaping holes, disappointing customers and damaging the stores reputation.
The sheer number of products that are carried and spread across numerous stores alone is enough to reveal what a colossal task this is.
Manual ordering procedures are limited and largely inaccurate. Regardless of how well planners perform, it is impossible to consider every single factor that affects sales when making decisions.
On top of this, every staff member also has their own individual perspective and bias – which can lead to very different estimations of demand.
With modern sales forecast systems, predictions are different. Clever mathematical algorithms based on objective data that has originally been developed for high tech analysis of big data collected by large international research organisations like CERN, deliver results that are based solely on actual demand: unerring and unaffected by any random factors that inevitably exist.
Whereas classical automated goods procurement already proves to be a powerful tool, modern methods enable retailers to avoid unnecessary costs and benefit from the full extent of purchasing power.
ERP software as a solid basis
It goes without saying business software – or at the very least an Enterprise Resource Planning (ERP) system – is a must-have for every retailer. Yet this type of software is unable to provide retailers with reliable recommendations for their goods procurement.
Despite the fact business intelligence tools are usually shipped along with inventory management systems and ERP solutions, these tools are not designed to look into the future. Instead, their role is to look back and analyse periods that have already happened, feeding the results and insights obtained into strategic planning.
They use ‘dashboards’ to allow users to get a picture of the current situation, however it is almost impossible to make reliable predictions using these traditional methods for every article in every store every day.
Predictive analytics can balance the equation
Dynamic markets such as the retail food industry demand tools that take numerous different influential factors into account.
Forecasts for product sales and predictions for customer and supplier behaviour can only be achieved by using software solutions that look to the future.
Against this backdrop, it is important to note in today’s fresh foods industry up to 40% of the purchase order proposals generated using traditional automated goods procurement systems have to be corrected manually when it comes to goods with a critical best-before date (dairy products, meat, fish, baked goods, etc.).
Conventional forecasts are simply not effective where complex sales planning is concerned. However, the situation is altogether different when a company chooses to enhance its ERP system by adding highly-scalable forecast software.
After all, modern predictive analytics help retailers to achieve empirically accurate forecasts in the form of complete probability distributions, therefore providing an estimate of the most probable demand figure as well as the uncertainty of this number.
They are then perfectly positioned to be able to make thousands of optimal day-to-day decisions – again automatically using clever mathematics and with immense accuracy. The accuracy of the predictions can be monitored every day – and this is very comforting and convincing.
Why automated goods procurement pays off
By introducing predictive analytics, purchasing departments can really stand out from the competition. Using highly reliable forecasts as a basis, companies are able to plan for the actual demand. From perishable, long-life, and fast-moving products to goods that are on offer and less popular products – sales across the entire product range are forecast with the same impressive level of quality. So what kind of figures are we dealing with exactly?
Let us take a typical UK supermarket chain as an example:
- 20,000 products
- 1,000 stores throughout Europe
- 20 million product items to be forecast every day
- Annual turnover of £8bn
When quantities as big as this are involved, even small improvements in sales planning have a significant effect on the turnover and write-offs achieved. As each material planner is usually responsible for a large number of products, there is usually little time for performing a detailed analysis. Especially when there are so many factors that have to be considered at each individual product level.
The bottom line: perfect planning
Having a detailed overall view enables retailers to centralise their procurement activities. Synergetic effects and bundling purchasing volumes allow them to reduce procurement costs. It becomes far easier to establish the best order value and time to order from a financial perspective. In doing so, they are able to avoid both write-offs and out-of-stock situations. Since they are now able to know purchase quantities much better in advance, marking down products is much less required.
As the order quantity for every single store is identified automatically and passed on to the retailer, this gives them the assurance needed to be able to run the store with competence. There is also a huge time-saving element for the retailers in question, since they no longer have to check the filling levels of their fresh counters or shelves. Instead, retailers are able to focus on their actual duties: creating an inviting shop floor, managing their personnel, and improving customer service.
With predictive analytics, retailers can effectively analyse their big data and interpret the results in a way that will allow them to make insightful planning decisions and maximise turnover. By increasing the accuracy of sales forecasting by up to 40% and the automatisation level to more than 99.8%, retailers can reduce surplus stock and out-of-stock-situations simultaneously and significantly increase in efficiency and profitability.
Blue Yonder contact details:
Tel: +49 (0)721 383 117 0
(A Retail Times’ sponsored article)