Promotional forecasting is complex. The knock-on challenges it creates along the supply chain – from ordering raw material and packaging, through to managing stock levels at depots and at individual stores – beg the question of why supermarkets run promotions in the first place?
The answer is quite straightforward. A buyer’s objective is to maximise profit through growing sales, reducing costs, and managing the percentage margin mix. Promotions help to increase sales, and can influence the mix of products sold.
Promotional mechanics are varied, and the different types of promotions available – from price cuts, to multi-buys and extra fill packs – can seem complicated. So beyond simply driving sales volume for the short-term, why else might a supplier or retailer decide to promote?
The main reasons are:
- To increase retailer market share
- To increase supplier or brand category share
- To grow total category sales, such as multi-buy meal deals.
- To increase consumption
- To encourage new product trial
- To increase average basket price, by encouraging consumers to trade into premium brands
- To manage the margin mix, by trading consumers into higher margin SKUs such as own-label equivalents
Different promotional mechanics can therefore be used to achieve these varied objectives – it is definitely not ‘one size fits all’.
Promotional volume uplifts can be affected by many factors and calculations with many complex variables, which makes forecasting volumes for retail promotions a complicated undertaking. With anything from the type and position of the promotion; what else is on promotion at the same time; time of year; and external influences (sporting occasions, weather & season; even TV advertising) all playing a role, predicting volumes in advance is a huge challenge.
With many products needing to be produced several weeks or months in advance of a promotion going live, the challenge is less about forecasting the total promotional volume, but more about managing fluctuating sales across the retailer store estate to keep 100% availability, and achieve 0% waste, consistently throughout the promotion period. In theory.
In reality, these fluctuating volumes present huge issues to the supply-chain. Suppliers need to produce more, depots need to hold more, more stock needs to be allocated to stores ahead of a promotion (so that they don’t immediately run out), and stores need to find more space on the shelf to put the stock. For short-life products such as produce, bread, and many others – this issue is compounded as you cannot simply ‘stock-up’.
And this is not just about the start of a promotion when volumes increase. It is also critical to plan the exit. Retailers do not want to be left with extra unsold stock at the end of a promotion, but also do not want to have empty shelves on their power-aisle for the last week.
Managing The End-to-End Process
So, what can be done to improve promotional forecasting?
The solution requires 2 things:
- accurate granular forecasts.
- the supplier to manufacture and deliver on time.
In the short-term, retailers and suppliers can make better use of data that already exists. Most large supermarkets provide access to daily sales and supply-chain data. Start by using the information that exists.
Secondly, it is important to look at promotions ‘in-flight’. Use tools that allow you to quickly check promotional performance (is it going to plan?); to check depot stocks daily – before they run out; and to be alerted to potential issues. These all allow you to be proactive in managing stock levels, and make life easier to communicate with your retail supply-chain colleagues. By reviewing data daily you can take action and minimise problems.
In the longer-term, retailers need to take responsibility for improving their forecasting methods. They have a wealth of historic data to use and also have the knowledge of what else is going to be promoted at the same time as your products. Artificial Intelligence and Machine Learning techniques are allowing constant improvements to forecasting methods and algorithms, and improvements in technology are allowing us to work with much larger data sets. This should ultimately help to provide what suppliers really need – accurate, daily, depot forecasts.
If it were easy for retailers to provide accurate daily product-by-depot forecasts then they would. Working out what will happen in the future can only be based on information we have about the past, mixed with known changes (weather, promotions etc.). Even with all of this to hand, there are still extreme cases, such as the CO2 shortage back in 2018, which drastically affect forecasting and are difficult to predict
By using best-practice visual analytics tools and techniques, supply-chain and sales data can be blended to highlight indicators for demand (such as sales spikes, low-depot stocks etc.), and provide an easily understood picture of recent and historic sales patterns. Having a system that is able to visualise daily data at the click of a button will help ensure you see dramatic improvement in forecasting accuracy; meaning that, over the coming few years, forecasting (and promotional forecasting in particular) should be the biggest area of improvement for retail supply-chains.