Top challenges in demand forecasting

Jun 16, 2015 4 min

Forecasting is difficult and, as generations of soothsayers and pundits have discovered, forecasting the future is much harder than forecasting the past.

Seriously though, many retailers find forecasting challenging but they prioritize it because it’s generally accepted that better demand forecasting helps improve cost effectiveness and availability in the supply chain.

But what is it that retailers find especially difficult when it comes to forecasting? We commissioned Martec International to carry out a broad study to find out which forecasting challenges UK retailers find most difficult. Martec found that retailers listed the hardest challenges as follows:

  • Forecasting new product introductions
  • Coping with challenges related to sales volumes
  • Forecasting promotions and promotional lift

Read the latest Martec International 2018 report here: ‘State of the Retail Supply Chain 2018

The Challenge of New Products

Forecasting new product introductions is becoming increasingly important as product life-cycle shortens and assortment turnover increases. It poses a particular challenge in speciality retail sectors such as electronics, fashion, books and gardening, where new product introductions and heavily refreshed seasonal assortments account for the bulk of sales. One of the most extreme is the book industry where, as a rule, more than 90% of items sold are new that year. So, if you do not get the forecasting of product introductions right, you can forecast only a small fraction of your business.

Good examples of RELEX customers who have cracked the new product forecasting conundrum are two European book retail chains, both market leaders in their respective countries (you can read the case of Suomalainen Kirjakauppa here). Working with RELEX they’ve implemented a process where forecasting also for new items is carried out automatically – it is done by searching the historical assortment for items with similarities to those being introduced and building a forecast from that. The results in this high assortment turnover and highly-seasonal business have been great – availability has been increased several percentage points to near perfect and, simultaneously, inventory values and mark-downs have been reduced considerably. Additionally the whole business of managing introductions is far simpler as staff do not have to spend huge amounts of time hunting down reference products. With a large assortment that process is, in practice, almost always very inaccurate as there are simply too many new products to map them correctly.

How to Tackle Variance in Sales Volume?

Coping with changes in sales volumes is pretty much inevitable in modern retail environments. This should be a standard feature of most retail forecasting processes. The key to tackling these change situations effectively is to bring together statistical forecasting with human insight. Typically humans react too quickly to changes and give too much weight to recent events even though they are probably well within the bounds of normal statistical distribution and best interpreted as such. However, sometimes a change can indeed be traced to wider market conditions and that’s where our ability to infer a causal link from events actually serves us well; it’s where humans still outperform computers.

Having a solution that can map these exception situations and present the data for effective manual checking and commenting enables supply chain teams to react far more quickly. A good system gives a dispassionate overview of the situation, directly comparing recent and historical changes, and that context means that managers can quickly make a judgement whether a change is ‘just one of those things’ or linked to specific external or internal factors.

The Importance of Promotions

Forecasting promotion demand is one of retail’s big challenges not least because of its growing importance in driving sales. For example in the UK grocery market as much as a third of all sales are promotion driven. So if you’re not forecasting promotions well, or they are not included within your solution’s forecasting process – one of the most critical aspects of your business is excluded from your core processes.

However, it is entirely feasible to calculate promotion demand forecasts using your past sales data with a considerable degree of accuracy – especially if you are willing to take on the task of collating data from promotions such as display details (such as what fixtures were used and where in the store the promoted items were displayed), promotion related advertising, the promotion price and the nature of the offer (two-for-one, 25% off etc).

A good example of this is the case of an UK based food retailer, who got significant benefits by implementing our solution and taking on the hard graft of capturing its historical data. As a result the supermarket boosted forecast accuracy which helped both improve promotion item availability and cut down on excess stock which in turn made room for the new promotions.

As the Martec research makes clear, many retailers still run promotions forecasting and replenishment as largely separate processes without dedicated promotions forecasting support from their systems. However our experience suggests that it’s an area ripe for improvement through good processes and good technology.

So new product introductions, handling demand changes and managing promotions forecasting are the three most cited challenges by UK retailers. What are yours? Do you already have the answers or might it be worth exploring the latest forecasting technology and techniques?

State of the Retail Supply Chain 2018

Read more about the challenges retailers face today and how they plan to overcome them in the report: ‘State of the Retail Supply Chain 2018

Written by

Mikko Kärkkäinen

Co-founder & Group Chief Executive Officer