Empty shelves also caused by covid-confused ordering systems


Prague, October 5, 2021: One of the most difficult tasks any retailer faces is to ensure correct supply. Nowadays, it’s a matter of course that smart and automated technologies in the form of so-called inventory management and replenishment systems assist with this process, using algorithms based on statistical and probability models, predicting demand and ensuring that an adequate amount of goods is ordered. As a result of the turbulent sales development during the covid-19 pandemic, these systems have begun to fail, causing further complications in retail chains during these already complicated times.

Being able to ensure correct supply is a crucial skill for a retailer to have, since both surpluses and shortages of goods pose a big problem. While in the past retailers had to largely rely on their own judgment and a lot of handiwork in planning, nowadays there are sophisticated systems helping predict demand, taking into account past sales, modelling future trends using additional data inputs, and automatically ordering a substantial portion of the goods. The covid-19 pandemic, however, threw demand into such disarray that ordering systems haven’t been able to respond, resulting in significant issues.

Demand for certain goods dropped suddenly while raising sharply for other ones, making it impossible to predict sales even for goods where predictions used to be relatively accurate. In such conditions systems aren’t able to calibrate without human intervention. However, manual calibration requires not only an enormous amount of extra labour, but also highly precise information about the market which, at the moment, isn’t available. The entire problem is further exacerbated by the current issue with delayed orders from overseas. Even with 100% accurate predictions, delays are another parameter which can confuse the systems.

“Replenishment systems are based on historical data and trends. But the coronavirus pandemic caused huge fluctuations in demand, brought about by the closure and re-opening of stores. Moreover, covid-19 has dramatically changed customers’ needs and consumer behaviour, which also makes it impossible to rely on pre-pandemic data. This inability to correctly predict demand has inflicted huge problems and losses on retailers, be it in the form of lost income due to unavailable goods, or expensive liquidation of unsold items,” said Roland Džogan, CEO of Ydistri and a long-standing specialist in stocking and replenishment systems. 

Incorrect predictions can result both in empty shelves, which have a very negative effect on stores’ income or reputation among customers, and a large volume of unsold goods. The so-called deadstock is associated with a number of risks itself. For retailers, the accumulation of unsellable goods means blocked cash and a lack of space to exhibit other interesting products. Substantial discounts are usually chosen as a solution, though they have a negative economic effect and impact the brand’s image. In addition, not even enormous discounts guarantee sales. Fortunately, technology can assist with resolving the problem.

“Smart redistribution across the outlet chain is a solution to deadstock, i.e. goods lying around at an outlet where their sales potential is dropping. Most of existing deadstock is quite likely to sell at the original price if moved to a different outlet that is more suitable for some reason—differences in socio-demographics may make sale more likely, or the chain is simply overstocked at one place and understocked at another. This makes it necessary to use technology to find a more adequate final destination for items from the original outlet, one with a shortage of the relevant goods which sell well there,” said Džogan about the solution.

It's becoming increasingly clear that retail chains cannot rely on demand prediction solely for initial purchases but must instead be able to correctly use and regulate their stock after the fact. Surpluses or shortages of goods at individual outlets need to be balanced out, using smart redistribution. In Ydistri’s long-term experience, over 90% of products suggested by the algorithm for redistribution sell within two months of their stocking at a different outlet. This allows retail chains to not just address surpluses, but thanks to new technology also use their own unutilized stock across the outlet chain, refraining from needlessly ordering goods. 

“Prediction models used by ordering systems are top-notch but they can never be 100% accurate, much less in the current situation. Retail chains must therefore constantly regulate their stock level, taking into account actual sales. After all, supply chain issues have resulted in significant extensions of delivery dates, making the possibility of utilising pre-existing inventories even more important than ever before,” concluded Roland Džogan of Ydistri, a deadstock specialist.