Artificial intelligence in Retail – What can Wholesale Learn from Retail?

In this post, I will describe a real-world example of a case in the realm of artificial intelligence in B2C retail. You will learn about the opportunities and challenges that were encountered when using AI systems.

Of course, it is particularly interesting to see which factors can be transferred to B2B wholesale.

Artificial intelligence in Retail – Practical Example Based on a Shoe Store.

A small, owner-operated shoe store specializes in the sale of sports shoes for running. This includes conducting motion analysis on the treadmill with corresponding measurement and analysis of the rolling behavior while running, including corresponding video recordings. Malpositions are also detected here. The trained sales staff are aware of the differences between running on a treadmill and running on asphalt or natural ground and include them in the analysis. This enables brand-independent advice to be given to the customer.

From the customer’s side – if he agrees – the following data is collected:

– Address
– Telephone numbers
– E-mail address
– Shoe size
– Shoe width
– Sold shoe
– Other sales (e.g. clothing)
– Date of purchase
– Various usage data (e.g. type of sport, distance run, frequency of use, etc.)
– Malpositions in the foot/joint, etc.)

This data is then recorded in the ERP system. In the event of future visits by the customer, the data is available to the seller. There is no further use of the data.

Opportunities and Limitations of AI Systems in Retail

The use of AI with predictive analytics offers numerous possibilities for lifting this treasure trove of data. This is especially with regards to the fact that running shoes are actually used for running have to be renewed regularly. This is not a factor for customers who buy the shoes purely out of fashion considerations. The behavior of customers who buy the shoes as a fashion accessory is also much more difficult to predict.

Newsletters and advertising via email and/or SMS can be used to retain customers. Using the data, the AI can make suggestions for the next pair of running shoes both in advertising and when the customer visits the store. This makes advertising much more personalized. Targeted actions for cross-selling and up-selling are also possible.

Thanks to predictive analytics, the purchase date for the next pair of shoes can be narrowed down fairly precisely. Advertising activities can therefore be limited to this period. This not only increases the hit rate, but also reduces costs at the same time. Fewer but better targeted ads also mean that advertising is no longer seen as a nuisance.

Another advantage arises in purchasing. As a businessman once learned, “in purchasing lies profit”. Since, thanks to predictive analytics, it is possible to predict quite precisely what will be sold, it is also possible to derive what needs to be purchased from the figures, plus what the owner thinks he can sell in addition or needs to hold in reserve. In this way, you can counteract a supply shortage – or, the exact opposite: less merchandise that has to be sold at special prices to clear the warehouse. The bottom line should therefore improve.

Even without the collection of personal data that takes place here, an AI system with predictive analytics makes perfect sense in other areas of retail. Using past sales data, the system is able to make a fairly accurate prediction about expected sales. Here, too, the figures for purchasing can then be derived.

In fashion, however, an AI system cannot replace the instinct for fashion that the owner should have.

Another limitation is the irrational behavior of customers or external influences that cannot be predicted. For the latter, I only recall the pandemic. Neither the pandemic nor the restrictions on trade imposed by the government were foreseeable.

Regarding irrational behavior, one can cite, for example, influencers or persons of public interest. If an influencer recommends a pair of sneakers, for example, and these suddenly become “must haves” among thousands of followers, this behavior is not predictable for retailers either by humans or by AI systems.

Transfer to B2B Wholesale

Many of the points that apply to retail naturally also apply to wholesale and can therefore be transferred.

However, there is One very important difference. Should the wholesaler also use an AI system with predictive analytics, the data basis will differ significantly from that of the retailer.

The wholesaler’s data is based on previous orders (historical sales data). The retailer’s data (in the case above) is based on information provided by customers – supplemented by the number of units that the owner additionally expects to be able to sell. But both data bases lead to the goal of predicting expected sales and customer behavior through predictive analytics.

It is also true for the wholesaler that the sales team has a strong instinct through their customer experience, which should not be underestimated. No AI – which is, after all, a computer program in the final analysis – can have this instinct at its disposal. Humans will always be at the helm.

And the last point, which can also be transferred from retail to wholesale, concerns unpredictable events and irrational buying behavior. Neither can be predicted by the best AI system.

 
CALCULATE NOW THE ROI OF QYMATIX PREDICTIVE SALES SOFTWARE
 

Artificial Intelligence in Retail – Conclusion

An AI system can offer significant advantages in retail and wholesale, both in sales and purchasing. However, this is limited to the “standard assortment” and a highly specialized assortment in a small target group. Unpredictable events cannot be recognized and taken into account by AI either. Nor can any AI replace the instinct or feeling for fashion trends. These are areas that only humans can continue to provide.

I WANT PREDICTIVE ANALYTICS FOR B2B SALES.