Analysis and optimization of the assortment thanks to machine learning

The stakes linked to the assortment are considerable for the stores, because the quality of the assortment depends on the results of a store. Indeed, the size of the assortment, its price and its availability will have a direct impact on the sales and on the turnover. The optimization of the assortment must take into account a certain number of elements in order to meet the needs of the consumers and to reach the objectives of a sales area.

So how do you analyze and optimize your assortment? What techniques should be used to methodically analyze the assortment? Which solutions allow to optimize an assortment according to the seasonality or the local economic context? We propose you to discover how artificial intelligence has become the only tool allowing to analyze and optimize efficiently an assortment.

Four steps to analyze your assortment

The analysis of an assortment is a complex process that can be decisive for its optimization. Indeed, a bad analysis can distort the optimization of an assortment and can degrade its performance. So, in order to safely analyze your assortment, it is necessary to follow four steps.

Step 1: Setting goals

Before analyzing or optimizing the assortment, it is important to set objectives. Indeed, it is by determining precisely the objectives of the assortment that it will be possible to set up concrete and effective actions. The objectives of an assortment can be multiple:

gain market share;
develop turnover;
implement a new brand;
improve customer satisfaction;
etc.

Step 2: Collecting data

To analyze an assortment, companies need data. They may differ from store to store, but overall, the analyses will be based on:

sales receipts;
abandoned baskets;
cancelled orders;
the socio-economic context;
the offer proposed by the competition.

It is therefore a question of internal and external data, because the analysis of the assortment can only be done thanks to the combination of all this information.

For example, during the coronavirus health crisis, we observed changes in buyer behavior. Online purchases and drive-throughs were favored. Similarly, sales of certain products were halved, while others increased fivefold.

The weather also plays a determining role in customer behavior. Some products sell better in summer than in winter. But during the summer, a cool, rainy day will change the choice of customers, who will no longer automatically turn to the favorite summer products such as ice cream, cold drinks, etc.

An optimized assortment is therefore an assortment that takes into account the needs of the consumer, but also his psychology and the situation in which he finds himself: his purchasing power, his geographical situation, his family situation, his health situation, etc.

Step 3: Implementing the data

Once the internal and external data have been collected, it is important to use them well. The question of how to do this automatically arises: how to sort all this information and how to use it efficiently and quickly?

In order to sort and analyze all the information that can improve the assortment, companies use artificial intelligence. Indeed, only artificial intelligence is able to process such a large amount of data simultaneously. This is machine learning.

The collected data is integrated into a predictive intelligence platform. The machine learning algorithms perform calculations. The results of these calculations can then be integrated directly into the company's tools. Similarly, new data can be added on a regular or ad hoc basis. The machine learning results are then updated according to the latest information.

Step 4: identifying potential assortment problems

The performance of an assortment can be degraded due to:

the prices;
product quality;
stock management;
competition;
merchandising.

Thus, the assortment may be too poor compared to the competition, it may also be too expensive or of lower quality. Furthermore, the turnover rate of certain products is too low and can degrade the overall performance of the assortment. Finally, a position too low on the shelf can reduce the visibility of an article or a range of articles. The profitability of the range is then significantly deteriorated.

Assortment optimization with Kepner's 6R's

In 1963, Charles Kepner stated that the role of assortment is to offer the right product, at the right time, in the right place, in the right quantity and to the right customer. Later, he would add "the right information" to this list. According to Kepner, a product presented with the right information is an item that the consumer will easily find in the store and be able to buy on their own.

Offer the right product in a particular sales area

An assortment should be made up of the right products. So what is a " right product "? In order to answer this question, we need to add up all of the "R's" that Kepner stated. But to begin with, we can say that a good item is an item that meets a need and an objective. It is therefore an item that has a place on the shelf in a sales area.

Addressing the right customer through content and merchandising

All consumers are different and will not all buy the same categories of items. Mass retail, online, retail and convenience stores are implementing marketing strategies to address the right customer. The segmentation of the database allows to address targeted and personalized marketing offers.

In stores, merchandising can be used to attract the attention of certain buyers. Indeed, some shoppers will be more sensitive to certain promotional messages. The way in which customers are addressed is one of the key elements in perfecting the assortment and highlighting a certain category of items.

Find the right time to promote your offer

Seasonality, the economic context, calendar holidays, vacation eves, winter weeks and spring break are all times of the year that trigger certain purchases. Similarly, in a store, buyers do not behave in the same way on a Monday or a Saturday, at the opening in the morning or at the end of the day.

The assortment must therefore be thought out according to each time of day, each moment of the week and each occasion. Companies have to find the right time to offer certain items to certain buyers. Otherwise, they run the risk of not making the number of sales they had planned. Merchandising adapts to these different times of the year by highlighting certain product categories on the shelves.

Providing the product in the right quantity

An efficient assortment must take into account the quantities and the stock management. Indeed, an article, or a category of articles, must not be out of stock or overstocked. The first case would cause customer dissatisfaction and a decrease in the number of sales, while the second would lead to additional storage costs.

In addition, storage techniques for certain sensitive products can generate significant financial costs. In the pharmaceutical or food industry, certain storage rules must be respected. The stock level of the references must correspond to the sales forecasts in order not to reduce the company's performance.

Set the right price for each product

In mass distribution, convenience stores, retail and online stores, all products must be priced right. The price of an item must include the notion of profitability and objective. Indeed, some products can be offered at very competitive prices if they are, for example, appeal products. On the contrary, other products can have a higher price depending on the competition.

In short, the right price is not necessarily the lowest price. It is rather a price that takes into account different parameters such as the purchase cost, the competition's offer, the stock level, etc.

Add the right information for each item

To sell an item successfully, a brand's marketing department works on the packaging and on the presentation of a range. In stores, in the middle of the shelves, the colors of the packaging can capture the consumer's attention. Within the same shelf, thanks to POS, brands can also add information.

This information can take the form of a usage advice, an exhaustive list of components, etc. This information must be taken into account in the assortment or merchandising plan.

Forecasting: an efficient method to analyze and optimize your assortment

Although Kepner's 6R's can be used to build an optimized assortment in a store, they are based on a lot of information. This information is all essential and must be taken into account to analyze and optimize the assortment of a sales space.

The predictive intelligence platform allows companies to perform a large number of simulations based on a large number of data. Thanks to the forecasts, companies are able to make decisions more quickly. They know how to react to all types of situations. Moreover, since these forecasts are based on machine learning, they are reliable and accurate.

In short, assortment analysis and optimization requires the mastery of an innovative and powerful tool: the predictive intelligence platform. Thanks to it, consumers will make sales that correspond to their needs and companies will reach their objectives.