How to assess the impression of visitors from the store - the case of "Perekrestok"

Customer loyalty guarantees the company's future earnings. Therefore, the buying experience becomes an arena for competition. How can a retailer manage it?

Retailers use in competition:

Prices. Price competition is getting worse: stores and chains are divided into price niches (premium/middle class/economy, etc.), and each has its own competing stores. In order not to lose margins, prices for some goods have to be made a little lower, and for some - a little higher than those of competitors. As a result, disloyal consumers, for example, either carefully monitor where buckwheat is cheaper and where they need to buy rice, or they stop feeling that prices are lower in some stores and higher in others.

Service. There has been a lot of talk about him in recent years. Many people in RUSSIA still have memories of trying on clothes while standing on cardboard at the clothing market. Now the basic level of service is almost everywhere:

the goods are arranged in the trading floor by category, which means that the client has the opportunity to quickly find and “feel” what they need; different payment options are available: cash, card, phone and even biometrics; The covid-19 pandemic and related lockdowns have forced all players in the market to provide customers with the opportunity to shop online and receive orders in a short time.

For a consumer to come back, you need to be memorable. If a consumer goes out of habit to a particular store, another reinforcement of this habit is needed; if faced with a choice of where to go, one must be on his list of those who first come to mind. Therefore, companies spend a lot of resources on designing and analyzing customer experiences.

How to analyze the impressions of store visitors? You can start with qualitative research methods. For example, build a client journey (CJM), drawing his emotional curve as he travels along this path. Or do a lot of JTBD interviews (Jobs To Be Done - an interview that helps you understand what motivates customers to buy your product), or even full-fledged ethnographic research. Such methods are good for getting hypotheses for further testing in quantitative ways, but they are difficult to collect statistics to say with confidence that you have not identified the "pain" and "pleasure" of only certain very specific people.

What to do? it is worth starting with quantitative methods, understanding and analyzing the data that you already have. This requires analysts with an inquisitive mind.

Let's analyze the case of the Perekrestok network (part of the X5 Retail Group)

Given: The results of an NPS study, which was conducted by calling customers of food chain stores. In particular, they were asked about what they liked and disliked during the visit. The customer comments were then divided into ten categories: assortment, product quality, visit convenience, and so on.

The task was to find out what and how influences the perception of customers of certain aspects of the experience of interacting with the store. In other words, it was necessary to determine the degree of influence of certain aspects of the store's work on customers.

Stage No. 1. We chose which specific aspects we want to analyze We formulated 58 ideas - triggers of customer emotions to digitize the ten main categories of customer impressions, for example, “the customer noticed an expired product”, “couldn’t find the product after rearranging the store - I don’t like it when products change places." For each trigger, we worked out a list of necessary data sources for digitizing these ideas; within each category, ideas were ranked according to complexity and expected effect. We selected 20 ideas that seem to be the most promising. We chose for digitization those of them that belong to the categories "staff" and "product quality". Stage number 2. Collected data from different sources

We chose a data depth of one calendar year. Data warehouse tables and downloads from different systems were used as initial sources, on the basis of which we compiled a showcase of attributes for analytics. It included: