According to Wikipedia, a decision tree is, in general terms, “a prediction model used in various fields ranging from artificial intelligence to economics”
In the world of FMCG, the term “decision tree” is applied to the process that leads the shopper from the planning of the purchase to the selection of a specific reference. Three (main) types of purchases can be distinguished:
- Planned purchase: “I will buy Coca-Cola Zero in cans”
- Planned purchase, but final decision in front of the shelf: “I will buy soft drinks or Cola or Coca-Cola and decide precisely what in store”
- Impulse buying: “I had not planned to buy soft drinks, but the availability and / or the offer generate an impulse and I buy 6 cans of Coca-Cola”.
Several surveys agree that 70% of purchases are NOT planned in terms of the final decision of brand and variety. It is therefore very important to understand how the shopper makes his decision to facilitate the purchasing act, hence the interest of modeling it in the form of a decision tree.
Historically, the hierarchical positioning of the different categories has often been based on their industrial characteristics, purchasing logic or exposure constraints. For example, for the shopper, there are no such categories as “frozen” or “chilled”, but pizzas or fish that can be bought in one or the other form. In the same way, “sliced cold cuts” is not a category: ham can be bought whole, cut, or packaged and pre-sliced, but what the customer is looking for is above all ham and not sausage.
While some categories such as kitchen paper or water are relatively simple in their offers, in complex categories, such as beauty or hair care, it is important to understand the buyer’s entry key into the category to facilitate:
- The planned act of purchase (“I can easily find the reference of the Pantene shampoo for dry hair that I usually use”)
- The understanding of the category and the choice of the appropriate product for shoppers who make their decision in front of shelves sometimes packed with several hundred SKUs divided into dozens of subfamilies.
A categorization based on shopper insights and sell-out data per store has a fundamental impact not only on the organization of the shelf (or planogram) and the definition of an effective assortment, but also in the very definition of categories and store organization.
Creating a universe dedicated to babies gathering everything that may be useful for new parents or implanting SKUs of certain categories in the shelves of complementary products (“cross merchandising” of alcoholic beverages and snacks, for example) are now common applications of this type of learnings.
How can a decision tree be defined? The most effective is to listen to the shopper to understand his decision-making process. For this we can use ad hoc studies based on focus groups or questionnaires, but the most interesting methodologies are based on observation in store (real or virtual) followed by questions to the shopper.
The analysis of sell-out data at check-out receipt level also makes it possible to define products cross-purchasing patterns and link them to shopping trips types and shoppers’ profiles which is particularly interesting to refine the decision tree.
If they are expensive for complex, these techniques make it possible to collect live information: time spent in front of the shelf, product manipulations or “eye tracking” of the linear with special glasses, but also explanation of the rationales (or not) of the purchase process directly from the mouth of the shopper.
For the manufacturer or distributor, researching and understanding the shopper’s purchasing decision tree means often changing from an industrial or commercial categorization that is surely easier to manage in terms of implementation. However, the integration of this key element of shopper marketing and category management opens up a thousand possibilities for improving shopper experience in store with very positive consequences on sales and loyalty.
Behind the decision tree hides a forest of potentials!
If you’re reading this, you’re one step away from selling more and better! Did you enjoy the read? Please, tell us more!
ThinkBI SLU processes personal data in order to address your request or enquiry with your consent. To learn more about how we manage your personal data and to exercise your rights, please refer to our privacy policy.