Since the introduction of electronic point of sale, retailers have had at their disposal an incredible amount of data. The challenge has been how to leverage this data to produce business value. Most retailers have already figured out a way to consolidate and aggregate their data to understand the basics of the business: what are they selling, how many units are moving and the sales amount. However, few have ventured far enough to analyze the information at its lowest level of granularity: the market basket transaction. The main reason for this is, perhaps, the preconceived notion that looking at data at this level of granularity is expensive and has limited business value. This article will explore the business value of market basket analysis through real scenarios, outlining along the way why the users don't need a strong statistics background to understand it and benefit from it.
Market basket analysis, or MBA for short, is the process of analyzing transaction-level data to drive business value. At this level of detail, the information is very useful as it provides the business users with direct visibility into the market basket of each of the customers who shopped at their store.
The data becomes a window into the events as they happened, understanding not only the quantity of the items that were purchased in that particular basket, but how these items were bought in conjunction with each other. In turn, this capability enables advanced analytics such as:
- Item affinity: Defines the likelihood of two (or more) items being purchased together.
- Identification of driver items: Enables the identification of the items that drive people to the store that always need to be in stock.
- Trip classification: Analyzes the content of the basket and classifies the shopping trip into a category: weekly grocery trip, special occasion, etc.
- Store-to-store comparison: Understanding the number of baskets allows any metric to be divided by the total number of baskets, effectively creating a convenient and easy way to compare stores with different characteristics (units sold per customer, revenue per transaction, number of items per basket, etc.).
Affinity Analysis
As previously discussed, affinity analysis is used to determine the likelihood that a set of items will be bought together is. There are natural product affinities in the market place. For example, it is very typical for people who buy hamburger patties to buy hamburger rolls, as well as ketchup, mustard, tomatoes and other items that make up the burger experience.
While there are some product affinities that might seem trivial, there are some affinities that are not very obvious. The classic example is diapers and beer as husbands who are sent to the store for diapers cannot pass the opportunity to buy beer to compensate for the emotional stress of being seen with a diaper bag.
Another classic example is toothpaste and tuna. It seems that people who eat tuna are more prone to brush their teeth right after finishing their meal. So, why it is important for retailers to get a good grasp of the product affinities? This information is critical to appropriately plan for promotions because reducing the price on some items may cause a spike on related high-affinity items without the need to further promote these related items.
A good understanding of the affinity of the items might lead to customer friendly planograms by re-accommodating the products in the store. Already a number of hardware stores stock items by "project" along with their regular categories. This facilitates things for beginners who are trying to do home improvements project themselves but are daunted by the thought of knowing what items to buy and where to find them in the store.
Indentification of Driver Items
Identifying the items that drive the traffic to the store is always a challenge. It is becoming increasingly difficult to strike the right balance between product depth and breadth regarding inventory. With only a couple of units on the shelf, the probability of running out of stock is very high. If a particular customer was drawn to the store for this particular item and there are none in stock, it is possible that this customer leaves the store immediately or makes a mental note not to come back in the future.
Identifying the driver items will also help to distinguish the main item from the related items when doing product affinity. For example, discounting the burger patties might increase the sales of rolls, veggies and ketchup, but the reverse will not hold true as discounting the ketchup will not bring additional sales.
Unlike filler items, shoppers are usually very brand sensitive when buying the driver items. If retailers are planning to introduce private labels, this information will be critical to determine the initial price point and the target market for these private items, otherwise they run the risk of one failed retailer who wanted to displace the leading brand of detergents with a product of "similar" quality at the same price point. Needless to say the results were a disaster; the national brand did not loose any market share and this retailer was eventually forced to severely discount their private label. It was not until someone realized that they had positioned the product for the wrong market and changed the market strategy to position the product for consumers with low and moderate incomes that the private label started moving at a decent pace.
Trip Classification
The concept of basket or trip classification is not new, but it has received renewed interest over the last couple of years as retailers struggle to determine the format for their new stores. There is no magic behind trip classification. It requires a real understanding of how to properly classify the contents of the basket to profile the shopping trip. Taking into consideration variables such as total basket value, number of items, number of category A vs. category B items, rules can be derived that help map each of the baskets to a previously defined classification.
Understanding what kind of shopping trips a customer performs at a particular store at a particular time is critical for planning purposes. This data provides a unique window into what is happening at the store and enables advanced applications such as labor scheduling, product readiness and even temporary layout changes.
Let's take for example a grocery store, given that most of the grocery items have a short shelf life - it is important for the store manager to understand when the items are going to be consumed to have enough product in stock. With some preliminary analysis he learns that not many people will buy beer during the early part of the week As a result, he calls the beer dispatcher and asks him to stop deliveries on Monday and Tuesday and come twice a day during Saturdays (when he is always running out).
Other retailers are using MBA to understand their customers shopping behavior during a particular day of the week and at various times from morning to afternoon. A particular hardware store used MBA to analyze the demand on certain consumer departments and found out that on certain days of the week, some employees were sitting idle while the contractor department was short staffed. By implementing on-demand systems (e.g., call buttons), this retailer was able to reduce labor costs by redirecting the employees to where they were needed and keep an electronic eye for any customers outside of the pattern.
As a Basis for Store-to-Store Comparison
This is a very simple but effective use of MBA - count the total number of baskets for each of the stores and use metrics that can be normalized so stores can be compared to each other.
Let's take, for example, a retailer that has big stores and small stores; the big stores have more employees, more customers and more sales than the smaller ones. One day this retailer decides to create a contest across the whole chain where all the stores will compete against each other on dollar sales and volume. The store managers for the small stores do not want to play ball, arguing that they will never be able to compete with their big brothers. An analyst reviews this concern and finds it to be valid.
Fearing a showstopper for the contest, the analyst remembered that he read an article about MBA where the author suggested dividing store metrics by the total number of customers per store. This metric could be used to compare results from store to store independently of the size. The analyst explained the idea to the disgruntled store managers with a practical example: Assume store A sold $540 worth of product x, and store B only sold $188. At first glance it seems that store A did three times better than store B. However, once you factor in MBA - you discover that store A had 400 baskets (customers) while store B only had 80 customers. This changes things. If you divide the $540 for store A by 400, you get $1.35 per basket; store B divides $188 by 80 for $2.25 per basket. Store B is getting a full dollar more per customer than store A. The store managers are not disgruntled anymore; corporate found a way for all the stores to compete on the same basis so every customer matters.
MBA is indeed a great capability that can revolutionize the retail business as we now know it. MBA provides an excellent way to get to know the customer and understand the different behaviors. This insight, in turn, can be leveraged to provide better assortment, design a better planogram and devise more attractive promotions that can lead to more traffic and profits.
