0Market Basket Analysis (MBA):
Market Basket Analysis (MBA), also known as affinity analysis, is a technique to identify items likely to be purchased together. The introduction of electronic point of sale systems has led to collection of large amount of data. Simple, yet powerful - MBA is an inexpensive technique to identify cross-sell opportunities mostly in CPG industries. A 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.
Market Basket Analysis (MBA) is a data mining technique which is widely used in the consumer package goods (CPG) industry to identify which items are purchased together. The classic example of MBA is diapers and beer: "An apocryphal early illustrative example for this was when one super market chain discovered in its analysis that customers that bought diapers often bought beer as well, have put the diapers close to beer coolers, and their sales increased dramatically. Although this urban legend is only an example that professors use to illustrate the concept to students, the explanation of this imaginary phenomenon might be that fathers that are sent out to buy diapers often buy a beer as well, as a reward."
The example may or may not be true, but it illustrates the point of MBA.
The analysis can be applied in various ways:
- Develop combo offers based on products sold together
- Organize and place associated products/categories nearby inside a store
- Determine the layout of the catalog of an ecommerce site
- Control inventory based on product demands and what products sell together
- Credit card transactions: items purchased by credit card give insight into other products the customer is likely to purchase.
- Supermarket purchases: common combinations of products can be used to inform product placement on supermarket shelves.
- Telecommunication product purchases: commonly associated options (call waiting, caller display, etc) help determine how to structure product bundles which maximize revenue
- Banking services: the patterns of services used by retail customers are used to identify other services they may wish to purchase.
- Insurance claims: unusual combinations of insurance claims can be a sign of fraud.
- Medical patient histories: certain combinations of conditions can indicate increased risk of various complications.
Three common terminologies are used a lot in the market basket analysis, which is mostly based on classical definition of probability:
i. Support, Confidence and Lift:
There are several measures used to understand various aspects of associated products. Let's understand the measures with the help of an example. In a store, there are 1000 transactions overall. Item A appears in 80 transactions and Item B occurs in 100 transactions. Items A and B appear in 20 transactions together.
a. Support: The simplest one, Support is the ratio of number of times two or more items occur together to the total number of transactions. Support of A = P(A) = 80/1000 = 8% and Support of B = P(B) = 100/1000 = 10%.
Support of a product or product bundle indicates the popularity of the product or product bundle in the transaction set. Higher the support, more popular is the product or product bundle. This measure can help in identifying driver of traffic to the store. Hence, if Barbie dolls have a higher support then they can be appropriately priced to entice traffic to a store.
b. Confidence is a conditional probability that a randomly selected transaction will include Item A given Item B. Confidence of A = P(A|B) = 20/100 = 20%.
i.e. to measure the probability of bundling/selling propensity of a product A when it is bundled with B.
Confidence can be used for product placement strategy and increasing profitability. Place high-margin items with associated high selling (driver) items. If Market Basket Analysis indicates that customers who bought high selling Barbie dolls also bought high-margin candies, then candies should be placed near Barbie dolls.
c. Lift can be expressed as the ratio of the probability of Items A and B occurring together to the multiple of the two individual probabilities for Item A and Item B. Lift = P(A,B) / P(A).P(B) = (20/1000)/((80/1000)*(100/1000)) = 2.5.
Lift indicates the strength of an association rule over the random co-occurrence of Item A and Item B, given their individual support. Lift provides information about the change in probability of Item A in presence of Item B. Lift values greater than 1.0 indicate that transactions containing Item B tend to contain Item A more often than transactions that do not contain Item B.
In order to gain better insights, differentiate Market Basket Analysis based on:
- Weekend vs weekday sales.
- Month beginning vs month-end sales.
- Different seasons of the year.
- Different stores.
- Different customer profiles.
Based on the content and value of the basket, it is useful to classify the trip. Variables such as total basket value, number of items, number of category X vs. category Y items, help in developing rules to 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.
Not Just Retail:
Although Market Basket Analysis reminds pictures of shopping carts and supermarket shoppers, there are many other areas in which it can be applied. These include:
For a financial services company:
- Analysis of credit and debit card purchases.
- Analysis of cheque payments made.
- Analysis of services/products taken e.g. a customer who has taken executive credit card is also likely to take personal loan of $5,000 or less.
For a telecom operator:
- Analysis of telephone calling patterns.
- Analysis of value-add services taken together. Rather than considering services taken together at a point in time, it could be services taken over a period of, let's say, six months.
A predictive market basket analysis can be used to identify sets of products/services purchased (or events) that generally occur in sequence — something of interest to direct marketers, criminologists and many others.
Advanced Market Basket Analysis 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 promotions that can lead to more traffic and profits.
