The Multichannel Purchase Index (MPI) And Multichannel Forensics

Let's take a look at three examples of the interplay between the Multichannel Purchase Index (MPI) and Multichannel Forensics.


Example #1: A brand utilizes Online and Retail channels.



This is a typical situation, one where customers transfer from the online channel to the retail channel. Retail customers, however, stay in the retail channel, not being as willing to shop online.

Because the online channel transfers customers to retail, multichannel marketing is important. The Multichannel Purchase Index (MPI) is 1.229, suggesting that this brand is much more "multichannel" than the average brand.


Example #2: Another Online and Retail example.



In this example there is very little interplay between the online and retail channels. The online channel is in equilibrium mode, sending a small number of customers to retail. Retail is in isolation mode, not sending many customers to the online channel. Predictably, only 4.7% of the file is deemed "multichannel", yielding a Multichannel Purchase Index (MPI) of 0.844, below average.

When both channels are in isolation mode, you'll see MPIs between 0.400 and 0.750.


Example #3: Catalog and Online channels.



In this case, each channel is in equilibrium. Each channel passes customers to the other channel. Customers are fully immersed in the "multichannel" experience.

Even though the average purchase frequency is low (1.82), and the percentage of the customer file with multichannel status is only 9.8%, the MPI is very high, 2.522.

This is the beauty of the MPI. It is able to simultaneously adjust for purchase frequency and multichannel status, telling you what is really happening.

Multichannel Forensics and the Multichannel Purchase Index are highly correlated. When channels are in isolation mode with each other, the MPI will be low (<> 1).

Since is it so easy to compute the MPI for your two-channel situation, why not give it a try?

(PERCENTAGE OF 12 MONTH FILE BUYING FROM MULTIPLE CHANNELS LAST YEAR) divided by (0.00559 + 0.05536 * LN (AVERAGE PURCHASE FREQUENCY LAST YEAR)).