Customer Value in the Brave New World of Big Data and AI

AI and Big Data are now everyday words in the business lexicon as organizations strive for what is its actual impact on their business. Mobile technology has accelerated the use of these technologies where the customer now controls the marketing paradigm. The notion of customer experience in many instances begins with the customer searching for a product and service.

Entire industries are being disrupted with the insurance sector as one example. Rather than the typical scenario of an insurance representative or broker proactively marketing a variety of products or services from multiple companies, the new face of the broker or insurance rep is now digital. These digital platforms allow the consumer to be proactive in selecting the specific product or service that meets their insurance needs and at a very efficient price. 

This impact of customer empowerment is now a fact for all industries and organizations as technology continues to empower more companies and start-ups in being able to provide a more seamless and to some extent a more enjoyable customer experience. But what is the common theme that is a key outcome of this increased customer experience: DATA. Analytics and data science are no longer “nice to have” skills but requirements for corporate success in this new paradigm. The more recent significant successes of AI have pushed organizations to utilize this technology at almost every touch point of the customer journey.

But are we as analytics and data science practitioners too immersed in the trees amidst all this technology rather than adopting the “forest” perspective.

What do I mean by this? One fundamental question which was true thirty years ago is even more prevalent today with all this data. Do organizations know who their best customers are or MVC (most valuable customers)? The answer for the most part is no. This may seem surprising given how everyone views its importance. But let us think through what exactly this means. Technology is not the panacea here but rather a facilitator in the calculation of customer value and the determination of a best customer or MVC.  The actual process of calculating customer value and determining best customers is a very labour-intensive process.

Defining value of a customer will vary from industry to industry and in many cases from organization to organization. The definitions can be quite complex especially for the banks where customers may have multiple relationships with the bank (ATM, credit card, mortgages, wealth, etc.). This also does not look at the cost to service that customer which essentially looks at every touchpoint that the customer has with the bank. Keep in mind, we are not even assessing the level of risk or the inability of the customer to pay for a given service or product which represents another cost component to the value equation.

Although the generation of customer revenue might be simpler in many other industry sectors, the notion of servicing that customer and customer risk still remain as other key customer value components.

Besides these three elements of revenue, servicing costs, and risk costs, organizations still need to figure what is the appropriate time period in developing the customer revenue component. The time period is critical in determining what an appropriate purchase time window is for an average active customer.

Why is this important? The typical purchase window for an average active customer purchasing groceries (one week) is very different than someone using their credit card (one month) which again is drastically different than someone purchasing a car (5 years). But in all these cases and for all organizations, analytics should be conducted to determine the appropriate time window. How is this done? Listed below is a chart which looks at the total % of active grocery customers (i.e. have made some form of grocery purchase activity) within that time period.


From the above, we look at where the line begins to elbow to determine the appropriate time window. In this case, we observe that in 2 weeks, 80% of all active grocery customers make a purchase within 2 weeks. The elbow actually represents that point where the marginal increase significantly flattens, often used by economists as the point of optimization.  If we compare this to the typical purchase window of the tire customer, we observe the following:


In the above case, it would appear that 2 years might be an appropriate purchase window for active tire customers if we adopt the elbow theory.

In both above cases, analytics was conducted to produce this type of report in order to determine what a typical or appropriate purchase window activity is.

Once this “window” is determined, we then have a way to effectively calculate the customer revenue component of customer value which is essentially the customer purchase activity over that time period. But how do we determine best customers. Again, analytics is conducted to produce a Pareto type decile chart where customers are sorted in descending order by value into deciles.  The actual cut-off points in determining the value segments would be a joint exercise between the analytics/data science area and the appropriate business stakeholders.  Here the top 2 deciles (green) represent high value and account for 69% of all value across the 223K customers while the pink deciles (decile 3-6) represent medium value and account for 30% of all value across the 223K customers while the blue deciles (decile 7-10) represent low value and account for only 1% of all value across the 223K customers.


Decile


Value Segment


# of Customers


Avg. Profit

1

High

22,317

$592

2

High

22,317

$248

3

Medium

22,317

$156

4

Medium

22,317

$107

5

Medium

22,317

$69

6

Medium

22,317

$39

7

Low

22,317

$16

8

Low

22,317

$2

9

Low

22,317

$0

10

Low

22,317

-$11

 

Total

223,170

$122

In our increasingly digitized world where information is harnessed with AI tools, the ability to integrate the value component provides additional insight that allows marketers to simply further enhance the customer experience.

So far, I have looked at defining best customers and value segments based on actual observable behaviour. But given our increasingly digitized world, should we be incorporating digital behaviour of a non e-commerce nature into our value-based solutions.  The answer of course is yes but what about potential value and how should this be integrated alongside our actual value solutions?


Richard Boire, President
boire@boireanalytics.com
Direct: 647-500-8053
www.boireanalytics.com