How to increase customer profitability with machine learning
So, what is the best way to improve customer profitability? The simple answer is to make your customers more successful.
For a gross average, you could take your total revenue and divide it by the number of customers, giving the total margin per customer. But pretty much every business has some kind of 80/20 rule by which roughly 20% of customers provide 80% of revenue. Or at least where different customers contribute very differently to the bottom line. High-touch customers that need a lot of customer success managers and hand-holding will naturally have a higher cost to serve. But sometimes, customers that don't contribute much revenue also cost a lot to manage. Sometimes, higher-revenue customers can have lower overhead costs. You need to have a clear picture of what your best customers look like.
Here's where machine learning (ML) and pattern recognition come in. This analyzes the business interactions that have led to better-performing customer profitability segments, mapping the customer journey and profitability at an account level. There will be a distribution that can be analyzed based on sales data, renewals, customer support, and service yield patterns in customer responses, behavior, and outcomes.
Better data will give better signals that yield more intelligence and better results. But even basic data will produce useful insights. The data can come from existing sources such as customer experience or health, process mining, Net Promoter Scores, and interaction quality metrics. The key is to then combine and aggregate the sources together correctly.
ML models find patterns in the customer journey and map these to profitability segments using customer revenue and cost to serve (see figure 1). By studying the map, your business can learn which customer journey patterns lead to better customer profitability performance and then enhance the successful outcomes that optimize customer profitability.