Anyone can reduce churn through analytics



The last time you stopped supporting a business, how did you do it? Did you send in a ‘Dear John’ letter, explaining why you are upset and what they did wrong? Did you declare it across social media? Probably not. Chances are you did make your unhappiness known through some correspondence and online gripes. But, ultimately, the relationship faded out. You found an alternative or perhaps felt you just don’t need that particular service anymore.

And a bit like a negligent partner who suddenly finds a letter on the fridge, the business that enjoyed your patronage only noticed you were gone when your contribution didn’t appear on the bottom line anymore. By that point, it’s too late.

“Losing a customer is more financially damaging than trying to attract new business,” says thryve’s Force Solutions Manager, Riaan Bekker. “Studies show that it can cost up to three times as much to get a new customer than to retain a current one. It’s also widely accepted that you should spend at most a third of customer lifetime value to acquire that customer. If they leave your service before you meet that ratio, you also take a big loss.”

Making sense of churn

Churn is very dangerous to a business. Yet, while companies are often concerned about churn, they don’t have the means to truly understand this phenomenon and counteract it.

This is because churn changes depending on the company, the products and other variables. Churn can be studied from various different vantage points. You could weigh churn from a customer numbers view or through revenue. Both are very relevant. The factors that impact shrinking customer numbers can differ from those eroding revenue. For example, losing customers may be a result of under-staffing – not enough people to serve customers.

In contrast, revenue shortfalls might relate to the value of a product and not customer service. The two areas can also overlap. With cohort analysis, you can find different patterns impacting different customer groups, such as how much they spend.

Some call this finding the ‘scent’ of the customer, and it used to be very hard to understand due to all the data variables.

“Churn involves a lot of different information, and you can spend too much time trying to figure out what’s right for your questions,” says Neer Rama, thryve’s Force Solutions Product Manager. “You might not want to keep every customer you have, or you might realise you should focus on more certain services or products. Investigating churn can answer many different questions, and that’s where we find companies struggle. It seems like too much to take in. They then regress to making calculations on the back of paper scraps. But doing that is wrong for the current customer environment.”

Today’s customer is much more vocal, impatient, and educated about their choices. This creates a paradox: it’s more vital than before to understand why customers might be unhappy, and yet it’s considerably harder to do so thanks to all the different information on offer. How do we square this particular circle? With analytics.

Reduce churn with AI

Analytics is not new. The practice literally means to analyze data or statistics systematically. But it’s received a terrific boost in recent years through artificial intelligence and cloud computing.

Today’s prominent AI models such as machine learning can track and mull the many data points generated around any customer. The cloud makes this affordable and accessible by pushing down the cost of computing power and providing easy access to AI services.

Their combination enables companies to work more creatively with customer data, asking different questions and following multiple hypotheses, but they don’t need to own the software that generates those insights. They might not even rely heavily on key skills such as data scientists.

thryve’s Rama demonstrates the difference with modern cloud software, in this case, Salesforce. Using tools such as Salesforce’s Einstein Vision and Discovery, he cleans customer data to be usable for an AI system. Then he sets specific attributes in the software:

  • Business goal: What is the broad goal of the analysis?
  • Outcome metrics: Which outcomes determine success or failure?
  • Levels of granularity: How deep or specific should the data go to match the business goal?
  • Data subset: What range of data should be used?
  • Independent variables: Are there variables that might influence the outcomes, such as a recent marketing drive?

Once these elements are set, they can generate surprisingly specific information about customers, including cohorts and individuals. They then offer remedial choices. For example, would a specific customer’s risk of leaving go down if they were offered a special discount? One click can show the probability of that.

“You can see all the relevant information there,” Rama adds, referring to the software screen. “It’s an interactive dashboard, so you can drill down or change parameters. The AI behind the scenes then does much of the boring work, connecting the dots so you can test your assumptions.”


Enhancing human decisions

This point is important: AI analytics doesn’t replace analysts and other business experts. Rather, it lets them test different ideas and theories. It stops them from wasting their time. For this reason, says Bekker, it’s important to get the right software and solution provider behind the project.

“Analytics software is pretty easy to find, and you already have customer data to use. But you need to align the software with your business and that’s where this stuff tends to fail and become more work than it’s worth. But if you collaborate with a solution provider who wants to understand your business and what you want to achieve, then analytics is one of the biggest advantages you can get for a business.”

The beauty of modern analytics is that it’s for every size business. From one-person shops to global multinationals, cloud platforms make it affordable and easy to use analytics that scale to operational requirements. Anyone can reduce churn by using analytics. Using information such as historical trends and social media sentiments, they can stop a customer from jumping ship and find ways to make them even more loyal.

Retaining happy customers and no more surprise ‘Dear John’ letters in the bottom line – why wouldn’t a business use analytics to reduce churn?