Gope Walker explores the difference between churn prediction and churn prevention.
Churn prediction is a type of technique that businesses use to analyse customer behaviour, predict the likelihood of them leaving, and then implement a strategy to retain them. It can be very effective but in the tech industry, a much more sophisticated approach is required. In this blog, we reveal why.
What is churn prediction?
The official definition of churn prediction is “modelling techniques that attempt to understand the precise customer behaviours and attributes which signal the risk and timing of customer churn.” In short, it is an approach used by businesses to predict customers before they leave with the idea being that if you know a customer is going to leave, or churn, you can put a process in place to stop them from doing so.
Churn prediction technology might assess how active a customer is, how much they are paying each month, when they last logged in, whether any complaints have been logged, and how much of the product or service they typically use in a given month. By evaluating this information, churn prediction can effectively determine how likely a customer is to leave within the next 28 days. But, as we’re about to detail below, it’s not always as black and white as that.
Why doesn’t churn prediction work?
The answer to this is simple: just because you know that a customer is likely to leave, it doesn’t mean you’re going to be able to retain them. When you do realise that a customer is about to leave, there are a few ways you can approach it. You could call them and offer them a new contract at a better rate. Or, you could offer them a discount if they lock in for longer. Alternatively, you could reward them with some vouchers. Knowing which approach will work best for you, and your customers is easier said than done and you only get one chance at this.
What’s more, with churn prediction in place and customer retention strategies in full force, you run the risk of alerting those customers who are neither here nor there. Perhaps you have a customer who has been paying for certain features but not using them or maybe you have a customer who has been meaning to cancel each month but never got round to it. While it’s important to meet every customer need, churn prediction models do come with a risk of an increase in churn.
There’s also the issue that churn prediction doesn’t always give a complete picture of a customer. Current technology relies on behaviour patterns and concludes that a specific action means a specific result. But, customers are different. Just because one customer doesn’t engage with the software or certain features all the time doesn’t mean they’re going to cancel – whereas for others it might. What’s more, churn prediction might predict that a customer regularly in contact with customer service might be unhappy but they could just want to know more about the tool.
Finally, the time and cost involved in marketing to new customers, training them up on your service or product, onboarding them and dealing with any requests in the first few weeks is often much more compared to if you simply focus on doing everything you can to please existing customers. There’s also the idea that if you’re worried that customers may be leaving, it’s probably too little too late. But that’s not to say you can’t implement effective strategies to help retain your other customers.
How can you make churn prediction work?
One of the best ways to prevent a customer from leaving is to get to know them – and this starts from day one. Learn about what they like, what they don’t like and why they left their last provider. If you know why your customers are joining you, you’ll have more information as to why they might leave in the future. By asking questions at the start of your relationship with a customer and constantly checking in throughout you can focus on providing a service or product that meets their needs. Not just that but you can assess whether the product or service you’re providing is meeting their needs as well as how likely they are to stay. Being a part of their customer journey will help you to be aware and stay aware of any issues that could result in them churning.
Ultimately, to do churn prediction right, you need to get close to your customers and really start building a relationship with them from the moment they express interest. While that might be easier said than done, CRM, usage segmentation and customer profiling can help. What’s more, you need to ensure any customer data tools are effectively set up and analysed to really use churn prediction to result in churn prevention.
Contributed by Gope Walker, CEO of Data Kraken
After working for blue chip companies for 16 years, Gope was disheartened by the lack of innovation in the analytics arm of most businesses. The desire to innovate to optimise and improve businesses using data-driven techniques was rarely seen at the level Gope deemed appropriate – hence the birth of Data Kraken. Today, the Data Kraken team is working with clients across multiple continents, offering data-driven insights that allow companies to manage their business as effectively and efficiently as possible.”
