Artificial intelligence for marketing
This is a guide on how to use an artificial intelligence for churn prevention.
Every company spends a huge marketing budget on acquiring new customers each year. While acquiring new customers is essential for all companies, retaining these customers as long as possible is what drives the profitability of the company. It is 5-20 times cheaper for your company to retain a customer than to acquire a new one. This is why an increase in customer retention by 5% can increase your company’s profitability by 50%-75%. If you know in advance which customer will churn you can take measures to prevent this customer from leaving.
Which of my customers will churn soon?
We can build an artificial intelligence for you that will predict which of your customers will churn in a specified timeframe. We push this data as labels into your CRM so your marketing team can then create precautious customer journeys to prevent these customers from churning. These experiences could be to remind these customers of the value they get from your product or service, an unexpected present, a thank you letter, an invitation to an event, a trial of a different product or premium service, a customer survey to show that you care about your customers or any other experience your team may come up with. If you can retain your customers longer, this will increase their customer lifetime value and boost your revenue. Building an automated churn prevention journey based on the data from the artificial intelligence we build for you, is one of the strongest levers you can pull to increase your retention rate.
What are the drivers for customer churn for my product or service?
The machine learning algorithms we use to create the artificial intelligence to predict your churners will also give you information about the drivers of churn. You will understand which variables cause churn and the importance of each of them. Equipped with this information your marketing team can build or fine-tune experiences to retain your customers. This process will furthermore identify suboptimal customer experiences or flaws in your service or product and give you the necessary information to optimize your overall customer experience.
Use Case: Artificial intelligence for churn prevention
A day in the life (before)
Scene or situation:
Maria M is the Head of Marketing of a financial service provider. Due to changes in the marketplace she is confronted with a report that shows a dramatic increase in the churn rate of the customers. Maria sets up a meeting with her customer lifecycle marketing team. She shows the report to the team and tasks it to take measures to increase retention in their customer base.
Desired outcome:
Optimize the current customer experience or build a new one that reduces the customers’ churn rate. This will require Maria’s marketing team to understand the drivers of churn and to know in advance which customers will churn soon.
Attempted approach:
Maria’s team orders a report on the churners from the analytics team. The demographic report shows no significant differences in characteristics of churners compared with non-churners. The analytics team points out that they realized that most customers that cancel their contract do so after receiving their annual bill for the next year. Equipped with this insight Maria’s team brainstorms ideas to prevent their customers from churning when they receive their annual bill. As most customers have an annual plan, the experience Maria’s team wants to design will be sent to a big percentage of the customer base of the company. Maria’s team rejects many ideas as too expensive and eventually settles for a thank you email from the CEO. Her team sets up an automated email in their marketing automation tool that will be sent to each customer one week before they receive their annual bill.
Interfering factors:
The problem with this approach is the very rough targeting based on a heuristic. Although only a very small percentage of the customer base is churning each year, a big percentage of customers is targeted with the churn prevention measure. This means that although only a few customers will really churn, the budget for the measure has to be split between a huge group of customers. Therefore neither costly communication channels like phone, SMS or direct mailing can be used nor a costly experience be built to incentivise the customers to stay with the company. With this approach a lot of budget is wasted on customers that would not have churned anyways.
Economic consequences:
Six months after the implementation of the new churn prevention experience Maria receives a report from the analytics team. The churn rate has not changed significantly since the new experience has been implemented. Maria is not happy and calls in her lifecycle marketing team again. After their discussion Maria and her team feel certain that the CEO thank you letter was not strong enough as an incentive. Maria’s company has lost six months to counteract the high churn rates and is therefore bleeding money.
A day in the life (after)
New approach:
Maria has hired aaimo to build an artificial intelligence for churn prevention. Using all available customer data, aaimo builds and trains a custom machine learning algorithm for Maria’s company’s specific situation. The output of the machine learning model is transformed into an easily interpretable score that is then sent into the CRM of Maria’s company. Now each customer profile in the CRM has a churn probability attached to it. Her marketing team is excited by the new score and decides to build three automated churn prevention journeys. The first one will contain the strongest incentives and is used for the customers with the highest 5% of churn probability. These customers get a personalized invitation to an event via a direct mailing and a discount on their next annual payment. The second experience will be sent to the customers with the next highest 10% of churn probability. Those customers receive a free trial for a premium service of the company via email. The third experience will be sent to the customers with the next highest 25% of churn probability. As this group is big, Maria’s marketing team decides to reuse the thank you letter from the CEO for this group, as this communication via email causes no additional costs per customer. Her marketing team is full of ideas they would like to test to prevent their customers from churning.
Enabling factors:
With the churn prediction algorithm from aaimo the customers can be split into segments based on their probability to churn. The budget for churn prevention can therefore now be focused on a small group of customers with the highest probability to churn. As the experiences can be automated with a marketing automation tool and the churn prevention score is calculated on a daily basis, the churn prevention score acts as a trigger for the churn prevention experiences. Once set up the experiences run continuously and can be subsequently optimized.
Economic rewards:
As the customers with the highest churn probability get the strongest treatment now, the churn prevention measures are successful. When Maria receives the next report on churn from the analytics team, she is delighted to find that the churn rate has decreased by 15% since the introduction of the new churn prevention customer journeys. She clinks glasses with her marketing team and tasks them to use the insights aaimo recently delivered on the drivers of churn to experiment with other incentives and experiences. Using the churn prediction score in their CRM, Maria’s marketing team will further reduce the churn rate in the future.
Many companies do not know if their first party data is sufficient for AI use cases. We make companies understand how to best leverage their data and how to improve their data collection efforts. We create machine learning models as a full service and deliver the output straight into the CRM so our clients can profit from artificial intelligence within weeks.
