Using Machine Learning to Flip the Script on Customer Retention

Customer retention is more cost-effective than acquiring new customers. We imagine the journey of recognizing the need for machine learning from beginning the process to the execution of the plan in a sales environment.

As Sales Manager, customer retention is always top of mind. Getting new customers can cost five times more than retaining current customers. But, staying on top of keeping customers happy and accounts open takes priority. You feel like you spend all your resources fighting the pop-up fires instead of being proactive. It’s the first Monday of the month and time to review the account closures from last month, and this routine is getting old fast. It’s like it follows a script:

1. Review the accounts on the ‘Closed Last Month’ list.
   • Look to see if any accounts on the list completely surprise you.
   • Scan for the accounts you “knew” would close. Of course, “knowing” is mere intuition, but it occasionally serves you well, so you check if you were right.
   • Spend time figuring out any signs leading to unexpected closures. Then, spend even more time trying to understand why only half of the accounts you expected to see are on the list.
2. Prioritize the accounts you can save.
   • Determine how to allocate the limited amount of resources (and time) available to spend on closed accounts.
   • Spend more time determining if it’s worth the effort.
3. Dig deeper into the accounts to find trends and commonalities.
   • Attempt to prioritize keeping these accounts. But the day gets away from you, and despite your request, your team can’t afford to prioritize this list either when active accounts need their attention.
   • Reside to the fact you’re not getting to it again this month.

How Do You Improve the Script?

You start thinking about how to write a better script for customer retention. There is valuable information hidden in the report that can not only keep customers from closing their accounts, but it can improve processes and gain more time for employees. Ideally, you’d acquire a list of the customers likely to close, then proactively work with them to resolve any issues, instead of reacting after they close the account. In fact, your dream script would include the following steps.

1. Obtaining a proactive list or lists covering multiple pain points
   • Accounts likely to close the previous month and didn’t
   • Accounts likely to close this month
   • Insights into these accounts
2. Reviewing the distribution of accounts to team members
   • Making sure there is a game plan for these accounts and getting them any additional resources they need
3. Reviewing insights provided into accounts on each of the lists
   • Working with your team to focus on process improvements to improve the department and keep customers happier

This script moves beyond the administrivia clogging up your calendar and keeping you from making progress on departmental goals. This set up allows you to focus on actionable work versus planning.

Putting It into Action

Now, you need the time to create and implement this script. But, theorizing about the solution doesn’t create the solution. You need someone who can take your ideas and write the script for you. You need a ghostwriter.

How do you find them, and what will they do? You know there must be tools that can build the lists you want. A quick web search on predicting customer retention shows articles on predictive models and various services around machine learning. Does this require a level of specialized work? Maybe the solution is a project manager.

Those are all areas you can support. You can answer the questions around the business processes, and you know who on your team can clarify any data questions.

You remember to loop in the IT department because accessing the data and getting any resources set up requires their support, as well. You wonder briefly about the results and output from a model and hope that you can see what data goes into the predictions as well as the potential impacts from using the results.

You need a ghostwriter with skills in both business analysis and machine learning. You’ve worked with enough data scientists to know this is a tricky mix. Finding the right fit for this job will take more time and energy than you have to spare. You push the whole project to the back of your mind.

But, you can’t ignore the need for a better solution. You find the next available time in your calendar and reach out to your preferred consulting company to schedule a meeting with your Data & Analytics contact. They always have the resources for the projects you need, and they can provide the perfect ghostwriter to get this done. A discussion on the topic reveals they have done this exact work before and have developed a stepped approach for these Customer Retention Prediction models:

• Conduct a business objectives workshop.
• Build a trial model to validate predictive capability exists using available data.
• Conduct a controlled experiment to apply model-directed interventions in a real-life operational setting.
• Expand on the trial model with new data and features to improve accuracy.
• Integrate the model into operational systems and incorporate interventions into the business process.
• Adapt the model as you achieve business benefits. Focus on new issues or continuously improve on the existing model.

This process addresses all the areas you found in your web search, and you know this team will provide regular updates and stay engaged throughout the project to ensure that the model aligns with your processes and meets your needs. Feeling confident in the decision to move forward, you reach out again to schedule the workshop. You’re ready to flip the script on customer retention.

The post Using Machine Learning to Flip the Script on Customer Retention appeared first on Centric Consulting.