Ideas, concepts, solutions, news
An AI-Driven Approach to Customer Success Playbooks
Searching “customer success playbooks” will give you quite a few answers. Almost all of them are based on dated ideas, rooted in the pre-data era. Most of them are based on best practices which in many cases are not even from your industry.
Your playbooks should be based on your data and not on best practices and expertise of other people.
Here are the steps that should be followed:
- Define success – Examples could include minimal churn, maximum expansion, both.
- Define account segments – Most companies have some level of segmentation. That’s the way that they assign Customer Success Managers to accounts. As an example, for large accounts they would assign one CSM per 10 accounts whereas for small accounts the ratio could grow to 100 or 200 accounts per CSM. Your data and AI can also help in recommending segmentation. That’s in a separate blog.
- Collect data – get historic information about (a) past engagement that your team (CSM, Sales, Delivery) had with your accounts. (b) In addition, capture data about your product usage. For SaaS companies it would be the breadth e.g. number of active users and depth e.g. use of certain features and for other it could be the amount of orders that you had along the time. (c) And the last major piece of content is customer service cases.
- Additional possible data points – such as, NPS survey, Subjective CSM input. Based on our experience we found these signals to be much less significant in the overall impact on account health and thus playbooks.
- Let your data speak – given the goal, the accounts segments and the data an AI model can provide a list of actions that drove success in the past. Each recommended action will include recommended target and its weight (significance) to the overall success.
So now you have a set of recommended customer success playbooks for each account segment that is based on your success goal, your account segmentation and the behavior that led to success in the past. And not behaviors that led other companies to success.
Finalizing your Customer success playbooks
Now that you have your data driven customer success recommended playbooks you can apply your knowledge to refine them.
You can remove recommended playbook items that you believe are not important. You can change the target of the playbooks items based on your knowledge as well as based on the current state. For example, the playbook may guide you to have monthly call with your accounts or have at least 2 executives engaged. But when looking at the current state you may notice that 70% of the accounts do not comply with this target. So rather than guiding your team to non realistic targets you can lower it for an intermediate period until it gets to a better shape.
The advantage of Data driven customer success playbooks is that they are not one time project. They continuously monitoring execution on one hand and success on the other hand and recommend changes to the playbooks when necessary.
Here is an example of Komiko’s customer success playbooks environment. It allows you to change the target, add/delete playbook items and view the current state of these playbook items versus existing accounts.
Customer Success Playbooks – Summary
The advantages of AI driven customer success playbooks include:
- Best fit to your environment
- The AI model will recommend the right target for each metric. No need to run a data science project.
- Quick to compare with your current state and try out
- Save weeks and months of discussions and debates
- Basis for a data driven customer success health score – more details in the customer success health score blog
Check out Komiko Customer Success Solution for more info and starting a free of charge trial.