Leave a comment
Get the GH Bookmarklet

Ask GH

  • PV

    Philippe Vdhd

    over 3 years ago #

    Define "about to churn"?

    The moment someone cancels their account (or doesn't renew), is not the moment they actually churn. Churning started when they stopped using your service. (or probably more accurately, when they stopped receiving value from you)

    So the best way to identify who's going to churn, is by looking at their activity.

    If they've used your tool once every day and they only log in once a week... is that a bad thing? Think about how often someone needs to use your product to actually get value from it.

    For example, if your customer is a performance marketer and your tool helps them identify the best performing ads, they should be logging in every day because your tool should be (one of) the most important tool for them. On the other hand, if you're AirBnB, almost nobody is going to log in daily to book a new stay.

    So first look at what the usage of a "good customer" looks like. Then see which users deviate from that behavior and take it from there

    • ST

      Stanley Tan

      over 3 years ago #

      Well said. How about if you are an eCommerce store and your customers don't actually cancel their subscription? How would you identify if a customer is about to churn?

      • AM

        Asia Matos

        over 3 years ago #

        Now I'm curious - are you looking at churn as cart abandonment? Is the ecommerce store managing accounts or people in some kind of way?

  • AM

    Asia Matos

    over 3 years ago #

    At Hull, our Head of Customer Success actually has a grading system based on certain behaviors that indicate churn. He discusses every risky account almost on the daily with the Sales, Product, Engineering, and Marketing teams to help contribute as well.

    For us, we predict churn based on actual product usage, but also qualitative information from the actual onboarding and implementation.

    Our customers tend receive the most value from Hull after the first month of using the platform, so the first 30 days are crucial.

    We look at factors like:
    - Was their implementation configured in time?
    - Were there any issues with implementing the platform?
    - Did the onboarding go well?
    - Did they login or someone in the account login to use the product?
    - Have they logged in recently?
    - Do they have any open support tickets?
    - Were the tickets addressed in time?

    We also use our own platform and tech stack to trigger Slack notifications when an account hasn't logged in for 30 days, if they've added any new users to the platform, and more.

    During the onboarding phase, there are certain actions we take to mitigate any churn risk. We really give our customers a "white-glove onboarding" - going out of our way to ensure they have an excellent experience with us and they ultimately get the most value out of their investment.

    But it doesn't stop after 30 days - we do have the regular cadence and touch, but it's also important for Customer Success to periodically investigate any issues in the platform or learn the customer's new goals.

    Our Head of Customer Success goes into more detail here:

  • AH

    Alex Holmes

    over 3 years ago #

    The best way to predict churn is to ask the right questions.

    We had an issue where we'd ask clients 'hows it going' and they'd almost always say 'good.' Saying 'good' is way easier than saying 'I have a problem.' Many of these clients ended up churning.

    So, think about questions that can better define a customer's success.

    NPS is a classic example of this. 'How likely are you to recommend this service to a friend.' You're not actually looking to ask them to refer you to a friend, but in order to refer something to a friend, you have to really like/understand it. So, some smart people figured out NPS as a proxy for 'how happy are you with your service.'

    Other questions you can use:

    - What's your number one issue with our service?
    - What do you think about our pricing?
    - How does our service compare to the competition?
    - How will your business change in the next six months?
    - How many hours per week do you spend configuring/troubleshooting our service?

    Things like that..