How to Improve Customer Retention with a Strike System

When we first built systems to help our customer success team with their book of business, we defaulted to traditional customer health scores. These scores seem simple but quickly fail because of two essential challenges: they’re hard to understand beneath the surface and are not actionable. They create questions like:

• What is the difference between a health score that’s a “B” or “C”?
• What’s the difference between a “91” and “85” health score?
• What does a “B” or “85” mean, and how was that grade created?
• Once we know that a customer is a “B,” what do we do now?

Health scores tell you when something is wrong but don’t help you diagnose the problem or the solution.

Another issue we had with health scores is that they often come “too late.” In the world of SaaS, one day of no action is too late. We wanted to create a system where we could take action as soon as a customer becomes high-risk.

After talking with our customer success team, we realized that several common factors signified a customer was at risk. For example:

• The customer hasn’t met with our team in over five months
• The customer hasn’t used our product in the last two months
• The customer hasn’t created more than two total use cases in the product

Customers risk churning if they meet any of these criteria, so we must act immediately. Rather than abstracting these factors into a health score, we keep it simple: each is a “Strike” (like in baseball.)

For our customer service team to be able to take action on Strikes, we identified several criteria for our new system:

1. Customer Success teams must be able to create “at risk” audiences (based on each Strike) without asking for help from the data engineering team. They should similarly be able to easily add and test new Strikes.
2. Strikes must automatically sync to tools like Salesforce for tracking and reporting
3. Severe Strikes must automatically trigger Slack notifications to alert the team for immediate action

In order to meet this criteria, we first use Fivetran to consolidate our customer data in the warehouse. Then, our customer service team uses Hightouch to identify audiences of customers with a Strike and then syncs those audiences to Salesforce and Slack.

B2B companies have some of the most complex data, and Hightouch is no different. When we create Strikes at Hightouch, we need data from Salesforce, Intercom, and our production database. We use Fivetran, the leading data movement platform, to consolidate data from all of these sources into our data warehouse.

Next, we used Hightouch to associate all of this data together. This was a one-time setup step by our data team so our customer service team could use self-service tools to build audiences with all the data they need. We have a parent model defined in Snowflake to represent all accounts we want to monitor. Underneath that, we have a variety of related models for various custom objects we define. Here’s the data schema setup we used for our customer data:

This is a real screenshot from the schema builder in our production Hightouch workspace.

Our customer success team uses Hightouch’s Customer Studio to create new Strike audiences. They use our code-free Audience Builder to use nested and/or logic to build an audience for each Strike.

For example, here’s a Strike audience our team built. This audience includes “Business Tier” customers who have not created a new sync or model within the last two months:

Once our success team defines these Strikes in Hightouch, they can take this audience and send it to Salesforce, Slack, or any of the 200+ other integrations we support. Implementing this framework has completely changed how our sales and success teams operate. For example, as soon as an account qualifies for a Strike, both teams are immediately alerted in a Slack channel with a message that looks like this:

We also send Strikes to tools like Salesforce to monitor what accounts have Strikes currently or have had Strikes in the past. Below is an example of a chart we built to showcase the number of Strikes for accounts within a specific quarter. This allows us to proactively understand what went wrong so our teams can mitigate risk in the future.

Each bar represents different at-risk accounts. We can see how many Strikes our at-risk accounts got last quarter.

Since adopting Strikes, both our success teams and our sales teams have been able to work in closer alignment to leverage actionable insights for greater success. The program is early, but it’s already led to several upsell opportunities and lowered our churn rate. In our very first week with the strikes program, we identified an account that was underutizing Hightouch. We wound up helping them find more valuable use cases, earned a powerful ally on their senior leadership team, and expanded the contract upon renewal.

If you’re a B2B company struggling with Gross Revenue Retention (GRR) and finding that you’re reaching customers a little too late, schedule a demo with Hightouch so we can help! (And if you’re not using Fivetran yet, you should definitely check it out.) Let’s do away with health scores and start a new era of Strikes.