In building the customer success function at a previous company, I faced a common challenge. We knew our prime directive: keep customer churn under 7% annually. We obviously knew that customers had to renew their contracts in order for that to happen. And we generally knew that they needed to be happy and engaged, with both our product and our company, if they were going to be retained. But that wasn't enough. It was too murky to establish a program around, to measure, or to impact. How could we know whether or not they were happy or engaged? How could we tell which were likely to renew, and which were headed off the rails to churn-land? And better yet, how could we use that knowledge to get out of reactive mode and instead proactively address churn risk before it was too late? Our solution was to create what we called "triggers" - pre-defined signs that a client might be in trouble (unsuccessful with our product, at risk of non-renewal or debook) which, when detected, kicked off an established, often automated or partially automated course correction process. And the secret sauce for our "triggers"? Carefully zeroing in on the right combinations of signs to trip the wire.

In the first post in this series, we discussed establishing business objectives - how your company defines success, from a revenue perspective - and business outcomes - the things your customers do that constitute that success. Now that we've set and prioritized the objectives we're aiming to achieve and specified the outcomes that support them, let's discuss how we can know when those outcomes are happening or likely to happen. For starters, a definition:

• Business Outcome Indicators - Signs or patterns, in the form of actions, events, attributes and/or metrics, which have a tendency to lead to a defined business outcome. For example, for the business outcome "renew for additional term", a likely indicator might be "utilize product regularly."

Channel Your Third Grade Science Teacher

So how do you know what kinds or combinations of behaviors and characteristics are indicative of the outcomes you're after? Amid the sea of all of the possible customer attributes and actions, how do you determine which ones hold the key to the mysteries of conversion, retention or expansion? The simple answer is: You know more than you might think. Give your intuition some credit, and recognize it as the perfect place to begin. Remember the good old scientific method from grade school?

• Start with a Question: What leads to our customers doing Business Outcome Z?
• Form a Hypothesis: If customers are like X and do Y, then they are likely to Z.
• Experiment: Analyze your data to see if the correlation you anticipated is valid. (If so, get proactive! If not, adjust your hypothesis and continue to experiment.)

Sounds great, right? Very third grade science fair. But, wait…

Watch Out for the Potholes

Where do you get the data you need to do the historical analysis on the hypothesized indicators and anticipated outcomes? Your company has likely been collecting at least some types of information about your customers' attributes and actions. But there are usually a few common hurdles to overcome here:

1. Incomplete Data - Even if your application's backend database has thorough activity logging and your company's CRM system is neat as a pin (and one or both of those are often NOT the case), you might find that it's difficult or impossible to retrieve historic attribute information, that the particular product usage event you are interested in has not been captured, or that a certain type of data (ex. billing) wasn't available prior to a recent implementation (ex. new finance system).
2. Siloed Data - What's more, even if your data is relatively complete, you are likely to find that it's strewn all over the place like a deck of cards in the wind. You might need to go digging in several disparate sources and departments to get the complete data set that you need for your analysis.
3. Lack of Analytical Tools or Resources - Finally, even if you're able to track down all of the needed information and gather it in one place, you might find yourself crunching sheets in Excel on nights and weekends, simply because you don't have a better tool or someone dedicated to doing that sort of work.

Are you facing one or more of these non-ideal scenarios? Never fear! Proceed with identifying the indicators you want to track, and use that list as a convincing argument to make the case for establishing a solid data pipeline and analytical toolset. And in the meantime, zero in on a few hypotheses to test that do have the necessary data available, and anticipate getting very hands on with the analysis.

Ask the Right Questions

To begin building your list of hypotheses, and in doing so identifying the relevant indicators, start by asking yourself the following questions:

Question 1: What things are true about a customer account and/or what have they recently done prior to renewing their contract for an additional term? (Insert each of your business outcomes in the green highlighted area.)

Question 2: What things are true about a customer account and/or what have they recently done prior to terminating their contract at renewal time? (Insert the opposite of each of your business outcomes in the red highlighted area.)

It's helpful to organize your thinking as you're listing indicators by categorizing them into three major groups: what customers are doing, what customers are saying, and what customers are achieving.

What Customers Are Doing

In this category, product usage is king. Although it's certainly not the only thing they are doing, it goes without saying that whether and how your customers are utilizing your application is central to whether they are likely to continue paying you for it. Make sure that any usage indicators you are considering as factors in your hypotheses are consistently tracked and available to you, or that you have a plan to ensure that they are in the future.

Besides activity in your product, there are other important usage behaviors to investigate, namely utilization or consumption of customer resources. These generally fall into two categories:

• Service Resources - Direct interaction with your customer-facing teams, such as:
  • Support Tickets
  • Account Management
  • 1:1 Training
• Self-Service Resources - Interaction with un-staffed and/or generally available supportive assets, such as:
  • Recorded Training / Online Tutorials
  • Knowledge Base
  • Customer Community
  • Company Website

To continue with our example above, the next step is to formulate hypotheses in answer to your questions, identifying indicators in the process. Let's start with a couple "what customers are doing" types of indicators. For instance:

Hypothesis 1: Customers who have added two or more users in the past year are likely to renew their contract for an additional term.

Hypothesis 2: Customers who have never submitted a support ticket, attended a training, or searched the knowledge base are likely toterminate their contract at renewal time.

List as many hypotheses of the "what customers are doing" type as you are interested in potentially testing. Highlight (as in blue here) the behaviors you are interested in examining. Those will become your list of indicators to track and measure. As you can see from the example above, they are often behavior trends over time or combinations of behaviors rather than stand-alone actions. However, there are exceptions to that generality, such as:

Hypothesis 3: Customers who have visited the account cancellation pageare likely to terminate their contract at renewal time.

The best hypotheses are devised when you put yourself in the customer's shoes and imagine what path and actions you would take if you were considering or approaching, in this case, renewing or terminating your contract.

What Customers Are Saying

In addition to your customers' behavior, it is essential to take into consideration their voice. This category of indicators probes into items such as:

• Survey Responses
• Feature Requests
• Online Reviews
• Social Media
• Referenceability

Continue building your list of hypotheses, this time focusing on "what customers are saying" types of indicators. For example:

Hypothesis 4: Customers who have given an NPS rating of 9 or 10 in the past six months are likely to renew their contract for an additional term.

Hypothesis 5: Customers who have posted a negative review are likely to terminate their contract at renewal time.

Your list of indicators is now growing and becoming more diverse. You may have an intuition that there is an interdependency between indicators, and want to test a hypothesis that includes more than one category, such as:

Hypothesis 6: Customers who log in at least three times per week andserve as a reference at least twice per year are likely to renew their contract for an additional term.

You might also layer in attributes that are not directly related to customer activity, but are important characteristics of the account, such as health index (as assessed by an account owner, usually a distillation of what the customer is saying) or upcoming renewal date. For example:

Hypothesis 7: Customers who are within three months of upcoming renewal and have a health index of yellow are likely to terminate their contract at renewal time.

What Customers Are Achieving

Don't be lulled into assuming that what your customers are doing and saying, while both critically important to capture and consider, are alone an accurate representation of how they are feeling about your product or how likely they are to stick around. Defining and measuring the actual value that your customers are attaining through utilizing your application is the best proxy for determining what they are thinking, and must be weighed the most heavily. Stay tuned for our upcoming discussion on customer value.

Until Next Time…

We've now explored the business side of the "circle of success": stating objectives, specifying the outcomes that support those objectives, and identifying the indicators of those outcomes. In the coming posts in this series, we'll cover how to define customer value and identify its indicators. Doing so will enable you to create a robust list of hypotheses to test that frame success both from your company's and from your customers' perspectives. Operationalizing the process of creating, evaluating and iterating on those hypotheses will become the core of your customer success initiative and serve as the basis for demonstrating its impact.