Sadly, pretending churn doesn’t exist doesn't make it go away. If this were the case, then SaaS startups would be littered with enough success stories to fill the pages of every book in the Library of Congress. To face this unpleasant problem head on and measure churn, most people just go with the prevailing wisdom and employ cohort analysis.

You’re probably familiar with it, but if it’s been awhile, here’s a top-level refresher. Cohort analysis simply takes data from a given data set — say, an an eCommerce platform, web application, or online game — and breaks that data into related groups for analysis. In the SaaS world, these groups are generally grouped in two ways:

  • Acquisition cohorts are time based, and usually separated by the week or month in the year that users signed up for a product. These cohorts are meant to show you how long users stick around -- and when they churn.
  • Behavioral cohorts are groups divided based on their behaviors and actions in your app. Where acquisition cohorts are meant to help you determine the who and the when, people believe that behavioral cohorts enable you to dive into the why.

For example, a behavioral cohort might be “the group of users who enable push notifications upon first customizing their settings.” You might find out that this behavioral cohort, let’s call them Notification-Flockers for fun, has a 97 percent retention rate at day 90 in the adoption funnel. That’s totally amazing, right?! After all, you’ve probably been led to believe that a finding like this means you should now have ALL users enable push notifications upon first customizing their settings. This, according to the prevailing wisdom, will boost retention rates for your entire user base.

But there’s a problem with this line of reasoning: It falls into the classic trap of associating correlation with causation. This type of logical fallacy can lead to reasoning like this: “Everyone who plays World of Warcraft will eventually die. Therefore, World of Warcraft causes the death of those who play it.” Of course this is an obvious example of mistaking correlation for causation -- everyone dies eventually, whether or not they play WoW. But the same faulty logic is at play when you assume a causal connection between an active group of users who enable push notifications and a high retention rate. The only way to establish causation is to figure out the real reason why Notification-Flockers have a 97 percent retention rate and then map that behavioral insight back to an explanation -- in their own words -- of why they stick around and the successes they have achieved. And behavioral cohorts simply can’t offer that kind of definitive insight.

But there's a better way to leverage cohort analysis -- a method that will allow you to engineer the ideal scenario where adoption is an inevitability rather than a guessing game. And it all hinges on an elegantly simple concept we call “progress markers.”

Simply put, a progress marker is a description of when and why a user feels like they were able to get something done successfully. It clarifies the purpose of their actions. It shows you what they care about. And in doing so, it helps you truly see which aspects of your product your users find valuable. Think of progress markers as tiny “yay” or “aha” moments -- in a word, successes -- that must be written down in the the mother-tongue of the user’s own language. And, for your own sanity, they should probably include a description of the user’s motivation and desired outcome.

To nail down these user-defined moments of “yay,” we perform qualitative research in the form of SLOW interviews with your customers. These interviews dig deep and uncover user motivations and the ways that your users define success. The information we gather in these interviews will help piece together each “aha” moment users feel they need to achieve to accomplish their goals as they make progress through your product.

By establishing a step-by-step pathway of “yay” and “aha” moments for your user base, adoption becomes more of a foregone conclusion rather than a murky possibility. To begin to optimize your pathway, first you’ll need to connect the progress markers to their corresponding events using code. (As far as instrumentation goes, it’s fairly quick and straightforward -- your developers will thank you!).

Next, measure the average time in between the exclamations of “yays” and “ahas” for each time-based acquisition cohort (the ones separated by the weeks of the year). Let’s say you find that it takes an average of 5.2 days for the “Week 7” cohort to reach their first ”yay” and another 7.4 days for them to reach their second “yay.” Now it’s the product team’s job to move these exclamations of joy from a light plopping drizzle to an unrelenting downpour. This downpour means that new users are likely to have a series of “yay” and “aha” moments in quick succession, thus setting off dopamine fireworks in their brain that bring smiles to their faces. These rapid-fire “ahas” will help drive their progress on a wave of “yays” through every stage of adoption, all the way through renewal.

To return to our Notification-Flockers example, now you understand why your goals have shifted. Instead of merely trying to get everyone to enable push notifications and hoping for the best, you’re now focusing your efforts on having the rest of your user base achieve the same level of success as Notification-Flockers. That’s why progress markers matter.

Progress markers will also allow you to identify areas where users are slowing down in their progress and beginning to disengage. This will allow you to proactively iterate on solutions to address and correct specific struggles, lapses, obstacles, and workarounds. Instead of playing a guessing game with lagging indicators and false causalities, you’ll be positioned to propel users towards avid engagement by creating a rapid succession of wins using techniques such as:

  • Using triggers to drive usage
  • Clarifying design principles and aligning them with your unique value
  • Removing causes of struggles
  • Addressing workarounds and adding value
  • Easing barriers or obstacles
  • Providing the the rapid series of successes your users need

Practicing cohort analysis by simply using acquisition and behavioral cohorts essentially amounts to performing a postmortem while wearing a blindfold. But if you’re committed to driving engagement and and correcting your course before your customers churn, take a more proactive approach and incorporate progress markers into your analysis. Using customer-defined progress markers that you connect to events in your product will finally take the guesswork out of your efforts to implement solutions and foster a base of avid users.