Make Better Sense of Your SaaS Company’s Data

Discover how you can break down and better utilize the massive amounts of data your SaaS company collects by organizing it into cohorts — and how doing so will lead you to more meaningful and actionable insights.

In the world of SaaS, data is everything. It allows a management team to gauge the health of its company, better understand the impact of key performance indicators like customer acquisition cost, churn, and monthly recurring revenue, and make more informed, cost-effective decisions based on hard evidence, rather than gut feeling.

That’s assuming, of course, that SaaS businesses actually do something with the data they collect. After all, even the richest data is essentially meaningless if it’s not properly analyzed, put into context, and leveraged in a way that drives productive change.

In that way, the relationship between data and data analysis is similar to having the best sports equipment. Can it help you perform at a higher level? Absolutely — provided you understand the nuances of that equipment and deploy it in a way that is truly productive.

So, what should SaaS businesses be doing to better harness and leverage the massive volumes of data they collect every week, month, quarter, and year?

In our experience working with expansion-stage SaaS businesses, we’ve found that organizing data into cohorts — groups that share certain characteristics, similarities, or parameters — can be a highly effective way of better understanding the context and impact of many critical SaaS metrics. And, ultimately, that context is what makes data meaningful and actionable.

Two Important Types of SaaS Cohorts

While there are many different cohorts SaaS companies can create, those cohorts can typically be placed into one of two categories: time-based cohorts and segment-based cohorts.

At a high level, the difference between the two is simple. Time-based cohorts center around customer data that begins and ends within specific time frames (e.g., customer join date, customer cancellation date, etc.). Segment-based cohorts focus more on specific characteristics with customer datasets (e.g., customers who sign annual contracts, customers from specific regions or countries, customers who use specific products, customers acquired through specific marketing channels, etc.).

Both cohort types are helpful in gauging a SaaS company’s health and performance, but the context in which you would use them will vary. For instance:

1) Time-based cohorts

These are particularly effective if you want to examine metrics within a certain window and compare them against other time periods. In other words, if your data is showing that customers who signed up in 2012 are churning faster than customers who signed up in 2011, you can create customer cohorts for both time periods, and examine data that points to factors that may have influenced their experiences. While time-based cohorts won’t explicitly tell you why something happened, they can inform your questioning and allow you to dive deeper into the root cause of it.

2) Segment-based cohorts

These cohorts, on the other hand, can help you compare and contrast customers with specific characteristics or traits. For example, you might use segment-based cohorts to compare customers who use different product types or features, or you could use this method to better understand the impact of specific sales and marketing strategies on individual customer segments. That information can then be used to determine the level of resources you allocate to specific product lines or marketing channels.

Ultimately, you should use both types of customer cohorts to fully understand your data and put it into usable context. Doing so will allow you to step back from typical financial, sales, or marketing reporting, and examine your business in a way that tells a more complete story about customer interaction and experience.

Best Practices for Conducting Cohort Analysis

Whether the data you have on hand is more conducive to time-based or segment-based cohorts, the process you use to conduct a cohort analysis should be similar — identify a specific problem to study, collect and stack the data necessary to do it, and then begin looking for common trends, problems, opportunities, or failures.

That being said, here are four best practices that you should always follow when conducting cohort analysis:

• Ensure that your data is clean, relevant, and reliable: Redundant or inaccurate data is the arch-nemesis of cohort analysis. It’s absolutely critical to ensure that your data is up-to-date, consistent, and properly filtered every step of the way. If a specific data set looks wrong or out of place, make sure you examine it and look for holes. Not all data sets will be perfect, which is why you have to continuously clean, test, and filter your data after every step.

• Start small and work your way up: While it’s true that the more cohort data you have, the easier it will be to paint a more complete, robust picture, that doesn’t mean you should begin your cohort analysis with all of the data that you think may be relevant. Instead, start small by selecting the data points that you know are relevant, and then add data along the way if it adds context to your analysis.

• Perform a trend analysis and look for anomalies: Once you have all of the cohort data in place, you should perform trend analyses and try to identify holes or anomalies in your data. For instance, if you’re examining churn among customers who signed up in 2010, 2011, and 2012, and you notice a dramatic spike upward or downward in a specific time period for one of those cohorts, that typically indicates that there could be something wrong with your data. When you see those anomalies, make sure to stop your analysis, dive deeper into that data set, and identify any problems within your data.

• Create graphs and charts that clearly convey the output of your analysis: Data on its own can be dry, and it’s unlikely that everyone in your organization will understand the context or meaning behind a row of numbers in a spreadsheet. Creating graphs and charts is a great way to boil cohort data down to its essence and quickly communicate what it means for the business.

The real beauty of cohort analysis is that it allows SaaS companies to look at data through different — and often more meaningful — lenses, as well as remove the guesswork that tends to plague operational decision making. But, as mentioned above, your company’s ability to derive that value truly depends on clean data.

If you have that, then the process of boiling and filtering it down should be relatively simple. And from there, your SaaS business can use cohort analysis to light a path that reveals roadblocks and opportunities, and allows you to focus your energy and time on the right activities to drive efficient growth.