“This is a really underutilized segmentation technique.” An industry colleague shared that sentiment with me recently, when he and I were discussing the types of analyses made possible by cohort analysis using Web analytics data. Cohort analysis is a popular way to segment volumes of data to make more sense of the identified trends. In a purely Web analytics, visitor-focused context, cohort analysis means grouping visitors who demonstrate similar behaviors during a specified time period.
For an online publisher, an actionable cohort analysis may consist of comparing the behavior patterns over time of visitors who subscribed in October versus those who subscribed in March. Are the more recent subscribers better engaged with site content, given the UX enhancements made and the focus on continually producing better content? This is the type of analysis made possible with a cohort-based approach.
I’ve been really bullish on this type of segmentation, first discussing it in a column last November on no-nonsense Web analytics. I think it can bring incredible insight to digital marketing programs, and effectively reset our collective understanding of performance. It would prevent the pervasive question, “How does our performance compare to the industry average?” from ever being asked. It oozes with so much context and insight unique to the particular use case.
I also believe it holds the key to agile marketing success.
Rapid experimentation, failing fast, growth hacking – these terms have become the everyday vernacular for startups and an emerging group of elite digital marketers. Eric Ries, in his book “The Lean Startup,” even advocates for the use of “actionable metrics” to continually learn more about your operating environment, and recognize when to “persevere” or “pivot.” This rapid iteration and marketing agility is made much simpler via cohorts.
The legacy digital marketer is defined by his focus on “optimization.” Optimization is refinement on a microscopic level, based on individual keyword or creative performance. Cohort analysis enables optimization on a grander scale. It looks at user behaviors over an extended period of time, and provides macro-level guidance on what works and what does not.
In the case of agile and “real-time” marketing, where constant refinement based on newfound knowledge is required, cohorts show insight into the performance of the entire experience. Cohorts are especially useful for sites requiring user authentication (publisher sites, primarily), web and mobile applications. By analyzing user behaviors based on join/subscribe date, an organization can quickly create engagement baselines from which to judge subsequent refinements or the addition of new content.
The possibilities here are limitless, really. Technology startups get it. The growth hackers they employ get it. And as more demands are made of digital marketers, embracing the agile marketing mindset and its accompanying toolset will become requisite. We all need to get it.