Understanding Metrics: Engagement

Data Science
SiteCrafting Systems Admin Joe Izenman Joe Izenman
01/03/18

Understanding Metrics: Engagement

We’re ringing in the new year with part five of our series on Understanding Metrics. For a high-level overview of the topic, check out the introduction blog.

The Goal

So far, we’ve covered several sales-focused measures of website success: online purchases, in-store purchases, and one-on-one sales. While cash flow is probably the end goal for your organization, the role of your website is often much more abstract.

“Engagement” can mean a lot of things, but here we’ll stick with a simple core: people actively doing things on your website. “Active” goes beyond the basics of clicking links and reading content. Engaged users comment on blogs, post on forums, and like and share content. They review your products and create user profiles, which provides valuable feedback and free content.

The Metrics

We can quantify user engagement several ways, though the details vary widely with the exact nature of your website.

  • Action volume: Simple counting works well for discrete, true-false events such as likes and shares. The user either performed the action or they did not. Future decisions are informed by both the engagement level of users across all content, and the engagement fostered by a given piece of content, across all users.
  • Numeric engagement content: A website with product reviews, or similar mechanisms, can track a score for site content to help guide later changes. Applied to users, this is a view into just how passionate they are about your product or where the risk factors lie that might drive their business away.
  • User value: Depending on the action, different users may provide a measurably different value. For example, a Twitter share by a user has value proportional to their volume of followers.

We can also apply these metrics another layer out: If a user has generated content, such as a forum post or blog comment, how do other users engage with that content? Our most valuable users will generate content that fosters engagement in others.

Measures of engagement feel intuitively meaningful. We can point to several logical explanations for how an engaged user base is good for business goals. But the truth is more complex. These intuitive links are not reliable, and assumptions must be made to bridge the gap. We assume that higher engagement is good for business from a few angles:

  • Increased time on-site. Engaged users spend more time on your site. This means more opportunities to sell, advertise, and contact. This seems straightforward, but it’s an assumption that higher engagement rates are causing this and not the other way around.
  • Brand loyalty. A user engaging with a brand feels a greater connection to that brand. Users with high engagement are likely not only regular customers, but strong evangelists for your brand out in the world. Again, we have to make some amount of assumption about causality, that the users we see as highly engaged wouldn’t have this brand loyalty anyway.
  • Customer profiling. Engaged users will have created a user account and profile to interact with the content, which can provide a solid base of information. But the real power of engagement comes from profiling the actions users take. A data-savvy organization can build a remarkably useful picture of their most dedicated and valuable customers.

With sufficient variety of content, actions, and user details, these profiles can also be predictable, helping us recommend likely engaging content and products of interest to our customers.

It’s possible to directly monetize these profiles, but it is no easy task. Even an industry titan like Facebook took years to solve the problem of turning engagement—and social network sites that are 100% engagement machines—into actual profit, by means of targeted advertising and the sale of data to third parties.

Next Up

You can probably tell from the growing length of these posts that the further we stray from direct revenue measurements, the harder things get. In the final post of this series, we’ll wrap up with the simplest kind of website and the hardest thing to measure: communicating information.

See below for the previous posts of the series: