“Engagement” in the web analytics world is about as emotionally-charged a word as it might be with someone you’ve been dating for a week. At best, it’s a conversation-killer. At worst, it’s a nuclear warhead. Marketing and analytics experts have a hard enough time agreeing on what exactly engagement is, let alone finding the metric(s) to illustrate it.
But this confusion among smart people makes sense when you think about it. When was the last time you had a face-to-face conversation with someone, only to realize they weren’t listening? How can we expect to measure engagement with metrics, when we often can’t tell if the person right in front of us is truly engaged? In fact, the only people who can reliably tell when you’re tuning out are your friends, family, and significant others. There’s a reason for that. They’ve seen your behavior before, analyzed it, and suddenly, in their minds, you’re easier to predict than Paris Hilton.
Likewise, engagement means different things to different websites. Since each site has its own unique characteristics and purpose, engagement must be defined by your site’s goals — not by Amazon’s, eBay’s, or Ms. Hilton’s.
The first step is to define how an engaged visitor behaves in terms of your site’s goals.
So, what exactly does an “engaged” visitor do on your site? What are some of the clues that engaged visitors leave behind in your analytics?
Some sites will have an even harder time than others at capturing the elusive engagement in their analytics and may instead need to combine the quantitative data with qualitative analysis, like surveys. (Here are three great survey questions.) But proceed with caution. While many sites could benefit from using surveys on their quest to find missing pieces of the engagement puzzle, it’s easy to be mislead by what customers tell you in a survey. Ever take an online survey where the questions were fundamentally flawed? Do you prefer the taste of New Coke to CocaCola Classic? (The folks who were surveyed did.)
What’s even more dangerous is that only certain personality types bother to participate in surveys in the first place. (And good luck getting a Spontaneous customer to fill out a survey unless they’re either angry or bribed.)
A common approach to getting an initial handle on engagement is to take certain metrics that relate directly to your visitor’s main goals: those that measure if visitors are taking the actions you want them to. Monitor them closely, and see how these metrics play off each other when certain changes happen — e.g., changes in season, updates to a checkout process, special promotions, inactivity on a blog, industry trends — affect the site.
When selecting which metrics to use, keep in mind that it’s easy to be deceived by your own numbers. Proceed with caution by giving an in-depth look into the stories these metrics can tell you before placing your trust in them. In order to be sure that your metrics are an accurate reflection of engagement, you shouldn’t take one-off metrics at face value.
“Page Views” are a great example of a metric not worth trusting on its own. In this case, it may very well be that a visitor isn’t finding what they’re looking for. Perhaps they’re “pogo-sticking” from page-to-page in search of what they need. Now you’re keeping them on the site longer, thus increasing “Time Spent,” which, again, can be deceiving by itself. Although wasting the customer’s time — so long as they don’t leave the site — will increase the page views and time spent, it may not mean you’re actually engaging visitors. (Not in the way we’d hope, anyway.)
With your site’s goals in mind, and a rough understanding of how an engaged visitor behaves, here’s a sample of some metrics that may be useful relative to your site’s purpose:
Once a set of metrics is selected that directly relates to potential engagement on your site, constructing a weighted average of the set might help. This needn’t be some painfully complicated multivariate regression model, needing someone with rocket science experience like our buddy John to make sense of it; just some metrics that can serve as a collective vital sign to measure how well your site is engaging people while carrying out its core mission.
Jim Novo makes a potent case for using visitor recency to measure engagement and how to leverage it. If you can collect information relative to the history of each specific user, and the recency of their visits, his approach can send your ROI skyrocketing.
Novo’s approach shows how recency can explain a visitor’s potential value, given their propensity to return to your site frequently, as represented by the horizontal axis below. The vertical axis, meanwhile, shows how often the visitor has taken the action being measured.
Although fuzzy and directionally correct at best, engagement is vitally important to measure because it’s a predictive metric. If your current visitors are exhibiting behaviors indicating that they’re engaged, they’re likely to return soon — and often. If you see signs that visitors are becoming less engaged with the site, it’s safe to suspect that recent changes to your site or the flow of its traffic may be working against you. Either that or your competition’s finally outdone you. Regardless, it’s always good to know when to hang it up and try something new.
Engagement can also be a useful measure of the effectiveness of your branding. If visitors are showing signs that they’re engaged with your site, they’re generally showing affinity for your brand.
While engagement has become a heated buzzword, and arguably an excuse, it’s important not to be mislead. Since it’s a state of mind for your visitors, and therefore not easily quantifiable, there’s no simple way to measure engagement. But attempting to measure will help you to keep your site from proposing on the first date.
Do you have any unique approaches for measuring engagement? Let us know. We’d love to get a conversation going in the comments.