Posts Tagged ‘user engagement’

Audience Paradigms

Tuesday, August 7th, 2007

I’ve always tried to avoid using the terms ‘user’ and ‘users’ to describe the people who browse to, and interact with, websites. In my last post, I typed ‘user engagement’ and then changed it to ‘audience engagement’ as I have done before – when I remember to.  Anyway, that simple change triggered a little thought experiment that I thought I’d share here.

Take the term audience and think about what it means in various contexts… impersonal passive audiences exists within the print, television, radio & cinema paradigms.  At the theatre, audiences have a more personal relationship with the ‘content’ but are still mostly passive.  Live stand-up comedy audiences are almost always active participants in the content. Classrooms are perhaps the ultimate in participatory audiences and very personal. 

There are an almost infinite number of situations where you can place the audience/content relationship on the spectra of participation and personal connection.

What about the web? The relationship between web audiences and the content they consume speaks volumes about the evolution of the web in recent years.  We are moving ever so steadily along the path to from a print audience paradigm to one that approximates that of the classroom.

What does that mean? Do we all miss our classrooms so much that we want our web to resemble one?  Is the evolution of the web being driven by folks who still spend most of their time in classrooms? Is the classroom a healthy environment on which to model our online existence?

I don’t have those answers yet, but it has given me a lot to think about.

Page views are dead, now what?

Wednesday, July 25th, 2007

If video killed the radio star, then AJAX killed the page view metric. There has been lots of chatter lately about the irrelevance of the page view (PV). I have to agree with the conclusion that PV is no longer a reliable way to measure user engagement. I’m so convinced that, in fact, I’ve switched to a new metric (at least in theory) that I haven’t actually defined properly, though I did use it in a post yesterday.

Introducing… the interaction metric! Original, eh? It’s not, really. People no longer simply browse the web or ‘read’ websites – they interact with them. Luckily, there’s only one way that web user can interact with a website… do you know? Here’s a hint: It’s right under your index finger. The venerable mouse click!

That’s right, interactions is really just a fancy way of saying clicks. It’s an ideal way to measure the intensity of user engagement because it rolls form submissions (comments, etc), rating clicks, navigation into one metric and it includes what was formerly caught by the PV metric. It’s also AJAX-proof, which is important in the era of rich internet applications.

Are there other alternatives? The most obvious existing metric that could be used in place of PV is time on site. Earlier this month Nielson/NetRatings announced that they would be replacing the PV metric with time on site and total visits. While I think both these metrics are useful, neither has the depth to replace PV as a benchmark metric. Time on site, in particular, has a major drawback in that we have no way of knowing what the user was doing during that time. I must spend close to 24 hours a day on Google, for example, since I almost always have 2 or 3 search result pages sitting open – idle – I may not even be in front the computer! Update: Jeremy Liew talked about PV vs. time spent the other day and noted some change in how sites would rank. Unfortunately, interaction stats aren’t readily available to make the same comparisons.

The beauty of the interactions metric is that it measures activity within a visit or timeframe. It’s also extensible into types that could, in theory, be measured independently. Navigation clicks, for example, tell an analyst something different than a search submission. But I digress, it is the aggregate interactions that tells the analyst something similar (and maybe better) than the trusty old page view did.

So, a formal definition, then?

The Interactions Metric is defined as the aggregate number of clicks performed by an authentic user for the purposes of navigation, content manipulation, form submission, etc. Interactions with ad units are not included.

R.I.P. dear old Page View.

User Engagement Defined

Tuesday, July 24th, 2007

Apparently this is definition week. Yesterday I tried to define Web2.0 and tomorrow I’ll have another one. I seems an appropriate way to kick off a new blog, actually. Whatever I write over the next months would mean somewhat less if not written (and read) within the context of these definitions.

Dave McClure wrote this morning about measuring user engagement (UE). Problem number one, of course, is defining what it is that you’re measuring! Jeremiah Owyang offered his interpretation back in February:

Engagement indicates the level of authentic involvement, intensity, contribution, and ownership

I was going to write an entire post picking apart Jeremiah’s definition and creating my own, but I realize now that his is a really good one. I’ll instead take a spin at interpreting the four keys concepts in the definition to give them some depth and give some thought at how we’d try to measure each one:

  • authentic involvement – this is a 2-for-1 as it covers both the idea that we are talking about real users and that those users are involved – and not just passing through. The latter is the easier to measure, on most sites you could simply set a threshold of interactions per visit (greater than 1 should suffice). Any user meeting the threshold is involved. Authenticity is a little tougher to track – depending on the type of site – but IP blacklists and user-agent sniffing will go a long way to filtering out the non-authentic.
  • intensity – is essentially the level of involvement. Once we’ve established that a user in involved with the site, one way to measure intensity is by looking at how long each user stays on the site by both time and interactions. How often users return to the site is also a key metric for measuring engagement intensity.
  • contribution – is more dependent on the type of site than the previous two concepts. Contributions come in a variety of forms: forum posts, reviews/ratings, blog comments, questionnaires/surveys, form submissions. This is easily captured and measured, but can be tricky to interpret. Forum posts and blog comments for example have a considerable breadth in terms of how much contribution content is realized per contribution. The quality of the contribution is also a consideration and difficult to measure.
  • ownership – is the strength of connection that is felt by the user towards the site. This is perhaps the most difficult component of UE to track. Partly this is due to the fact that much of the expression of ownership is done off-site. Measuring ownership is almost an adjunct of reputation management. If you’re tracking what is being said about your site, you can measure how attached you users are to it.

It’s possible that I’ve completely misinterpreted what Jeremiah meant when he penned his definition of User Engagement. That’s ok, each strategist works from a slightly different set of experiences and will perceive the web, and its users, in a unique way. I’d appreciate any feedback!