December 10, 2007
New Engagement Metric: A Conversation Piece

On Monday morning, I was up at 3 a.m. working on a class project. Part of the assignment was to come up with an alternative metric for television.

I thought back to what I know about engagement and what it might mean from my friend Ivan Askwith's thesis, the Metrics & Measurement panel at C3's Futures of Entertainment 2 conference, and what we'd covered in class. There's been a lot of great discussion about a new metric, but few concrete suggestions about what might replace the much maligned Nielsen ratings and C3 (the commercial rating).

So, I decided to write a metric, put something on paper, and get feedback on it. That's what this post is about.

Now it's time for a major disclaimer. This is, with some refinements, what I came up while burning the midnight oil earlier this week for a class presentation. It is based on what I know about engagement so far, but it's not the product of rigorous research. It's untested. It's rough. And it's definitely not meant to imply by ANY stretch of the imagination that I think I've addressed the engagement question definitively.

What this is, I hope, is a tool to stimulate discussion about what an engagement metric may actually look like. I want to know what you think of this, whether you love it, hate it, or something in between.

I went into this with a fairly good understanding of what the key issues were in defining engagement and making it a usable metric, which I've tried to address. For me, the main ones are:

* Difficulty defining what precisely engagement is, and what type of behaviors to include
* Need for a single currency that advertisers will respond to and trust
* Understanding engagement in the program versus engagement in the commercials
* Measuring long-term versus short-term engagement with a product or media property

In a nutshell, what I'm proposing is a single metric based on the weighted percentage change, week over week, in various quantifiable audience behaviors. This really only works for programs that are being aired. Fan engagement in the longer term warrants a different approach, something that I'm currently working on for pieces in the C3 newsletter.

Because I think it's easier to explain these ideas in pictures than in words, I've put together a slideshow here:

If you have trouble reading the slides on the YouTube video like Mike, you can access a PDF of the slides here. Let me know what you think! And stay tuned...



maybe it's me, but the text in this video is mostly illegible. i'm very interested though. is there any alternative? a PDF maybe?


Yes, me too. Eleanor, the idea sounds really interesting.


I really like the idea of metrics taking engagement into account. It seems better from a fan point of view, and from an advertiser's (e.g: I just grazed with Biggest Loser 5 mins ago, but can't tell you who advertised during the show. I can tell you who sponsored Buffy when I watched it in 2003 in England).

My key question would be what would have to be done to sell the TV industry on metrics that require time. Right now, they seem to love how Nielsen tells you "overnight": even C3 seems painful to them. So add longer delays and though this makes for better science, I could see them not caring (?)

Other than that, though, my concern with this model is that it allows Facebook and Myspace to measure fandom. Not only is this a problem for those not online, but also for those whose fan communities take form elsewhere (whether Livejournal, personal sites, Television Without Pity, etc.). And beyond that, fan community membership might measure fandom as community, but not more personal engagements and textual love affairs that don't go online.

So those would be my immediate points to throw forward. But the interest in adding engagement, along with some of those other factors, is great, and exciting.