Flow Conference 2008: Thoughts on Media Metrics

After a day at the Flow conference, I’m reminded of how good this conversational round-table format can be. The Narrative Franchises panel was amazing: 40 or 50 people in the room, some of which were experts in the field who probably knew more and thought more deeply about the this phenomenon than anyone else on the planet, and instead of just throwing out self-indulgent tangents about their ultra-specific area of interest (as often happens at academic gatherings), they were contributing information or viewpoints that were totally relevant to the discussion, fleshing out the picture we had a franchises with examples, considering exceptions to rules and considering whether they constituted the re-write of those rules. I got the feeling we were building a larger understanding of the phenomenon than any of us came in with. It reminded me of the kind of collaboration that I thought could only happen online.

Media metrics (the most famous of which is Nielsen TV ratings) are to be thought of as currency, like money, or degrees on a thermometer. They provide a universal, transferable, singular standard by which we measure the value of something.

What creators really want to know is: what combination of textual characteristics and audience characteristics lead to more people buying an advertised product within that text and/or paying for the text itself? Whether or not certain people choose to watch a show or see a movie at a given time is an imperfect measure of that. The perfect measure would: not just measure whether or not someone was tuned in, but what was going on in their heads (look to transportation measures for this). It would also measure whether they viewed something at any given time (look to Tivo data for this). It would also measure the degree to which they talked about the content with their friends (look to optimedia’s measurements of buzz, or searches like Technorati that could tell you what people are blogging about, but also take valence into account). It would take into account previous viewing and purchasing experience. It would take into account characteristics of the text itself. Its not an infinite number of things you need to quantify and combine into some sort of quotient or scale, but its more multifarious than we thought it was before, or had the ability to measure before.

One of the dilemmas confronting researchers, ad people, and producers/networks will be whether or not they go with a census (in which case they're bound to lose a lot of the aforementioned nuance) or a sample (in which case you might not get the generalizability you desire). Maybe a flexible system like YouTube, where the stakes are lower, ad money-wise, will evolve new metrics quicker. They’ll weed out the ones that don’t tell you anything new about viewers and their desires and habits, but new tools for gauging that will evolve. Insight is just the beginning of that.