In today's LA Times, a spokesperson for TiVo was describing in granular detail the viewing behavior of the new Jay Leno show at 10pm. Apparently 20% of viewers who record the show watch it within an hour -- potentially cutting across other NBC programs. Never mind that time shifting can take all kinds of shapes: maybe these viewers wouldn't have otherwise watched the NBC shows at 11p or midnight. Maybe they were recording Conan while watching Jay. Maybe the fact that they watched within an hour is more about the need to keep up with the topical nature of a daily talk show (one that leverages the latest news for comedic effect) more so than, say, a drama. Watching a chat show a week later is like reading Sunday's paper on Thursday.
But the Leno/TiVo story got me thinking about statistics and analytics... And how few useful data points we have at our fingertips regarding the consumption behavior of music fans. When SoundScan came along in the early 1990s, it signalled the death of (highly subjective) record sales reported by music stores. At the time the labels were dismayed they could no longer manipulate the charts (gasp!) but in the end their corporate parents appreciated the reality check. It reduced some uncertainty when it came to projecting quarterly results, and in the mid 90s BDS and Mediabase brought similar science to radio airplay metrics.
But in an increasingly fragmented culture -- with people experiencing music across radio, CD, iPod, television, desktop, web, mobile, shopping malls and multiple other channels, it is surprising nobody has yet attempted to amalgamate all this data and attach value to it. Doesn't Music 2.0 depend on such quantitative analysis? Wouldn't a composite snapshot of the commercial impact of these channels add some heft to the valuation of, say, a tour sponsorship or a 360 deal with a brand partner?
Why is it that the TV guys have always been more sophisticated about audience measurement? Because they've had to be. That's how they've sold advertising. That's how they've monetized. The Nielsen-driven CPM model is woefully outmoded (arguments about "lean back" versus Hulu-style viewing behavior and the relative value of each abound at media conferences), and yet year after year the networks' upfronts are based on these archaic figures. And the amounts paid "per thousand" get larger each year -- despite the dithering of the digirati. Imagine the M&A analysts at Goldman using abici to determine the valuations of their transactions.
But the upfronts are only part of the story. Behind the scenes, integrated media specialists at the networks and cable/satellite providers are acutely aware of time shifting, Hulu, Apple TV et al, and are optimizing their advertisers' spend against these evolving behaviors. They have become adept at slicing and dicing each channel's audience and offering sophisticated pricing models for each.
So why not music?
Just this week a start-up called Next Big Sound got funded by Foundry Group -- both entrepreneur and investor are based in Boulder. It's no coincidence that the same VC is backing TopSpin Media -- a company dedicated to building clever CRM software that allows artists to engage their fans directly. Clearly part of Foundry Group's investment thesis is that music will at some point re-monetize, and they want to have their fingers in a lot of pies when it happens. Companies like Big Champagne and Band Metrics also measure new types of music consumption behavior -- NBS claims to have captured a quarter billion data points on a half million artists across platforms such as Amazon, iLike, Last.fm and Twitter.
When will we get as granular as the TV guys? When will a music analyst be able to tell a manager or a record label things like (a) when the music was consumed, (b) how much of the song was listened to, (c) what other music was consumed before or after it, (d) what products or websites were visited while the music was being streamed -- assuming a Slacker or Imeem-style playlist in the background while multitasking, (e) what was the click-through rate to related commerce links, (f) if other sites were visited, what were their page views and sales conversion rates, (g) what kind of device was used, and (h) based on the above, what kind of demographic or psychographic information can we infer -- by algorithm -- about said listener?
The technology is there, folks. We had a one-to-one relationship between content provider and music consumer -- or at least the ability to have such a relationship -- long before the TV guys went digital.
Until the data becomes this sophisticated -- and of strategic value to advertisers -- we're gonna keep hearing how the CPM-based "ad supported" music model is dead. It's time for the industry to put away the abacus and get serious about music metrics and valuation.