• by John Spiropoulos , December 15, 2005

Every once in a while it's interesting to sit back and take stock of "the situation." Today we stand in the midst of a situation that warrants doing just that due its interesting, and albeit frustrating, juxtaposition of modern and archaic elements.

The era of mass time-shifted television viewing is upon us. More accurately, it's the measurement of time-shifted viewing that's here. Time shifting itself has been around for years, although you wouldn't know it by looking at data coming from Nielsen's various television samples.

Households with the ability to time shift via today's time-shifting technology of choice, the digital video recorder (DVR), aren't currently included in Nielsen's national sample. It's only now that Nielsen is in a slow build-up process to add these homes, which are estimated at nearly 10 percent of the population. Measurement of time-shifted viewing? Now that's modern.

But what's the big deal? How much viewing is time shifted? No one really knows for sure, it's not properly measured. But we know households with the ability to time shift programming watch more television than their counterparts who cannot time shift. This fact doesn't have as much to do with the technology itself as it does with the household's interest in watching television. Why would someone who watches one hour of television a week be an early adapter of a DVR?

By fall 2006, DVR penetration is projected to increase to around 15 percent of all U.S. homes. Households with this technology will account for roughly 20 percent of television viewing. It's unclear exactly how much viewing will be time shifted, but we can be sure it will be between zero and the aforementioned 20 percent. If we take the Nielsen Media Research model as fact, time shifting will account for about 5 percent of total viewing on a nationalized basis.

Nielsen has decided to retrofit this new dynamic time-shifted viewing data into the current systems. Three static streams of data are supposedly the answer. There's a stream for live-only viewing (or, non-time-shifted viewing). There's a second stream for live viewing plus same-day playback. And finally, the third stream is live viewing plus seven-day playback (168 hours from the live telecast's start time, to be exact). This process is archaic. It's clunky and out-of-date. Worst of all, it creates yet another needless battleground between the buying community and our network colleagues.

The new streams are as unnecessary as they are frustrating. Progress is supposed to move us forward, not keep us stuck in the past. What we need is event-level data, which is less bulky than current respondent-level data and more flexible than tomorrow's multiple data streams (starts December 26th). With proper edit rules applied to the raw data, event-level data allows for exact commercial audiences and not today's standard of minute allocation. In addition, sub-minute events can be properly and implicitly accounted for regardless of overall impact.

Yes, if we took a minute to sit back and take stock of the situation, we'd see that our industry is data rich, but technologically challenged. We need to push for more data fluidity and better data collection techniques to foster technological change and improvement. Some at Nielsen call the new respondent-level data, the minute-by-minute data, as the "keys to the castle." We task them to lower the drawbridge as well by stepping forward with event-level data. After all, it doesn't help if we have the keys, but can't get to the castle.