Rentrak hopes to crack ratings

Digital information offers more detailed info

Ever since the advent of the digital cable box, TV entrepreneurs have been trying to build a better Nielsen.

The digital information recorded by the box and sent back to the cable operator represents a treasure-trove of viewing behavior far more detailed than Nielsen ratings generated from a sample of 12,000 homes nationwide.

The latest bid to collect and analyze that data on a national basis comes from Portland, Ore.-based Rentrak, a media measurement firm that made its name reporting box office results and DVD sales.

Rentrak, which expanded into measuring the various studios’ video-on-demand offerings, recently adapted its software to measure linear TV, and is conducting a trial of 500,000 households to see if they can reliably process census-level ratings data.

“I think the TV ratings business is going to move out of a monopoly,” says Rentrak prexy Kenneth Papagan. “It may be that Nielsen remains as currency, but there will be more measurement companies giving perspective on TV use over the next five years, and Rentrak will be one of them.”

Last year, Rentrak quietly began testing its TV ratings service, TV Essentials, collecting data from 500,000 cable boxes among the 45 million from which it extracts VOD data around the country on major cable systems like Comcast, Charter and Cablevision.

Papagan says it will be months before Rentrak can draw conclusions from the data, which includes trillions of TV remote clicks for everything from channel flipping to ad skipping. But already some trends are starting to emerge.

First, as expected, Nielsen’s sample appears to quite adequately measure the big hits, like “CSI” and “Desperate Housewives.” It’s among small networks and low-rated shows where discrepancies between the set-top data and the Nielsen sample are appearing.

Papagan says it’s too early to tell what the discrepancies mean. “When you get down to the more targeted programs and networks, let alone unrated networks, you have no idea what’s going on,” he says.

The road is littered with failed attempts to disrupt Nielsen’s monopoly on the TV ratings business, on which $70 billion in TV advertising is bought and sold each year.

TiVo rolled out its own set-top data service, a sample of 20,000 TiVo units called StopWatch, which was quickly adopted by ad giants Starcom USA and the Interpublic Group of Companies, a sign of how strong the marketplace demand is for data beyond Nielsen.

ErinMedia, a Florida-based startup that won some funding, extensive treatment in a New York Times Magazine cover story in 2005, and quite a few converts within the major media companies, recently shut its doors and filed an antitrust lawsuit against Nielsen.

Media measurement firm TNS Media Intelligence and Cleveland-based Everstream have also made inroads.

There are two hurdles: convincing the cablers and satcasters to part with the data, and then processing an average of 20 keystrokes per hour multiplied by 65 million homes.

“No other company I’ve heard of is able to manage the amount of data required as well as Rentrak,” says media analyst Alan Gould of investment bank Natixis Bleichroeder, who owns shares in the company.

Nielsen is also doing deals with cable and satellite operators to mine set-top data with its Nielsen Digital Plus service, but it’s in a particular bind in incorporating the new numbers. It’s looking to the data to augment its sample, but if it emphasizes the service too much, clients may start to wonder if the 12,000-strong sample, on which Nielsen’s rep is based, is good enough after all.

Nielsen is planning to bump its sample up to 36,000 homes by 2011, a major investment of time and money, and a commitment to the old way. It maintains that set-top data is far from reliable, and that no system reliably discerns when a box is turned on but the TV is off, as is the case in many homes.

At the same time, Nielsen’s hold on TV ratings is probably as weak as it’s ever been. Demand for set-top data is especially strong among agencies and advertisers, who want more granular info. The networks want it, too, but are concerned that added precision might further erode ratings and benefit niche networks and shows, which aren’t measured well by Nielsen.

But for marketers, the potential is mesmerizing. Rentrak can tell, for example, that an MTV viewer in Detroit came from Comedy Central and then left to watch a Shakira video on MusicChoice — details that a small sample used to extrapolate broad viewer trends just can’t provide.

Most observers predict that data from set-tops will be at least as important as Nielsen ratings in the coming years –perhaps more.

“They won’t be able to replace Nielsen as ratings currency — it’s too entrenched — but they will be able to add qualitatively, because at the end of the day, networks need a real sense of who’s watching,” says Thomas Eagan, who covers Nielsen for Oppenheimer Research and called Rentrak’s TV service “ready for its closeup.”

Papagan says, realistically, it will be another 12 months before the glitches in the data are worked out. Some digital set-tops provide time stamps (which makes the data-crunching task easier), but some don’t.

Unlike Nielsen, Rentrak doesn’t own the data, so it will be up to the networks to decide what to disclose to the marketplace and the public. But if discrepancies emerge between the Nielsens and the set-tops, and they most certainly will, it could make the last dispute over program and commercial ratings look, well, quaint.