Cable Companies to Log Viewing Habits–Is Privacy at Risk?

Ryan Radia

Cable Companies to Log Viewing Habits–Is Privacy at Risk?

According to Investor’s Business Daily, several major cable companies have launched an advertising start-up called Canoe Ventures that is developing the technology to record and analyze cable subscribers’ viewing metrics. As part of this plan, Comcast is reportedly building a “500 TB TV Warehouse” to store the aggregate habits of its 16.8 million digital cable customers. DCT6412 Cable Companies to Log Viewing Habits Is Privacy at Risk?

Canoe has the potential to bring television advertising techniques up to par with their online counterparts, which have for years enabled websites to tailor ads to individual visitors. While cable companies haven’t used digital set-top boxes to collect viewing metrics until fairly recently, the concept of tracking television viewing habits isn’t new. For several years, TiVO set-top boxes have logged these habits without any ensuing privacy violations. And last May, Charter Communications began selling set-top box data to Nielsen Media Research so it could analyze the viewing statistics of Charter’s Los Angeles-area cable customers.

It’s a near-certainty that Canoe will have privacy activists up in arms, warning us that consumer privacy is under siege and calling for regulators—or Congress—to act. But as always, privacy risks must be balanced with the wealth-creating upside of targeted marketing. Data mining can indeed co-exist with effective privacy safeguards, and disciplined firms must be allowed to experiment with intelligent methods of matching buyers with sellers.

The initial details of the Canoe plan as explained by Comcast reveal a fairly tame system unlikely to pose much of a threat to viewer privacy. The system would collect data on an anonymous basis, meaning it wouldn’t be possible to tie a specific individual back to a set of viewing records, and logs would be purged after a year. The specifics of this plan are still being worked out, but Time Warner Cable said in a statement that Canoe will have anonymization built-in.

Why should we take the cable companies at their word? For one, the very real threat of insider leaks (and subsequent public backlash) means that cable companies are sure to think twice before being anything less than forthright about their plans. Besides, cable companies already learned their lesson about transparency when Comcast’s concealment of certain technical details surrounding the Sandvine network management system spurred a massive public relations fiasco.

Aside from the PR angle, a privacy flub would have serious legal implications. As a recent article in Multichannel news points out, the Canoe system will not diminish customer rights as enumerated in Comcast’s privacy policy. Comcast’s policy allows data logging only if done on an aggregate or anonymized basis, so sharing any data with third-parties that can identify specific individuals is verboten. Under this privacy policy, if an accidental leak or data breach were to cause sensitive data to end up in the wrong hands, users would have legal recourse against their cable company.

Anytime a new data mining system is put into place, there are always accompanying privacy risks. Yet the benefits of data mining can be quite significant, and displaying more relevant ads can be a boon for consumers and corporations alike. Some people are really worried about privacy, and they may wish to err on the side of safety and subscribe to television providers (like DirecTV) that can’t track viewing habits. Or perhaps cable companies might allow users to opt-out of data collection.

Either way, those of us who can live with a remote risk of our viewing records being compromised should be able to enjoy lower cable TV prices, less obtrusive ads, or better television shows—all of which are potential upsides to targeted advertising.

January 12, 2009 | Comments |

Kenneth

Is it just me or does the size of that Warehouse seem excessive? If you stored equal amounts of data for even twice the number of users (33.6 million) for the next 15 years, 500TB (TB = 10^12) divides out to almost a GB (10^9) of data per customer. Can you possibly need that much space to just store viewing statistics?

Ryan Radia

I'm no mathematician but 500TB/16.9 million = 30 MB per subscriber, right?

rogerthat

You are indeed a mathematician, sir.

Kenneth

which just goes to show that I should always use Google to do my math for me instead of Excel :)

Thanks for correcting my error.

Jarred

I wonder how they factor into their data in the fact that most people leave their cable boxes on when they turn the TV off? Or is that just me? Maybe they give greater priority to those metrics measured after channel changes, with a tapering authority given to those hours upon hours spent on the same channel overnight?

Ryan Radia

Good point. Sorting through the "junk" data from the actual viewing data is going to be a technical hurdle, but I agree with you that the cable companies could probably develop some sort of compensating algorithm. Perhaps the viewing stats will place added weight on programs watched within, say, 2 hours of a channel change.

Although I'd love to leave my TV on "Monster Trucks 24/7" (if that's not already a channel, it should be) if that'd mean I would see endless advertisements for gigantic trucks senselessly plowing through stuff.

ciphone

indeed a mathematician

j.dean

Does anyone know of a way to allow consumers themselves privately monitor their own tv usage? Simply to allow people to realize how much tv they are actually watching? I am sure cableco's would not want us to figure that out and would be loathe to give us acces to the data that exists inside the box sitting in our homes. It would be basically the same data the wii smartly provides users. What's the big deal?

Cable Companies to Log Viewing Habits–Is Privacy at Risk?

Welcome

Archives

Search

Categories

Contributors

Tech Policy

Tech News

Bloggers We Like

Our Sites

Where We Work

Podcast Feeds

Sponsors