From our friend Brian Stelter at the NY Times comes news that Hulu is challenging Nielsen's audience counts for its site.
Personally, I take any publicly available web measurements with a bushel of salt, except those from Quantcast.com for "Quantified" directly measured sites. Our own Quantcast traffic numbers match our internal numbers very closely.
Does Hulu, the Web’s most popular place for TV viewing, reach nine million people a month or 42 million?
Millions of dollars in advertising revenue may hinge on the answer. But no one seems to know for sure how big the site’s audience is.
Any way the streams of shows like “Fringe” and “30 Rock” are counted, it is clear that Hulu’s growth has been explosive, up 490 percent year over year, according to Nielsen Online. Hulu executives, however, are fretting that the company, one of the leading purveyors of ratings data, is undercounting the site’s visitors. They say Nielsen’s numbers hurt Hulu’s perception among advertisers and the press.
While Nielsen reported 8.9 million visitors to Hulu in March, another measurement firm, comScore, counted 42 million. Exacerbating the confusion, Nielsen’s numbers for April show Hulu losing audience while still managing to add video views, also known as streams.
lots more where that came from at - NYTimes.com.
LET THE NIELSON CONSPIRACIES BEGIN!
I’ll start. Nielson is a joke of a company and i like this comSCORE more not only because it’s better numbers, but that it makes more sense!
8.9 million viewers a month? UHHHHHHHHHHH NO. 42 sounds like A LOT MORE REALISTIC ESPECIALLY WITH ITS VIRAL MARKETING CAMPAIGN WHICH IS FANTASTIC
I have been struck lately that the ads Hulu serves up to me are almost exclusively PSAs, all in the context of missed-an-episode viewing of current programs.
I’ve only been coming to this site for a little over a month. What was going on to cause your huge spike in December and January?
Scott, any very high spikes in December and January were one day heavy linking events from other sites (Drudge). in January it was Inauguration and the Idol premiere. the December spike was for this post:
http://tvbythenumbers.com/2008/12/16/jay-lenos-lies-jay-lenos-damn-lies-andstatistics/
BTW, this strikes me as diversionary tactic by Hulu that borders on desperation. As Boris notes, Hulu’s primary problem is it can’t even sell the inventory it has. Nielsen’s counting doesn’t necessarily help that, but it doesn’t hurt it either since Hulu can show any advertiser their real numbers.
I’d look for management change involving the installation of an experienced sales executive leading the company, spun as reorg due to Disney’s new ownership position in the company.
doesn’t hulu know how many visitors they get, they run their website, you don’t need an independent study, they have the data right in front of them?
Is this some sort of breaking news? I was at a conference over a year ago when people from Break Media and other huge companies were saying that ComScore and Nielsen both suck and are off by large factors.
IT is a joke that both companies charge the amount they do, it is almost a barrier of entry for any smaller site to get into the advertising realm since they cant (and shouldnt) pay the astonomical fees for reporting.
Nielsen crapped all over TV ratings for years, and i am appauled that the online community is allowing it to continue. Both companies break every statistics rule i ever learned in Business School and the fact that free tracking companies numbers are more accurate adds credence to the fact that we should all do without both firms.
Lets look for converging data .. if Google Analytics, Server Data and Compete or Quantcast all have similar numbers – then that is Fact. Forget the overpriced fat cats charging websites $10k+ a month to get innacurrate recordings.
Figure it the F out!!
Saltynuts, the problem is you can’t present those numbers to advertisers on such a large scale as this. How do they know you didn’t doctor the numbers?
That’s because Hulu is also counting visitors who get the “we’re sorry that material is not available in your region”
I could understand it being hard to monitor TV viewership, so although I shouldn’t give they a pass, I could see. But, internet viewers? How much easier could you get? Like they could even add one of the little number counters on the bottom of each show.
Between Comscore, Compete, Quantcast, and internal tracking, every site is able to get a reasonably accurate. Hell, even Alexa, which notoriously undercounts traffic, lists hulu.com as the 200thish most popular site.
Everybody’s raw numbers, including Nielsen, all likely say the same thing. It just comes down to how each defines a visitor. It certainly suits Nielsen to manipulate the data to lowball the traffic in this case.
@ Robert: I could understand Hulu not being able to sell its inventory might be caused by advertisers thinking they have a low number of views based on Nielson reports. It’s far better to advertise with a show that gets only 4 million a night, 4 times a week then to advertise with a site that only gets 9 million in a whole month.
If advertisers thought there were a lot more views on Hulu then they would be buying up that unused ad space.
This really speaks to the bigger issue – the rating system in general is antiquated and needs to be overhauled. Even the way DVR viewing is counted is off, way off. For instance, right now I have this entire season of Lost on my DVR and the second half of the 24 season as well.
According to Neilson, if I were one of their families, Lost or 24 would not be counted right?
I like to savor things, and watch serialized shows all together for maximum enjoyment. This should be factored in.
Jim, the problem with waiting so long to enjoy your serialized shows is that you obviously skip over the ads when you DVR it… which defeats the purpose of the advertisers paying for ad space entirely… and it’s kind of like stealing..
I personally think that DVR’s should be programmed so you can’t fast forward over the advertisements… it’s only fair, really..
How exactly does Nielsen count HULU’s visitors? I would think that HULU could easily dissect their own logs and use some custom code for the most accurate measurement.
@Joe: “I personally think that DVR’s should be programmed so you can’t fast forward over the advertisements… it’s only fair, really..”
That would be so unworkable. And besides which, some boffin somewhere would work out soon enough how to get around that “feature” and become very rich as a result. Also, it might mean people digging out their old VHS recorders!
Isn’t it curious that Hulu’s internal tracking for the month of March doesn’t appear in this story? Why wouldn’t they simply provide that data to Stelter to illustrate how far off Nielsen’s numbers are?
Without defending Nielsen’s number, I would say it makes perfect sense that Hulu might lose uniques while adding video views. It suggests that some users who tried Hulu after their Super Bowl advertising did not become repeat users, while repeat users have been increasing their number of videos viewed.
I don’t see how that exacerbates confusion, unless your presumption is growth in everything forever.
Can I offer a suggestion?
If you’re going to post a comment in which you rip Nielsen as a joke or antiquated, try spelling the name of the company correctly. It might help your credibility.
JIm, and even if you *don’t* skip the ads, the fact that you watch the shows so long after the fact means the ads in them are worthless to many of the time sensitive advertisers like movies.
why do they need to know visitors? Can’t they just count the number of streams and advertise that?