According to the New York Times, Hulu is angry, angry with Nielsen. ComScore, who probably doesn’t even like Nielsen considering their past, sides with Hulu. Hanging in the balance are the true traffic numbers of Hulu and the reputations of all involved.
The problem gets interesting when you look at the wide gap between the numbers – 42 million (ComScore) vs 8.9 million (Nielsen). Hulu was incensed that Nielsen ranked them so low and that they showed declining visitors March/April 2009. Yet in that same time frame they showed increased video viewing. So are less people finding the time to watch more?
ComScore is on the side of Hulu in this one but if you look at Quantcast (who lists Hulu right below it’s search box as an example on the front page) they have them at around 16.6 million a month worldwide (which would be about 2 million ‘sorry Hulu does not stream its video library outside of the US‘ messages per month). They do show Hulu as having continued growth though with a big drop in late April and a large spike in May. Now, Quantcast states that they are directly tracking Hulu and the Quantcast code is in place on the site (I checked the source code).
There is also soucre code in place for ComScore but I could find nothing for Nielsen. Maybe they are angry that Hulu isn’t buying their product and so is slandering them in the numbers. Just a thought I had, no proof at present so take it as pure conjecture on my part, more like wild speculation perhaps.
The thing about the numbers game is that if you pay for a traffic and demographics package it seems wrong. If we truly want an independent judge of internet traffic we need to look to a free service as they are not tainted by the need to keep their paying customers happy.
Personally, over at Analytics. Analytics and Quantcast never have the same numbers and in general, even when filtering out the staff at GDN, there is a disparity of approximately 8% with Quantcast always the lower of the two. It always puzzles me personally since code for both is included in every page served by GDN. How can there be a disparity of that much?we use , and Google
Of course that’s nothing like the 33 million visitor gap between ComScore and Nielsen or even the 25 million visitor gap between ComScore and Quantcast in regards to Hulu. This could be a case of traffic not being counted properly for video only hits. Perhaps, since Nielsen is out in the cold at Hulu, they aren’t getting the full access that would allow them to properly track the traffic. Maybe Quantcast doesn’t have the ability to track the Hulu viewer only visitors that might link directly to content and not in fact to a web page with the code in it.
I think some of this is a problem of how traffic is being tracked there and partly what is considered as traffic or a visitor. A unique visitor should be exactly that, a single person, but if the visitor isn’t served up an HTML page are they still a visitor? Not according to most stats packages I’m sure. After all, if you’re requiring code to be put in the page then that code needs to be displayed.
Frankly, if Nielsen isn’t directly recording every hit to Hulu, then in my eyes at least, their numbers are suspect. I had a similar problem with Quantcast in the past. They were reporting that we had far, far less traffic than Analytics said. Now that we have them both incorporated we still see a disparity but not the 50% that it was. When I contacted Quantcast about it at that time they said that they ‘pulled numbers from other sources,’ which was anything but helpful. Basically it meant they were guessing and I have a feeling that may be the case here as well with Nielsen.
The solution to this whole thing is easy: Hulu, show us the numbers. Show us some stats tracking logs that support your claims. If you want to take Nielsen to task over what you believe to be shoddy statistics generation, then put forth some cold hard facts and have them independently verified. Until you do, your traffic will continue to be a mystery and all of the reputations of the parties involved may take the true hit.