Hacker News Reader

Comment by zende | Hacker News Reader

just throw them all on Hulu and collect direct metrics. why on earth would you have people vote or rely on Nielsen? each view is a vote, and there is no limit of time-slots on Hulu

View on HN

It is highly non-obvious to me that the demographics who use Hulu and the demographics advertisers pay $$$ to reach are the same demographics.

Advertisers do pay to reach demographics on Hulu - just look at all the advertisement breaks in videos on Hulu. The problem is that there is no distinction between Hulu users who also watch Television and Hulu users who do not watch television; this compounds the problem of not knowing if the Hulu users who watch Television are a sufficiently representative sample of Television viewers to predict whether a show would succeed.

Which is probably what you meant, but your phrasing was (to me) a bit off.

The other problem is that people that watch shows online might not even consider watching the same show on television even if they're avid TV viewers. For example people who flock to "OMG this pilot is hilarious!!!" links on social media when bored probably aren't going to watch the series during prime time

I think it is a far more valid statement to say that demographics who use Hulu are a subset of the demographics advertisers will pay to interrupt with their message, probably enough of a subset that some advertisers just won't bother to advertise as at all.

However, targeting certain demographics by way of traditional advertising, like paying to interrupt everyone watching a broadcast television station, can suck as bad or worse, depending on the demographic.

As far as I can think of, it comes down to a couple of decisions: how much do you believe the medium, and not just the content, is going to affect the number of hits you get; and, how much is the feedback data you gather from your medium going to help your advertising campaign become more successful. There may be fewer people on Hulu, but your television does not provide a "thumbs-up" button at a per advertisement granularity.

So, while I think that the metrics would be bogus for trying to get a representative sample of a collective demographic who might watch motion video on any medium, I also don't think there is no worthwhile demographic information to be gleaned from posting one's media content on Hulu. It feels like a game of tradeoffs, and one you can probably only make an informed decision on by doing test runs, and seeing what data comes back.

Hopefully, this doesn't seem too pedantic; I just thought your assertion was too broad.

It's not far from the reasons you might choose to do internal betas and focus tests instead of test marketing every revision and variant you come up with from concept to product.

If pilots were all released on HULU, then a season filled with middling to poor pilots would be reviewed as such. And then no matter how much a studio might try to improve the best of the lot, they'd have those negative reviews dragging down performance of the ultimate product.

If pilot feedback is reasonably controlled, the studios have the option of polishing an almost-there concept (or more cynically: the best turd of the bunch) and not having reviews of its dodgy pilot dragging down performance of the final show.

there is no limit of time-slots on Hulu

This also means you can't manipulate the results by dumping shows in awkward time slots (interpreted another way, you avoid having to match each show to a time slot when its potential audience is likely to see it).

Hmm... There's only one of each, so to know whether I like it, I have to watch it. How are views going to indicate what people liked?

They indicate what people liked enough to, say, link on their website and get other people to watch. An inexact method, but then, so is like/dislike panels.

They can also measure things such as: what percentage of people decided to stop watching halfway through.

Reply on news.ycombinator.com