well they are a capricious lot yr 14-25s, they might happily say "yes mr marketer, yr magazine looks delightful and i would buy every issue" and then be like botherd when they actually have to spend £1.50 of their fag money on it, innit...
Sales volume research assumes automatically that something like 3/4 of the people who say "ooh yes I'll buy that" then won't. (It varies by category).
It's worth pointing out from experience that when dealing with new products and launches a hell of a lot of the time the people commissioning the research will just outright ignore it though! It might be that the research people said "This will bomb" and the backers just said "No it won't".
All the completely horrible and stupid products you see on the shelves are more likely a result of them being someone's pet project, or an R&D department wanting to look busy, than the research getting it wrong. Research gets it wrong in lots of areas but sales volume predictions are one where it's more often OTM.
Tom, you might be interested in this NY Times piece that Dave linked over on Cure For Bedbugs. Has a really interesting analysis of why consumer markets will always be unpredictable: You generally only get very well-known if you're slightly well-known first, and what gets you slightly well-known is idiosyncratic and somewhat random.
Of course, this doesn't explain the Popworld train wreck, given that you had something that was known. And bad writing (I assume ms_bracken is OTM here), would explain why it bombs in its second, third, or tenth week, but not why it crashes coming out of the gate.
If you're not signed up for NY Times they might not let you in - that isn't predictable either - but registering is free. However, in a few more days this piece will go over to Times Select, and that's something you'd have to pay for. If you have trouble accessing the piece, I'll email it to you. The piece is written by Duncan Watts, one of the men who conducted the original study; I'd read a number of news accounts (two or three or four) of the study when it originally came out last year in Science - I think one of you linked the Guardian coverage from poptimists - and (1) not one of the accounts gave a comprehensible account of Watts' procedures (which Watts was able to do in a couple of sentences) and (2) not a single one mentioned the point that Watts' considered most significant: that the fact that we influence each others' taste makes popularity somewhat random and hard to predict. The accounts stalled on the idea that listeners are influenced by other people rather than just by "quality" - if I recall correctly, the Guardian dope was using it to "explain" why Ashlee gets better sales than McCartney, not realizing that the piece applies to all popularity, Beatles and Beethoven as well. (No one's saying that "quality" has nothing to do with success, but rather that people pay attention to what other people are paying attention to, so popularity tends to clump around certain items. This applies as much to The Test Of Time as it does to new product.)
I'd need to re-read it again to actually understand the experiment they did (though I've read about it before).
Another current hot theory in research - which goes against this I guess - is "predictive markets", the idea that a large enough sample is a good predictor of things, but an even BETTER predictor when the sample is itself acting as a predictor, i.e. "What will you like?" gets you an OK result, but "What do you think other people will like?" gets you even better ones, and even BETTER better ones if there's something 'at stake' i.e. if you present it as a game.
I've seen the research - it did not say it would bomb, it was very positive. Or, at least the spin that was put on it was - there's a good chance the marketing people just cherry-picked the good bits, which as a former media researcher used to really piss me off, and is partly the reason I got out of research! Unfortunately I appear to have deleted it otherwise I'd send it over.
no subject
Date: 2007-04-19 01:55 pm (UTC)no subject
Date: 2007-04-19 02:02 pm (UTC)It's worth pointing out from experience that when dealing with new products and launches a hell of a lot of the time the people commissioning the research will just outright ignore it though! It might be that the research people said "This will bomb" and the backers just said "No it won't".
All the completely horrible and stupid products you see on the shelves are more likely a result of them being someone's pet project, or an R&D department wanting to look busy, than the research getting it wrong. Research gets it wrong in lots of areas but sales volume predictions are one where it's more often OTM.
no subject
Date: 2007-04-19 02:05 pm (UTC)no, i agree, why would anyone try and launch a music mag for that demographic in this day and age, it seems entirely counter-intuitive...
no subject
Date: 2007-04-19 02:53 pm (UTC)Of course, this doesn't explain the Popworld train wreck, given that you had something that was known. And bad writing (I assume
If you're not signed up for NY Times they might not let you in - that isn't predictable either - but registering is free. However, in a few more days this piece will go over to Times Select, and that's something you'd have to pay for. If you have trouble accessing the piece, I'll email it to you. The piece is written by Duncan Watts, one of the men who conducted the original study; I'd read a number of news accounts (two or three or four) of the study when it originally came out last year in Science - I think one of you linked the Guardian coverage from
no subject
Date: 2007-04-19 04:43 pm (UTC)Another current hot theory in research - which goes against this I guess - is "predictive markets", the idea that a large enough sample is a good predictor of things, but an even BETTER predictor when the sample is itself acting as a predictor, i.e. "What will you like?" gets you an OK result, but "What do you think other people will like?" gets you even better ones, and even BETTER better ones if there's something 'at stake' i.e. if you present it as a game.
no subject
Date: 2007-04-19 09:52 pm (UTC)no subject
Date: 2007-04-19 03:35 pm (UTC)