The Answer Factory: Fast, Disposable, and Profitable as Hell(wired.com)
wired.com
The Answer Factory: Fast, Disposable, and Profitable as Hell
http://www.wired.com/magazine/2009/10/ff_demandmedia/
17 comments
Just BTW: Regarding the time to prepare the article itself, it's standard practice for companies to send press releases to magazines/articles that are already ready for publication as is. It makes sense: they will benefit massively from the publicity, and they have the best access to the material. Who knows, in this case they may even have included the "too slick" passage for verisimilitude (with an ironic wink). Most probably, wired did extra work, and just used the press release as a base. I'm just saying that that base reduces the work they had to do.
In the long term, quality will improve, in tiny marginal steps. The company already talks about that extra $1 for fact checking; but if another company is doing the same thing, and starts ranking higher, there will be pressure to increase quality. (That is, assuming the ad-clicks are worth fighting for...)
It occurs to me that Google's pagerank is out of date: today, few people will add a link to their webpage/blog to one of these $20 videos, so pagerank can't rank them using links. (this isn't a danger to google; and their algorithm already uses many factors other than links). It would be a valuable to rank them somehow, using the behaviour of users. How to do that?
In the long term, quality will improve, in tiny marginal steps. The company already talks about that extra $1 for fact checking; but if another company is doing the same thing, and starts ranking higher, there will be pressure to increase quality. (That is, assuming the ad-clicks are worth fighting for...)
It occurs to me that Google's pagerank is out of date: today, few people will add a link to their webpage/blog to one of these $20 videos, so pagerank can't rank them using links. (this isn't a danger to google; and their algorithm already uses many factors other than links). It would be a valuable to rank them somehow, using the behaviour of users. How to do that?
Most likely Wired didn't do anything based on a press release. The P.R. industry would like you to believe they are able to drive the news agenda, but when it comes to the Wireds and The New York Timeses of the world, it's rare for a press release to lead to a story of this magnitude.
Most likely, Demand Studios' P.R. people freaked out when Wired inquired.
This story likely took weeks to report and write. A graphic artist probably spent at least a week preparing the chart. Good journalism takes a lot of effort, and so few people realize that.
As someone who has been paid to write and edit news for 30 years, this article was pretty depressing, but I already knew about Demand Studios. And you know what? It looks like they take good care of their writers, at least in the karma department. These seem to be folks who have always wanted to be paid to write but never had the opportunity to get paid for it, enjoy surfing the Web, and can bang out three or four of these articles in a day and have some good walking-around-money.
They seem to be mostly American, but at some point I'm sure the Third World will be recruited to drive the price down even further. Seen what they are paying on Mechanical Turk?
Most likely, Demand Studios' P.R. people freaked out when Wired inquired.
This story likely took weeks to report and write. A graphic artist probably spent at least a week preparing the chart. Good journalism takes a lot of effort, and so few people realize that.
As someone who has been paid to write and edit news for 30 years, this article was pretty depressing, but I already knew about Demand Studios. And you know what? It looks like they take good care of their writers, at least in the karma department. These seem to be folks who have always wanted to be paid to write but never had the opportunity to get paid for it, enjoy surfing the Web, and can bang out three or four of these articles in a day and have some good walking-around-money.
They seem to be mostly American, but at some point I'm sure the Third World will be recruited to drive the price down even further. Seen what they are paying on Mechanical Turk?
I wonder if we'll ever know, factually, who wrote to whom first?
I don't mean to slight wired; but I'm sure that many technology companies would like to be covered by wired, and if they have any sense, they'd prepare an article in the most convenient form for wired, and send it to them - and the guy running Demand definitely knows how to make things happen at this point. They wouldn't have any control over what Wired did with it, of course.
I wouldn't worry about it too much - it's just a drop in the ocean, and there are lots of other sources of information on the web. Probably, it's fair to say that having a crappy $20 video is quite a bit better than nothing. And if there's a few results to a google search, most people will look at more than one.
Example: I had a blocked toilet yesterday, google came up with a several excellent hits on it. I tried the suggestions in order, and problem solved. The internet is pretty cool.
I guess the depressing part is that higher quality can't be supported by adwords alone - it would seem, anyway. I think quality is often best rewarded in niches audiences and uses, where quality really matters.
I don't mean to slight wired; but I'm sure that many technology companies would like to be covered by wired, and if they have any sense, they'd prepare an article in the most convenient form for wired, and send it to them - and the guy running Demand definitely knows how to make things happen at this point. They wouldn't have any control over what Wired did with it, of course.
I wouldn't worry about it too much - it's just a drop in the ocean, and there are lots of other sources of information on the web. Probably, it's fair to say that having a crappy $20 video is quite a bit better than nothing. And if there's a few results to a google search, most people will look at more than one.
Example: I had a blocked toilet yesterday, google came up with a several excellent hits on it. I tried the suggestions in order, and problem solved. The internet is pretty cool.
I guess the depressing part is that higher quality can't be supported by adwords alone - it would seem, anyway. I think quality is often best rewarded in niches audiences and uses, where quality really matters.
I think you mean AdSense. AdWords are a way for content creators to <i>spend</i> money, not make it.
In any event, AdSense revenue is trivial compared to the cost of creating content that has any widespread impact.
In any event, AdSense revenue is trivial compared to the cost of creating content that has any widespread impact.
It would be a valuable to rank them somehow, using the behaviour of users. How to do that?
I'd be surprised if Google wasn't using Analytics to do that for many of the sites that use it.
I'd be surprised if Google wasn't using Analytics to do that for many of the sites that use it.
What is depressing is that Google likes that kind of content, when a lot of cheap articles are pumped in furiously. If they were after quality, they could de-rank Demand and that would make them out of business quickly. But they wouldn't do it, since advertising dollars are at stake.
But on the other side, it isn't that terribly depressing, since their site will occupy only one slot in the search results. So it is still a chance for others to come up with better content for such keywords. But as it seems, it is inevitable for online media producers to play the keyword game.
But on the other side, it isn't that terribly depressing, since their site will occupy only one slot in the search results. So it is still a chance for others to come up with better content for such keywords. But as it seems, it is inevitable for online media producers to play the keyword game.
At the end of the day, though, people want quality. If a low-quality SEOed article appears higher in the search rankings than a high-quality article, it is a flaw in ranking algorithm -- a weakness that other search engines can exploit.
I think Google is smart enough to realize that search dominance is their most valuable asset... that the money from search dominance dwarfs the money from favoring crappy content that matches ad keywords. This is just another step in the struggle between Google and the search engine optimizers.
(The deal between Demand and YouTube is curious, though... If the "real world" worked like that, prime time television would be filled with shows about cars, beverages, and pills.)
I think Google is smart enough to realize that search dominance is their most valuable asset... that the money from search dominance dwarfs the money from favoring crappy content that matches ad keywords. This is just another step in the struggle between Google and the search engine optimizers.
(The deal between Demand and YouTube is curious, though... If the "real world" worked like that, prime time television would be filled with shows about cars, beverages, and pills.)
"It's oft pitched as Wired readers, only about a decade younger (Wired is an extremely well loved brand by advertisers)." -kn0thing
http://news.ycombinator.com/item?id=906830
http://news.ycombinator.com/item?id=906830
There is a rattling great market opportunity here. Assuming Demand Media and friends manage to complete the race to the bottom - and why not, it's happened in every market - there's going to be a prestige knowledge-brokering market springing up.
Translate "personal shopper" to the limitless library and now you're talking. Tricky to scale outside niches - so it's most likely not a scale, VC business - but there's a million very nice consultancies waiting here for Internet safari guides...
Translate "personal shopper" to the limitless library and now you're talking. Tricky to scale outside niches - so it's most likely not a scale, VC business - but there's a million very nice consultancies waiting here for Internet safari guides...
Eek... Reading this article was like watching a nightmare become reality. I've been getting thrown into ehow.com quite a bit lately, now I know why the answers are so terrible. This sort of mathematical approach to maximizing profit is, to me, a cancer on the web.
The analogy in the article was perfect, they're like the kid that screamed out the answer to every question the teacher had, but with absolutely 0 insight.
The analogy in the article was perfect, they're like the kid that screamed out the answer to every question the teacher had, but with absolutely 0 insight.
As I understand it, the mathematics approach is for finding questions, and I don't think there is something inherently nightmarish about that. It's the quality of the answer that matters. I don't know, I just watched the "What is the best vodka?" video, and being an amateur in this area, I actually found it quite useful.
I hope this doesn't get buried. Fascinating article about how a media company is scaling itself in real-time (by analyzing over 100 high-value sources of what people are searching for) and algorithmically determining what obscure piece of content will generate a lifetime profit (at 55 to 60 cents per adsense click). They match the costs of production to the expected life-time profit, for example $15 for a few-hundred word article (fully acknowledging that most content on the Internet is cheap)
Instead of trying to raise the market value of online content to match the cost of producing it — perhaps an impossible proposition — the secret is to cut costs until they match the market value.
This is the reality that Rupert Murdoch is trying his best to ignore.
This is the reality that Rupert Murdoch is trying his best to ignore.
Interesting excerpt. But isn't Murdoch actually something of an "upstart" in media, who himself has been viewed as corrupting the journalistic standards of yesterday?
http://en.wikipedia.org/wiki/Rupert_Murdoch#Beginnings
http://en.wikipedia.org/wiki/Rupert_Murdoch#Beginnings
While it is a bit of a race to the bottom, it doesn't have to be.
Especially if you include the PG-style democratic effect, where the good eHow articles float to the top and the bad sink, which creates a feedback loop that rewards quality. As quality becomes more important, the eHow statistical analysis will demonstrate that and eHow will adjust the process to increase quality.
Alternatively, they might just stick to such simple topics that this is unnecessary. "How to Be Popular" is not a simple topic, but "How to Use a Can Opener" is. Using a can opener is a useful thing to have available and doesn't need expensive production values.
Over time, if the business is truly data-driven, it should adjust to reflect these realities and lend itself to fulfilling true human needs, not just yelling the loudest for short term profit.
Especially if you include the PG-style democratic effect, where the good eHow articles float to the top and the bad sink, which creates a feedback loop that rewards quality. As quality becomes more important, the eHow statistical analysis will demonstrate that and eHow will adjust the process to increase quality.
Alternatively, they might just stick to such simple topics that this is unnecessary. "How to Be Popular" is not a simple topic, but "How to Use a Can Opener" is. Using a can opener is a useful thing to have available and doesn't need expensive production values.
Over time, if the business is truly data-driven, it should adjust to reflect these realities and lend itself to fulfilling true human needs, not just yelling the loudest for short term profit.
Whoa, there. They do look at the data, and do what the data tell them makes the most sense. That's what that means. You can't say that you would do what your analysis tells you is right, and then decide what you'll conclude in advance!
Also, you might have noticed that the "democratic" parts of the web are mostly terrible.
Also, you might have noticed that the "democratic" parts of the web are mostly terrible.
What I'm saying is that the quality-effect would take time to become prominent.
First they have to profit off all the low-hanging fruit simply because there is no competition.
Then, once competition starts to build up, there is no more low-hanging fruit left and they are forced to compete on quality.
Over time, a democratic effect takes place and quality rises.
There is a difference between saying, "Cat pics are more popular than articles" and saying, "This eHow video is higher quality than this other site's HowTo video"
Cat pics don't compete with articles except on reddit and digg which love to compare apples and oranges. In a google search though, they are totally separate. So yeah maybe eHow will make 10 times more cat pics than informed videos but the point is that the cat pics will be automatically filtered out unless you search for "cat pics".
First they have to profit off all the low-hanging fruit simply because there is no competition.
Then, once competition starts to build up, there is no more low-hanging fruit left and they are forced to compete on quality.
Over time, a democratic effect takes place and quality rises.
There is a difference between saying, "Cat pics are more popular than articles" and saying, "This eHow video is higher quality than this other site's HowTo video"
Cat pics don't compete with articles except on reddit and digg which love to compare apples and oranges. In a google search though, they are totally separate. So yeah maybe eHow will make 10 times more cat pics than informed videos but the point is that the cat pics will be automatically filtered out unless you search for "cat pics".
I'm skeptical of this type of business, simply for the reason that I think you have to actually help people to make money. You can trick them in the short term or exploit some discrepancy, but this doesn't sound much different than spam, so I can't imagine it being successful long term.
There are some cool elements to this system - the algorithm for determining what content to create and the ambition of the founder but overall the system seems to be designed to produce the video equivalent of sludge.
Incentives are structured so that quality, in-depth content will be penalised while terrible, quick to produce movies will be rewarded highly. The only check on the terrible videos is the fact checking process but if you think you're going to get any kind of quality fact checking for a dollar then I have a bridge to sell you.
Given the rewards are much higher for the latter kind of video, these videos are going to grow to dominate the service. The producers of these will make videos about anything they can, not anything that corresponds to their field of expertise so essentially all you'll be getting (at best) is educated guesses about your topic of choice.
In fact there's probably a nonzero amount of money to be made in creating an automatic video generator which automatically composits videos and submits them under various names; remembering, of course; that it doesn't have to have actual information, it just needs to be good enough to get past the dollar per video filter.
The internet doesn't suffer from a shortage of information (we're already drowning in it); it suffers a shortage of quality. I can't see this doing anything to solve that.
That said, perhaps there's a market for sludge. If people want a few factoids about 'the best vodka' or something else that's utterly irrelevant, well maybe this will satisfy that demand. At least it's doing it cheaply.
Finally on an unrelated note this seems to be video for videos sake. Unless there's something being conveyed in the video that can't be done easily in text (for example recordings of the Dean Kamen robotic arm) I'd much rather stick to text - it's faster to consume, easier to parse and much more in depth.
Incentives are structured so that quality, in-depth content will be penalised while terrible, quick to produce movies will be rewarded highly. The only check on the terrible videos is the fact checking process but if you think you're going to get any kind of quality fact checking for a dollar then I have a bridge to sell you.
Given the rewards are much higher for the latter kind of video, these videos are going to grow to dominate the service. The producers of these will make videos about anything they can, not anything that corresponds to their field of expertise so essentially all you'll be getting (at best) is educated guesses about your topic of choice.
In fact there's probably a nonzero amount of money to be made in creating an automatic video generator which automatically composits videos and submits them under various names; remembering, of course; that it doesn't have to have actual information, it just needs to be good enough to get past the dollar per video filter.
The internet doesn't suffer from a shortage of information (we're already drowning in it); it suffers a shortage of quality. I can't see this doing anything to solve that.
That said, perhaps there's a market for sludge. If people want a few factoids about 'the best vodka' or something else that's utterly irrelevant, well maybe this will satisfy that demand. At least it's doing it cheaply.
Finally on an unrelated note this seems to be video for videos sake. Unless there's something being conveyed in the video that can't be done easily in text (for example recordings of the Dean Kamen robotic arm) I'd much rather stick to text - it's faster to consume, easier to parse and much more in depth.
I told a bunch of journalism friends about this over beers tonight. A few of them were really depressed by this.
One of them is actually on the margins of it. He just started writing for an AOL site and said that this is the direction they're going to start going with for their weblogs inc. sites. He gets a list of topics culled from Google Zietgeist and other sources that his editor wants him to write short posts about for $50/pop. They don't really care what he writes so long as they have content available on the topics that people are searching for right now.
This is the world that's been built.
One of them is actually on the margins of it. He just started writing for an AOL site and said that this is the direction they're going to start going with for their weblogs inc. sites. He gets a list of topics culled from Google Zietgeist and other sources that his editor wants him to write short posts about for $50/pop. They don't really care what he writes so long as they have content available on the topics that people are searching for right now.
This is the world that's been built.
While the lack of a strong editorial voice can be a negative, the pay surely isn't. As you note, $50 a pop for "short posts." For someone who's skilled in an area (as most Weblogs Inc writers were, at least) $50 to put out a 300-400 word post in 30 minutes or so is pretty good pay.
Word count is not an indication of the effort involved. Itg's not merely a matter of typing in words like a monkey. You have to know about which you write. The more you know, likely the more time it took you to learn it. Time is money, words are just words.
Word count is not an indication of the effort involved.
If it's for writing small news bits, it can be. Let's say you're a total expert with using GMail and you have to write a couple of posts linking to GMail news, the occasional tip, etc, each day - you can probably bash short posts out in no time.
You have the many hundreds of hours invested in actually become the expert in the first place, but I doubt anyone becomes an expert so they can write $50 a pop pieces - so that cash is extra value, not a payback for the time you spent learning (as is the case with a professional's fees).
If it's for writing small news bits, it can be. Let's say you're a total expert with using GMail and you have to write a couple of posts linking to GMail news, the occasional tip, etc, each day - you can probably bash short posts out in no time.
You have the many hundreds of hours invested in actually become the expert in the first place, but I doubt anyone becomes an expert so they can write $50 a pop pieces - so that cash is extra value, not a payback for the time you spent learning (as is the case with a professional's fees).
Well, OK, but that's an awfully narrow example. What happens when you are asked, or want, to write of something about which you know only a little? Or something you sort of know well but which requires a modest degree of fact-checking and verification?
Put another way, how much do you expect to earn per hour for this sort of work? How long does it take you to write an informative 300 words?
Put another way, how much do you expect to earn per hour for this sort of work? How long does it take you to write an informative 300 words?
Yes, but $50 in exchange for 300-400 words and a small slice of your credibility isn't a good trade.
However, if they don't have to put their names on the article then you're right, it is a good deal.
However, if they don't have to put their names on the article then you're right, it is a good deal.
It depends if they're writing crap that's below their station. If you're getting a heavy investigative journalist to pump out this stuff, it could hurt their credibility. If it's just a niche expert pumping out small news pieces that have some small insight behind it, I can't see how it could hurt.
It's the new domain squatting (kinda). Content producers are laying claim to keywords/searches that are under served but provide questionable value to those that they seek to reach.
Maybe eHow articles could serve as a flag to better content producers that moderately valuable yet under served "real estate" exists in a space. They might then claim some of that revenue by targeting the topic and providing higher quality content.
Maybe eHow articles could serve as a flag to better content producers that moderately valuable yet under served "real estate" exists in a space. They might then claim some of that revenue by targeting the topic and providing higher quality content.
This is at the same time really cool and horrific.
This is the modern information economy equivalent of a manic episode.
All hell will break loose when algorithms start writing these articles, imagine information just getting recycled with no new information conveyed. Of course, algorithms could detect correlations and patterns in text but would they be able to produce original and genuinely interesting articles?
Machine-generated sports stories via Northwestern's CS department: http://www.medill.northwestern.edu/innovation/page.aspx?id=1...
Paying someone a minimal amount to write about something for which they have no in-depth knowledge (i.e. they only scraped it from other online sources) is probably not much different from just having an algorithm write an article. In fact, an algorithm might do a better job.
I'm cautiously excited, but am withholding judgement until I see the quality of their products:
From the article:
"Howcast, one of Demand’s largest competitors, also produces explainer videos and how-tos. Unlike Demand, the company employs a staff of editors and writers and gets freelance voice-over pros. Filmmakers can earn a couple thousand dollars shooting the videos, and the difference is noticeable. (Howcast’s “How to Make Friends at a New School” includes such useful tidbits as “sit in the middle of the classroom to surround yourself with as many potential new friends as possible.” Demand-owned eHow’s “How to Be Popular in School” video, in contrast, offers such vague guidance as “be nice to everybody.”)"
From the article:
"Howcast, one of Demand’s largest competitors, also produces explainer videos and how-tos. Unlike Demand, the company employs a staff of editors and writers and gets freelance voice-over pros. Filmmakers can earn a couple thousand dollars shooting the videos, and the difference is noticeable. (Howcast’s “How to Make Friends at a New School” includes such useful tidbits as “sit in the middle of the classroom to surround yourself with as many potential new friends as possible.” Demand-owned eHow’s “How to Be Popular in School” video, in contrast, offers such vague guidance as “be nice to everybody.”)"
It's crap. Just take a look at eHow.com and judge for yourself.
Howcast actually seems fairly good.
I'm optimistic about eHow (and the rest of those). A mining company is happy to get 6 grams of gold out of a ton of ore. The world's internet community will collectively sift through the sewage looking for good content, and pagerank / Twitter will bring it to the rest of us.
I'm optimistic about eHow (and the rest of those). A mining company is happy to get 6 grams of gold out of a ton of ore. The world's internet community will collectively sift through the sewage looking for good content, and pagerank / Twitter will bring it to the rest of us.
> This, Reese says, is the ultimate promise of his algorithm: “You can take something that is thought of as a creative process and turn it into a manufacturing process.”
That scares the shit out of me.
That scares the shit out of me.
If it continues, this is what all music will sound like: http://www.youtube.com/watch?v=3oGFogwcx-E
See the previous discussion here:
http://news.ycombinator.com/item?id=892290
A similar, though less automated and small scale, variant of this scheme was discussed some time ago:
http://news.ycombinator.com/item?id=511935
http://news.ycombinator.com/item?id=511935
But then I observed that I was reading a more than 3000 word article, one which includes several interviews (with at least one on-site). And I wondered: is Wired itself a dinosaur? It's still apparently paying for pretty high journalistic standards. I imagine this article would have taken at least a few days to file, and presumably cost thousands of dollars. Why hasn't Wired figured out that the way to make money in online Journalism is to rip out dozens of 100 word articles per writer per day, consisting of facts plagiarized from others, with a few pithy statements and a linkbait title?
It's like Wired is some nostalgic holdover from the heady days of the print magazine, when they had people like Gibson and Stevenson writing grandiose, timeless (if a bit naive) articles on the coming digital wonderland. It made me want to search into Wired's archive and re-read Mother Earth Motherboard for the umpteenth time -- perhaps before Wired folds like the other dinosaurs and takes its archive with it. But then it occurred to me: the fact that I could, on a whim, pull up an article that was published in an ephemeral magazine more than a decade ago is itself a result of the Internet.
When that article was originally published, Wired's entire opportunity to make back the commission it paid was the shelf life of the issue in which it was printed. After that it would have spent perhaps a few years in library periodical departments, and would have eventually ended up on microfilm in archives, available to those who know the issue and volume number. I would have only been able to access their back catalog with great effort, and their ability to make money off of my doing so would have been nonexistent.
But now Wired's entire catalog is available online -- years and years of articles. And they all have fresh, revenue yielding advertisements around them. I think about how much time I've spent in Wired's archives. Just a few weeks ago I read a huge article about Project Xanadu that made the rounds. If you look at the link-broker sites like news.ycombinator, there is a good deal of blogspam -- no doubt -- but just yesterday there was a Dijkstra essay from the 1980s. It makes me think that perhaps there is a market for quality on the Internet. It's just hidden, because the quality isn't rewarded upfront by the initial wave of attention, but by the slow, slow trickle of back-issue readership -- still producing impressions on articles written and paid for many years ago.