Posted by: structureofnews | August 17, 2010

A Little Knowledge…

Off the point of Structured Journalism, but on the point of business models, some comments on this post that Ross sent me.

Basically, it’s about Demand Media and how, by algorithmically determining what people want to read, the company can predict associated ad revenue and build content to spec and cost – hence turning in some nice profits. It’s an interesting concept, although not necessarily directly related to journalism business models – but gives some interesting insights nonetheless on the notion of demand-led content and predictive models for ad revenue.

First, a side comment: At $15 a pop for content, frankly, Demand Media would be profitable regardless of what algorithm they use. Low content cost is probably a big part of their business model, never mind predictability of ad revenue. But of course being able to predict user demand means they waste less on content generation and drive margins up. But still.

More broadly, the question of being able to accurately predict ad revenue is an interesting one. Here’s a thought experiment: If we take the view that a journalism outfit wouldn’t want to produce only things that people are searching for, because there’s a public interest need, or a corporate mission, or a burning individual passion to cover something (not unlike an artist’s decision to create a piece of art he wants to, rather than what will sell), then let’s assume for the moment that the newsroom would still just work on what they want to work on. But what if the predictive model could tell you accurately which of those stories would make money and how much? How much would or should that change newsroom behavior?

What if we knew to the dollar how much each movie would make, before we made it? It doesn’t mean that we’d only see Transformers 5 in cinemas; lots of small indie movies make money too, and presumably would continue to be made. Will mediocre movies die out? Will people subsidize pet projects they know would be money-losers with more profitable ventures?

I don’t know, but it seems an interesting universe to inhabit.

That said, it’s a fair point to ask if journalists should just do what they want to do and not be influenced by the wishes of their readers. Of course they should know what people want – and that’s the point that Jay Rosen and others have made. A little knowledge is an important thing when it comes to reader metrics – even if journalists don’t usually like the answer.

But reader metrics aren’t everything either. And nor are algorithms, which nearly by definition measure what they’ve been tasked to measure. Google’s great, but at the end of the day it is a highly-tweaked engine that matches up some sets of human behaviors with word patterns and all that.

But that is useful knowledge – as long as we remember that there are limits to its wisdom.


Responses

  1. From Keith Winstein, and my responses below:

    Keith Winstein: If the WSJ only wrote about the stories that ginned up a lot of readers on the Web site, the paper would just be about puppies, new Apple products, and the funerals of prominent liberal politicians.

    Reg Chua: Of course, that’s just a thought experiment – but in any case, we do have successful indie movies and indie bands, and summer blockbusters that go bust, but it’s not always true that having the broadest mass appeal is the smartest or most profitable way to go. Still, it would be an interesting world (or maybe a terribly dull world) if we knew before we did anything if it would be successful or not…

    Keith Winstein: Reg, of course you’re right, and I was trying to be funny with my reference to puppies and Apple products. But I’m not sure the question’s implication about the appropriate way to value “how much money any story would make” is healthy.

    To th…e extent we’re still trying to get readers to believe their subscriptions are valuable — and to get more people to subscribe — I don’t think the value of a single news story is well-defined. Readers don’t decide to subscribe or cancel their subscriptions based on one story, and the contribution of each story to that overall decision isn’t unitary or orthogonal.

    The bargain is that the reader subscribes and we give them a palette of edifying material we’ve carefully selected and recommend that they read. That’s not a service whose value can easily be quantified story-by-story.

    Movies really do live or die on their own publicity (nobody pays Paramount Pictures a monthly check to provide them with a steady well-rounded diet of cinema), so maybe that’s part of my concern with the analogy.

    Meanwhile, for news outlets that survive solely on clicks and ad revenue (as opposed to subscriptions), the decision about whether to read a particular story or not is generally made by exactly the group of people least qualified to make it — that is, those who haven’t read it yet.🙂

    Empirically, the ad-driven news outlets in possession of the best data about reader/viewer behavior — e.g. CNN, cnn.com, Gawker — are also the crappiest. CNN and its competitors, for example, have the benefit of analyzing viewer behavior minute-by-minute. Their erudite, well-informed conclusion: the most profitable programming is cheap-to-produce live commentary that plays on the news uncovered by others about celebrities and hot-button issues. That’s not good.

    With the benefit of hindsight, would anybody have greenlit my four months spent getting an article in the paper in 2008 proving that Boston Scientific miscalculated the result of a high-stakes medical trial, incorrectly announcing the results as a success when it was really a failure? I doubt it.

    My concern is that if the paper really did try to evaluate value on a story-by-story basis, nobody would rationally invest in long-term watchdog-style journalism about semi-obscure topics. That’s really what I was getting at with my joke about puppy stories.

  2. And another response from me:

    Reg Chua:
    Keith, thanks for the comments, which are very insightful. The subscription/continuing goods and single product distinction is one… made very well in The Curse of the Mogul. I’m certainly not advocating a shift to valuing each story – I was commenting on Demand Media’s model. But even if we don’t like it – and even if we don’t change to this brave new world – the information about the ad-supported value/traffic on any story is rapidly becoming available. The question is what we do with it.

  3. […] circling back to an earlier post that mentioned Demand Media and its model of predicting what people are searching for and creating […]

  4. […] files (unless we’re Demand Media, who pay by the piece, but that’s another post.  Or two.)  So you can imagine what a newsroom does is file a series of these stories, some with higher […]


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Categories

%d bloggers like this: