January 1st, 2008

Can Pay-For-Performance Improve The Quality Of Content On The Web?


Nick Denton and Gawker Media are wrestling with the problem of content quality on the web — specifically, how to give bloggers incentives to create content that drives traffic based on quality rather than quantity. Gawker has announced that incentive pay for its bloggers will now be based entirely on the number of page views that each blogger’s posts generate, rather than on the total number of posts a blogger writes. (Vallewag has posted the entire internal memo, which anyone interested in the economics of publishing on the web should read and re-read.)

The downsides of this approach are obvious — the incentive rewards content that is salacious, titillating, slanderous, nasty, etc. — anything that appeals to the base interests of a mass audience. It rewards gaming of social news sites, i.e. creating content that appeals to the most parochial interests of users on Digg, Reddit, StumbleUpon, etc. And of course it rewards search engine optimization, writing content that is packed with keywords and that foots to the top searches, with headlines written for search engines rather than people.

Google AdSense is in many ways the most successful pay-for-performance content system on the web, rewarding content that ranks high in search and draws in readers who then click on ads. It’s easy to argue that this system has flooded the web with junk “Made for AdSense” content that has no purpose other than to generate ad revenue.

But the flip side is that Google AdSense has also financed innovation on the web, giving web start-ups and niche content creators an easy stream of revenue to help bootstrap new sites. AdSense may not be a long-term business model, but it does reward innovative new sites that can attract attention.

So what about Gawker? Nick seems to believe that pay-for-performance will lead to an overall improvement in the quality of Gawker content. He uses the adjective “linkworthy” to describe content that is of sufficiently high quality that other sites will link to it. Of course, the same salacious, titillating, slanderous, etc. content can also be linkworthy, but I think Nick is less focused on the problem of offensive or “cheap thrills” content, and more on the problem of mediocre content.

When blog networks like Gawker paid writers based on the number of posts, they provided an incentive for writers to post even when they didn’t have any interesting information or anything interesting to say. The result of this “infinite news hole” was that blog networks generated a lot of mediocre content.

You could argue that mediocrity is the real scourge of content the web, and that more “linkworthy” content means, on balance, higher quality content. Links are of course the principal driver of search engine ranking, so writing linkworthy content is ultimately the best SEO strategy. Search can and does reward the best content as easily as it can reward poor quality content engineered to game the system.

So in the final analysis, does pay-for-performance create incentives for better or worse content? I think the answer is both. The web in many ways turns a blind eye to quality — it rewards both the good and the bad.

What the web lacks most right now is a content filter that adheres consistently to a high standard of quality. If their were such a content aggregation system, it might be possible to significantly improve the quality of content on the web with pay-for-performance systems.

  • Scott, as you note, “[w]hat the web lacks most right now is a content filter that adheres consistently to a high standard of quality.” I think that’s absolutely right. Fundamentally, it’s why we’re building Brijit. But we don’t believe paying your writers based on traffic is the way ensure quality. Combining the best aspects of algorithm, user-generated content and traditional editorial control allows us to control for quality. Pay-for-traffic, more often than not, is a race to the bottom.

    More here: http://brijit.wordpress.com/

  • Freelancers who win cover stories for print magazines don't automatically have an advantage over paid staffers. An editor decides what's worthy and puts his/her neck out to fight for great work deemed so. Lazy or indifferent staffers can be scooped at their own publications. I've seen it firsthand.

    How can clicks and links provide a meaningful metric for quality web content? They encourage manipulation for the sake of circulation and ad revenue.

    Measure the length of visitor stays. Measure the quality of visitor paths. Analyze who among your contributors is most read. Measure ad click-throughs (novel concept). Then factor in raw visits and figure out pay.

    What's the difference between a 5-second web visit and a 5-minute web visit? A video rant? A compelling post that breaks news? A headline written purely to grab attention for a moment? Tough dilemma. But the editor shares in the burden. Temptation is to go for the easy way. I can't see that as a long-term business model.

    Responsibility, respect and trust are part of publishing -- no matter where. You have to consider the people who produce the work, the customers who read it, and the advertisers trying to sell amidst it.

  • Goto Ao

    Scott Karp states the following at the end of his piece:

    "What the web lacks most right now is a content filter that adheres consistently to a high standard of quality. If their were such a content aggregation system, it might be possible to significantly improve the quality of content on the web with pay-for-performance systems."

    What Mr. Karp apparently does not know (or chooses not to know)is that such a "content filter" (or rather and more properly an "information utility") with the quality of Web content that Scott rightly wants, is here and available now and has been here for almost a decade layered across the public Internet in the form if "Interchangeable Master Channels" (IMCs) that are basically a private commercial Internet with a built in "filter" called a quality driven marketplace.

    Scott was offered access (does he not now remember this?) many months ago to the IMC network whch has more than 700 dedicated IMCs in the dot-com name space that (a) bypass ICANN for practical purposes and b) can acommodate almost every conceivable brand name product or service, where the brand name sponsors then become the enablers for the "high standard of quality" criterion for content aggregation that Scott evidently (and rightly) wants to achieve.

    And, not incidentally, the network absolutely protects the PRIVACY of every single person who intends to receive such content on any consistent basis (RSS and so forth), the latter about which Scott has also written.

    Nor is Scott the only person who has access to the network. Anybody can access the network. It appears to be the case that persons like Scott who really are rather knowledgeable about such things, sometimes fail to see the forest for the trees - perhaps because they are not the ones who actually invent the solutions about which they now complain are unavailable?

    All Scott has to do is ask and his really rather meritorious and above-mentioned complaint about quality filtering is, for practical purposes, alleviated.

    I'm not saying that the EXISTING network is perfect, but the original concept which contemplated a UNIFORM AND COHERENT NETWORK with 700 plus dedicated channels in the dot-com name space was and is sufficient to accommodate all content considered worthy by in all probability any brand name sponsor anywhere in the word that wants to sponsor quality content.

    Think of it as a Web-based PBS, with the difference bring that content creators work with their sponsors while the network itself is laissez faire in the sense that it anables connectivity in the rott zone but otherwise totally "hands off".

    Sure, content filtering raises questions about who decides what is the quality criterion that triggers the filter. The best I could come up with in the original design was the marketplace. Sleazy and meritless content presumably would not easily find sponsorship - as if, for example, I wanted to program a Howard Sern nudity festival on PBS and tried to find a sponsor.

    I admire Scott very much. And I also think his would an excellent voice to have present in the management of the network. Nonetheless, I do think on this occasion that Scott protests too much.

  • The problem is the metrics being used to define success. Number of posts? Page Views? These are not metrics built to last in publishing 2.0 let alone 3.0.

    One we start to look at outcome based metrics to quantify and redefine what success means to advertisers this all gets sorted out.

  • M

    I think a policy change like this can also mean that mroe bloggers will have to hustle to make more connections, network, sell themselves outside their online contributions. I know a lot of bloggers focus as much on relationships with real people who can swap links or just help sell the blogger. I don't know whether that's good or bad for content. Or whether that's fair to other bloggers who may not have the personalities to network with real, live people.

blog comments powered by Disqus


Receive new posts by email

Recent Posts