Content Publishers: I Have A Business Model For You

Fri, Aug 14, 2009

Disruption

filter1OK, this is definitely a theory but part of me thinks it’s well worth exploring…and I’ve heard precious little about this to date.

So two experiences this week plus this Fred Wilson blog post got me thinking that content publishers are looking in the entirely wrong place for their business model. They are looking at the content.

But what if that’s the wrong approach entirely?

What if the thing you can charge for is the filter?

Let me relay my two experiences this week and then I’ll circle back to wrap things up.

Experience #1: Hacker News. IMHO, Hacker News is one of the best, if not the best, news/article filter out there. OK, maybe not for everyone but if you remove the coding and start-up articles (the focus of the community) what remains is a super-interesting collection of smart writing and interesting thinking. Kind of like what I thought Digg would ultimately become but hasn’t really (cute pictures and funny videos have largely overrun the homepage).

However, here’s my beef with Hacker News. I want to do more with it. I want to be able to see the most popular Hacker News articles of all time (the “Best” link which is buried two links in off of “Lists” in the footer is sort of this but only for the last 2 weeks). I want to see the most popular stories of all time. I also want to be able to see articles filtered in different ways (e.g., only startup, everything except “news” sites like TechCrunch, CNN, etc.). I want to be able to search in ways that I can’t unless I do some sophisticated stuff in Google.

I get it. HN isn’t set up for this. That’s fine. It’s an amazing resource the way it is. But would I pay $5/month or $50/year for a “power user” version of Hacker News. You bet.

Experience #2: Delicious. I’m learning to do more and more with Delicious these days but I’ll admit it, it’s not a user-friendly process. For instance, let’s say I want to be able to find all articles that are tagged “subscription” and “revenue” sorted by most bookmarked. Here’s the best process I was able to find this week:

Step #1: Install Greasemonkey and then install the Sort Visible Links Greasemonkey script in Firefox.

Step #2: Search by multiple tags in Delicious. This isn’t super difficult (Here’s the query: http://delicious.com/tag/subscription+revenue).

Step #3: Sort the results based off of the number of bookmarks using Sort Visible Links.

However, what happens when I have, say, 2,500 results? Well then I have to set Delicious to show 100 links per page (the max) and then I have to page through 25 pages of results and do a sort on every page. That = Royal. Pain. In. The. Ass. Did this for one query this week and it probably took me two hours to go through what I should have been able to do in about 10 minutes. That sucks.

Would I pay delicious a bit of money to make this stuff easier? I would. Absolutely.

filter2So the big question is whether I am just a strange edge case? I realize that I probably am. But guess what…so were the people trying to upload video to the Web and show it to everybody in 2003 before YouTube came along. Once YouTube made it easy for people to upload and share video that space exploded.

What we have right now, to steal Clay Shirky’s words is not information overload. It’s filter failure. We have an increasing amount of info, links, etc. being thrown at us through stuff like Facebook, Twitter, Digg, etc. and fairly crappy ways to get to what’s most important. And here’s where the publishers have an amazing opportunity to help us.

Build robust filters.

They have ridiculous amounts of data on what peoples’ preferences are. They have internal metrics out of the whazoo. They have the potential to help us get to the best and most relevant content. And what have they done? Largely outsourced the process of building content filters to Digg, Reddit, delicious, Hacker News and other folks.

I love those guys. But the problem is that delicious knows a heck of a lot less a specific piece of content than the content owner does. New York Times could do an insanely good job of giving me personalization and robust filtering. And I’d be willing to pay for that. Instead, they do an inane job of this. And therefore I don’t pay them. And I filter their ads. At least 5% of Internet users (FF users) do the same thing and I wouldn’t be surprised if 50% or more of Internet users do that within 3 years. (Camino is my half-time browser and it sports a freaking checkbox that blocks all web ads.)

So here’s the deal. Very few people will pay for your content (and increasingly fewer). And while more people will view your ads that number will fall over time as will CPMs likely. So how do you make money? By creating services that people will pay for (like next-gen filtering and personalization) and that don’t subtract value in any way (unlike pay walls which subtract huge amounts of value). These services will produce revenue that drops completely to the bottom line which is dramatically different from advertising sales which often have gross margins of less than 50% once the cost of selling the ads is factored in *and* detracts from the user experience.

OK, bonus round. Don’t even build the filters in-house. Do the Netflix Prize thing and outsource the building of the filters and recommendations to smart, hungry (and often young) hackers. Turn over your data to them (anonymously of course) and let them build the engines that can help you reinvent your industry.

filter3Sound risky? Damn right. But what do you really have to lose? Doing this adds money directly to your bottom line if people pay money for premium services and at a minimum makes your service a better user experience if they don’t pay. It allows you to understand better why people read and watch what they do and as we become an increasingly information-saturated society it allows you to skate to where the puck is going vs. skating to the where the puck is.

Humor us content producers. Start the revolution this time rather than being a victim of it.

This post was written by:

Jon - who has written 27 posts on JonBischke.com.


Contact the author

  • Interesting article. My e-mail is andreaswagner11@centrum.sk
  • Hmm, thank you for good article. I wonder if you will write more like this...if my contact e-mail andreas@gigacaststatic.com
  • Makes very much sense!
  • At the OpenViewing Foundation (www.openviewing.org), we are implementing your approach by creating/adopting/opening up protocols for the publishing of video viewing metadata, both better structural data as well as anonymous behavioural data.

    We believe that one approach is to have all content providers start to publish anonymous behaviour information so that third party filters can do a better job than they can with things like tags. Whether those filters are bookmarking services, search engines, or domain specific recommendation engines, there are many potential consumers of such enhanced data.

    If you step back for one second, wouldn't it be better if you had an engine that helped filter not just the NYT, but also the WSJ, or whatever set of online news sites you read?

    The challenge of course is how and why you publish this metadata. Content providers have to decide that the connections this type of open sharing would create will lead to innovation and more demand for their content (assuming their content is good, of course the bad content would always be filtered).

    Google has standardised what metadata is important based on pagerank etc. but the real questions is how do we start creating more useful metadata, in an open, authoritative way without leaving the system open to abuse. ala META tags of the 90's....
  • @Paul - Very cool. I haven't used StackOverflow yet but have heard good things. I'd love to spend some more time looking into how they've architected their API to support the other sites you reference (if indeed that's what's happening here).

    @run4yourlives - You're my hero of the day. But that also kind of proves my point. I've been reading Hacker News for six months and only now do I know that exists. :) It seems like there needs to be a better way to let people know that this stuff is out there. Similarly, it took me a while to figure out how to do the delicious thing I referenced above.

    @gordonmattey - Very, very cool. Will check out. I definitely think a metafilter would be better. However, as you state, the question is whether you could get all these companies to share data in a meaningful way. That's why my initial recommendation was for better filters within the various silos. I think your approach is more ideal but also more difficult to pull off.
  • If I recall correctly, NYT has adopted semantic web standards to publish their data in RDF formats, which could be integrated with semantic web applications and the Linked Data movement in the future. Hopefully, others will join the bandwagon - there is clearly value is making your data "share-able" assuming you still have control over how it is used.
  • run4yourlives
    http://top.searchyc.com/ - Your first issue is solved.
  • your post reminded me of stackoverflow.com - it's filtered by tags and you can search on it and now it's branched into serverfault.com, meta.stackoverflow.com and superuser.com - I think it's going to be the poster child of the Post-Search Web :)
  • Great post Jon! Are you thinking something like parse.ly?
  • @Name: I don't think it's going to be subscription-based. Content publishers have been playing with that for 10 years and haven't figured out a way to make it work. I don't see what's going to change for them on that front.

    @Erik: Sort of. But I want the content producers to be involved. NYT, WSJ, etc. have a huge amount of data that is closed off to the rest of the world about what their readers are doing on their sites. And they're doing very little in terms of making that data useful for filtering, recommendations, personalization, etc. My idea is to turn over that data (much the way Netflix has done) and allow people build filters and personalization on top of the data.

    Twitter has exploded because of their API. Imagine where Twitter would be without it.

    Major content publishers need to take the same approach. That or they'll continue to fall like dominoes.
  • Name
    Interesting article. Good ideas. Question. This seems to focus on increasing revenue for sites that filter from various news sites. But before investing too much in that, don't the news sites themselves need to improve their model so they are making money? And if that model ends-up being subscription based, does that do away with the filter sites? I'd be interested in your thoughts on that.
  • Good one, Jon.

    One of my blog readers suggested something similar for educational content. A search engine that build courses based on the needs and objectives of the student.

    Any replies from Rupert yet? ;)
blog comments powered by Disqus