Mar 15 2007

YAWP – Yet Another Widget Post

While the SXSW crowd is a-twitter, I’m still thinking about the range of private emails I’ve gotten about my widget post.  While some of the comments were assertions that I’m an idiot, there were a few that caused me to think harder about the opportunities (or non-opportunities) to build real businesses around the construct of widgets.  In addition, Mike Hirshland wrote a post challenging several people (including me) to debate whether or not someone could build a company around “widget management systems” that could generate venture returns and one of Mike’s partner at Polaris – Sim Simeonov – wrote an extensive post titled Widgets, Widgets, Everywhere on his view of this.

In the emails I got, several people misinterpreted my point of view so I figured I’d start with a quick summary.  I think widgets are an incredible distribution mechanism for web-based functionality.  I love widgets – and love the widgetization of the web.  However, I’m struggling to see where the real business opportunities are in “wrangling widgets” (or – more simply – “widget management systems and infrastructure.”)

Whenever I think of something like this, I start with the assumption that there are no fundamentally new business models.  In my experience, almost every business I’ve ever seen on the Internet either has a non-Internet real world analogy (an analog analogue), a non-Internet software analogy, or a previous incarnation on the Internet in some previous phase (Web 2.0 anyone?)  While the technology, implementation, distribution, user-interface, and infrastructure dynamics change, the business models rarely do.

To me, widgets are application packagers that enable you to embed specific functionality from a web site (or web service) into another web site.  Sim provided a similar view from a slightly different angle in Widgets, Widgets, Everywhere by saying “widgets are the next step in the trend towards disaggregation of content at the production end and aggregation of content by the consumer.”  My view of widgets today is that they are tightly coupled with the long tail – adoption seems exciting (and pageview (widgetview?) ramps seem huge) because of the long tail effect (vs. moving up the curve to mainstream media.)

When I stare at this and think about the different ways to build businesses to support this, I come up with four business models:

  1. a new form of ad network: analogous to DoubleClick
  2. a widget management system (WMS): analogous to CMS’s
  3. a content distribution network (CDN): analogous to Akamai
  4. an analytics business (Stats): analogous to pick-your-analytics package

I don’t buy that #1 (ad networks) is a big moneymaker.  Several people hit me over the head with a brick and have said “widgets will generate 10 zillion ad unit page views per day.”  I’m struggling with this – you’ve got issues around form factor, content rights, revenue share, long tail critical mass, and many incumbent ad networks that are dealing with different approaches to widgets.  Maybe there’s room for one new one to emerge – the way FeedBurner emerged around feed-based advertising, but that begs another question which I’ll leave for you to ponder.

#2 (WMS) is great and helpful to me as a publisher, but I don’t know how to monetize it.  I’d love a single WMS / widget container for all my widgets.  If I’m a Typepad user, I’ve already got this and I assume WordPress.com will nail this also.  I expect the CMS platforms can and should quickly integrate this – it’s not technically difficult and continues to improve the value of their CMS.  As a blogger, I’d be hard pressed to pay an incremental fee to a widget CMS company (I’d do a rev share on #1, but I don’t know where the ads would go!)  Most of the stuff I’ve seen fits in this category and I’m impressed with what a few of them are doing functionally, but I’m still struggling with “turn the money on.”

#3 (CDN) seems like there should be something interesting there.  I’m already having periodic performance issues with some of my widgets and if you’ve been following along with Fred “the king of blog bling” Wilson’s performance issues, you will appreciate that a real CDN could be helpful.  However, the long tail bites (or “tangles”) again – the only people that will pay for this are at the head of it so the real network effect doesn’t pay off big here. 

#4 (Stats) just won’t work.  Google commoditized the long tail stats market when they bought Urchin and made it free.  We already saw this with feed stats – the vast majority of them are now free. I believe long tail stats – like Wikipedia (and – according to the FSF – software) – really want to be free.

If you think I’m missing something – obvious or otherwise – I’m all ears.  And – if you think I’m really wrong, please keep hitting me over the head with a brick.