Mar 4 2007

Another View on Widgets

David Cohen has been appropriately nudging me to write my counterpoint to his Big of Bullshit: Widgets post.  Lots has been said about widgets over the past six months, including the ultimate counterindicator article The Year of the Widget? from Newsweek (nothing against Newsweek, but whenever something like this shows up it often means we’ve hit the apex of the phenomenon.)

Lest you think Widgets are new, let’s take a short walk back in history.  Do you remember Active Desktop?  Yup – a nice Windows feature dating back to IE 4.0 that never caught on.  I bet Microsoft has a patent floating around somewhere for this, even though it would be a bogus one because of prior art.  How about DoDots (The Web without a browser) – one of our failed investments.  Similar time frame, incredible technology, clearly ahead of its time, and misunderstood (e.g. a random quote I found in an old usenet archive when poking around – “I never understood their business plan. DoDots was creating software that allowed people to embed internet-based applications on the desktop — sorry, but isn’t that a web browser?”)  Oh – and in case you happen to think Microsoft invented this – think again – they were floating around MIT / Project Athena when I was there in the mid-1980’s (anyone remember “Widget Athena Front End”?)

In the spirit of “everything old is new again”, I give you widgets as a great example.  When Konfabulator came out in 2003, I loved it.  Yahoo ended up buying it (it’s now Yahoo! Widgets) – shockingly Microsoft, Apple, and Google now have similar “widgets” (or “gadgets” or “dashboard thingys”).  All basically the same – now just an extension of the operating system.

In 2004 when I started blogging (and became a “publisher”), I discovered this really interesting thing where I could include stuff on my blog using script type=”text/javascript”.  Some folks used script src=”…” and others just gave me a smart link back to their web service.  Eventually embed’s appeared (thank you Youtube for making this radically popular.)

Like Fred Wilson, I’ve gone through a cycle of putting up and taking down widgets.  I’ve gone through my hunger for more than just dynamic information – I love the social stuff that makes me a publisher that can connect with my actual community of readers.  I got excited about TypePad’s widget implementation, fell in love with MyBlogLog, was psyched about everything FeedBurner did with widgets, and even was intrigued with Widgetbox.

I knew I was in trouble when I started getting notes from entrepreneurs saying “we are going to be the FeedBurner of widgets.”  Then – in a matter of two weeks, I got about a half dozen emails / executive summaries of things that looked like Widgetbox (yes – a couple of them got funded.)  Then – suddenly, everything I saw led with “widgets widgets widgets widgets widgets widgets.”

Enough history – so in David’s words, “is this widget thing big or bullshit?”  This time around, it’s clearly big as an “application container for publishers” but it’s not so clear that all of the derivative businesses make sense.  If you create any web-based services, you need to provide a “widget-like” application container.  Fortunately, this is trivial (in general) although it’s still much harder than it needs to be to incorporate these into all the different types of publisher sites (e.g. just try to integrate your widgets into Microsoft Spaces.) The problem here is one of standards – or lack there of (ad-hoc or otherwise) – right now it is chaos and some folks are trying to tame this.  However, the monetization strategy for taming this is thin, especially given the half-life of the vast majority of widgets (e.g. very short) combined with the overall utility of most widgets (e.g. very small.)

As an investor, I’ve looked at and decided not to invest in “widget management systems.”  However, all of the companies I’ve invested in that provide web services to publishers (including NewsGator, FeedBurner, Lijit, ClickCaster, and Me.dium) are “widgetizing” their services.  FeedBurner is an obvious platform in my universe for this (given their broad relationship with over 350,000 publishers) as they have demonstrated with their integration of Headline Animator with Stats, SpringWidgets, and Lijit Search.

While there might be room for one or two “widget management systems”, there certainly isn’t the need for 23 of them.  In addition, the ability to actually build a real business based on a packaging and distribution system around the application container widget is unclear to me.  So: widget=big; widget-derivative-business=probably-bullshit.