The personal portal battle: Google's personalized homepage vs. Start.com
I've been objectively playing with and testing the personalizable portal concepts from Google and Microsoft a lot over the last week, and I'm not alone (
Ben Askins,
Eric Hammersley, and
Adrian Sutton have a few of the better blogged tests). I've been a big fan of the web parts portal framework concept ever since I saw it in early bits for ASP.NET 2.0, and I'm glad to see some production-quality apps finally available for public consumption. Trying to use the
"Remember the Digital Dashboard?" or
"Think of Rainbow, it's kinda like that…" or
"Why you should really consider Windows SharePoint Server" examples to a neophyte audience was getting nowhere.
(If you have no idea what I'm talking about, thanks for proving my point.)
Both
Google's personalized homepage and
Microsoft's Start.com initiative exhibit slick DHTML powering animation and drag-and-drop features that are very cool, and both ship with a ton of pre-configured content (weather, stocks, national news), in addition to supporting subscriptions to external RSS feeds. Each gets a slight nod over the other for certain aspects of its shipped content (Start.com supports more global locales for weather, Google's movie service is more localized; Google makes it way easier to add content to a page, while Start.com's preview pane is a really neat way to get an abstract of a feed without leaving the page, etc.).
Surprisingly, Google's portal naively allows a user to subscribe to the same RSS feed multiple times - whether inadvertently or deliberately - while Start.com detects the presence of a specific pre-existing feed and intuitively only permits a single subscription. Start.com also supports the importation of OPML files for auto-subscription to existing lists, which Google doesn't at the moment do.
As far as speed goes and in lieu of any formal and scientific testing, Start.com, it would appear, would be the better performing service. It just seems to execute more efficiently, carrying out operations and rendering content to a page faster. Maybe this implies a superiority of the ASP.NET 2.0 engine over Java/Python.
That having been said, accessibility is the determining factor that puts some real competitive distance between the two for me, and the reason I personally prefer Google. I really dig the fact that being a true membership-oriented web app, Google's homepage enables roaming and can be accessed with your customized profile from any PC (and hopefully sometime, any device). I normally jump around on at least 8 computers of varying OS and browser at work, so it's nice to be able to get at my GMail account, search history and RSS feeds wherever I am without any additional configuration or setup, other than a one-time login. Start.com evidently uses a per-machine server-side cookie to persist customized content, so a personalized portal using Microsoft's offering at the moment can't be accessed if you use a PC adjacent to you, in another building, or on the other side of the world.
Further, because the Start.com cookie is resident only on the web browser that first set it, a customized portal can't even be accessed on a different browser on the same machine without setting it all up again from scratch. The concept of network usability is lost. This means if I setup some content areas different from the default load, add some RSS feeds, and change some program parameters in MSIE, I can't see the same page in Firefox and have to repeat the actions. And continue doing so for cross-browser maintenance over time. This is a big letdown, in my opinion.
But the main thing that brings Start.com down for me also delivers an implied benefit I think is cool. I realize not everyone has multiple PCs and workstations at their disposal, and most users are content relying on a single browser. So it's nice that Start.com supports configuration without membership via anonymous personalization, and people can setup customized experiences without having to signup for anything. This is good, and I initially predicted unavoidable, as programmatic anonymous personalization was
one of the first things I tried to achieve with the early preview bits of ASP.NET 2.0.
With both being in beta, there's still work to be done. And while increasing customer feedback and incorporation/integration into existing products lines for both Start.com with MSN and Microsoft tools and Google's page with…well, Google, the more the products begin to differentiate from one another the more they'll become even more similar.