Archive for October, 2007

Measurement “experts” huh? “LOL”

Friday, October 12th, 2007

Om Malik wrote a post last night about a Facebook traffic drop but was skeptical of the validity of the information.  Today Om explained the drop in an update post where  Paul Sutter, co-founder & president of Quantcast told him the reason was that (Om writes) "ComScore has a panel that has a bias toward Internet users who log on
from home. The same is true of all measurement panels – Nielsen,
Hitwise or Quantcast. As kids go back to school, they vanish from the

OK, makes sense.  The measurement panels don’t accurately represent the level of traffic from college users.  Yet here’s what reps from a couple of the major measurement services had to say (from Om’s post) about traffic at one of the most prominent Internet properties in the world:

Andrew Lipsman of comScore:  "Last year, there was a similar dip in visitation at
Facebook, which suggests there are some seasonal factors at play.
Another important factor that has not been mentioned is that September
has 1 fewer day than August."

Hitwise: "There is a small dip for Facebook in terms of U.S. visits
it appears for the last few weeks, but looking back over time this
could be more seasonal than anything."

So, apparently, the real reason for the traffic pattern is because of the makeup of their measurement panels.  Yet, both of these companies chalked this up to "seasonal factors."

Why is this significant?  Well, traffic numbers from firms like these are often the "facts" upon which stories are built that help define entire markets in the emerging technology world.  Yet, when they don’t have a clue about what’s going on with traffic at a company as high profile as Facebook, one has to question how valid a basis for analysis this type of "data" really is.

Semantic Web is great, but I want to DO STUFF

Wednesday, October 10th, 2007

In my "Web 1.0" days the first software company I founded, Moai Technologies, supported many of the major "online marketplaces" of the late 90’s.  One of my colleagues in the early days of Moai who helped develop our sales and marketing strategies was Stephen Bove.  Stephen ended up becoming part of the founding team at Chemdex, a leading chemicals marketplace.  I remember brainstorming w/ Stephen about the challenges at Chemdex and what he and the team there felt were some of the key issues for online marketplaces.  Perhaps the biggest common theme that applied to all of our marketplace customers was the creation of good ontologies – creating hierarchies to classify, categorize, and organize information.

I was reminded of this by Alex Iskold’s excellent post on the Structured Web.  It really feels like a decade later we are largely struggling with the same core problems.  Brad Feld posted about Alex and Adaptive Blue in this post with links to more semantic web reading.

This is good stuff, but I think a lot of technologists get caught up in the theory of all of this and forget about the fact that what people really want is to DO STUFF.  Alex clearly understands this based on what he’s doing with AdaptiveBlue – using this type of information to help people find specific pieces of information based on context of what they are looking at and doing.  While elegant theoretical constructs for all this stuff are great, I think a key to success is developing these technologies with a perspective of the types of things users and developers will actually want to do with all that information in real-world scenarios.

We’ve learned a lot of those lessons while developing ClearContext IMS over the past few years.  A lot of the "coolest" stuff we’ve developed is related to our contact and message analysis and prioritization algorithms that mine email history to rank which contacts and emails are most important to you.  However, that information only becomes really useful to users once we put it in the context of actually being able to DO SOMETHING with it – in the case of email that means creating tasks and appointments from important emails and being able to link and view all this related information together.

I definitely agree with Alex and others about the powerful possibilities of a web built upon an underlying foundation of structured information.  But without keeping in mind what people will want to do and why this is so valuable, there’s a real risk of creating some beautiful standards that largely get ignored.

PS to Alex: I suggest you check out this 2004 Santa Rita ($14.99) and this  2004 Seventy Five ($18.99)

Sigh, BoardFirst stopped

Monday, October 8th, 2007

Well, I’ve been doing a lot of writing on tech and email lately, getting ready to re-launch this blog and keep it updated in a more dedicated manner.  I’ll be starting with the serious posts soon, hopefully this week.

But I just received an email that really irritated me, which inspired my return to the blogging world.

"Dear BoardFirst Customers:

It is with the deepest regret and sense of disappointment
that I must inform you that BoardFirst is being forced to discontinue the
service it has been providing to you for the past 2 years."

The email explains that this is due to an injunction from Southwest Airlines forcing them to stop this service.

My flying these days is mainly JetBlue to New York and various airlines for short trips like LA and Vegas.  I often didn’t consider Southwest because I didn’t want to deal with the lack of assigned seating and have to worry about remembering to check-in online or arrive early.  BoardFirst solved that problem for me, resulting in me choosing Southwest for more flights.  Being able to pay a small premium to not get stuck in a really crappy seat on Southwest was a service I was glad to have.  It sucks that Southwest isn’t allowing the marketplace to provide that added value service for them.