ecommerce was Re: [uf-discuss] Principles of Microformats?

[Benjamin West]

Mike,
Nice reply.

> > I've also noticed that many of the more successful
> > technologies I can think of first implemented use cases with
> > user-centric data: people, places, things, times, and events.
>
> I don't know that to be true, but it certainly makes sense.
>
I don't know if it's true either.  Tantek suggests I'm being a bad
scientist by allowing myself to look for patterns.  Nonetheless, I've
been working on investigating it more.  I call this data
"environmental data types" because their concrete place in the world
we evolved in.  Anyone have suggestions on putting this to the test?

> > That doesn't mean other use cases aren't of interest to the
> > community.  It simply means that time and energy are limited,
> > and people tend to spend most of it on things they are good at.
>
> Correct. The problem (as I've seen it) is the vision and process for
> microformats inhibits addressing those other issues.  Again, this is just an
> observation, I am explicitly no longer advocating they change it.

I respectfully disagree.  I'd like to start investigating the use
cases your business had trouble fulfilling, starting with: are the
problems you described with vendor communication common?

> > Can you elaborate a bit more on these kinds of use cases?
> > Are there some basic categorizations of ecommerce?  What are
> > the common things sites need to do? Where and how do they
> > need to talk with other systems?  High level answers are good.
>
> Let me give you a real world example.  I used to run Xtras.Net
> (www.xtras.net)  We sold products from companies like www.infragistics.com
> and www.componentone.com, but at certain points we dealt with well over 500
> different vendors. Had we need able to manage the logistics, we could have
> dealt with well 2500 vendors and our sales would have increased
> significantly.  Realize that most of these vendors were "starving artists",
> i.e. one or two man shops making less than US$100k/year in revenue.  They
> certainly were not going to be buying any ecommerce infrastructure, but they
> did update their websites whenever they had a new version.
>
> If an appropriate "semantic markup" could have been defined we might have
> been able to get most of those vendors to apply it and then we could have
> run crawlers to get a lot of our data. I actually had this vision in 1997
> when I first heard of XML, but for many reasons was never able to make it a
> reality (many reasons=lack of capital to fund the development :)  It would
> be that much easier with semantic markup with today scripting languages and
> other tools.
>
> But realize that Xtras.Net's business volume was a gnat on the back of a dog
> in a car on a ship crossing the Pacific ocean when compared to the world
> economy.  So take all the other gnats, and dogs and cars and ships and have
> them all start creating their own industry specific semantic markup and BAM;
> you got some serious eff-ing chaos, and I declare that will be a Bad
> Thing(tm) for the web.
>

Mike, this is real good stuff.  Can you continue elaborating?  The
thing I'm hearing is:
"We had trouble keeping our product list in sync with our vendors'
lists, even though they had it published on their own website."  This
is the kind of problem statement we know how to handle.  There are
some questions to answer:
1.) How many ecommerce sites have their inventory published on the web?
   - my guess is nearly all of them, although I suspect there are some
confounding factors.
2.) How many sellers have trouble getting product information from
their vendors?
  - this one kind of surprises me, I thought everyone had at least a
php script hooked up to their database to produce a CSV.
3.) How many esellers use vendors to populate a part of their online offerings?
4.) Would hlisting <http://microformats.org/wiki/hlisting> solve this
particular use case?

Finally,
What other challenges did you face/are aware of?

Ben