January 28, 2005

Amazon.com: Yellow Pages: Pacmed Cafe

So that's what those folks were doing the whole time... Interesting.

Here's a shot of the Amazon castle on the hill. Sometimes I go up there for a cheeseburger.
Amazon.com: Yellow Pages: Pacmed Cafe

I tried getting the URI for Amazon HQ, but it resolved to the sample center down on Rainier Avenue - not nearly as picturesque as the PacMed building.

Coolio - here's the coffee shop I walk to from work. I wonder if I could reference an individual image from the block shots they have. Using just the ASIN, I'm at the mercy of the community voting on the 'best' image. I'd like to piggyback on the data and have a direct URI to that block shot.

January 24, 2005

Gordon Mohr's Theory of Wiki

Gordon asks

Is there any information domain where the open, world-writable wiki model can't be beneficially applied?


I think the answer is clearly "yes" and is based on the meaning and usage of the information. When the information domain has an objective authority, then community consensus is harmful. For example, legal contracts come from a single source - the participants. Although it would be useful for me to modify the terms of my car loan ("I'm going to write me up a minivan!") whenever it was convenient, I doubt that will ever happen. Another example would be prices of items listed for sale, which is in essence a contract for an 'offer to sell'. There is a clear authority for the offering price - the seller of the item. This shouldn't be confused with the 'going market price' for an item. That kind of price has no objective authority and actually seems to be defined by community consensus itself.

Some data in an information domain is synthesized or summarized from a set of other data - for example, the market price of an item isn't an actual amount offered by any indivdual to buy something. I wonder if a world-writable wiki is beneficial only in areas where meaning can be preserved by selecting one value from many choices.

January 21, 2005

Google to provide AdWords API to Advertisers

This statement from Silicon Valley Watcher (via The Now Economy) caught my eye -

It is but a short step from the global delivery of simple text ads to carrying commercial transactions also. This would pitch it against companies such as eBay and other online retailers.


I don't belive it is a short step, but it does seem like the next logical step for Google and Froogle.

Google offers no controls over ad content and where it is delivered. It just gets scattered out through its network chasing keywords in web pages. It's inefficient, but, surely Google can just take notes and implement similar ad services packages and controls?

“They won’t because it’s not in their genes, they are engineers,” Podell says.


That's funny - they are engineers so they won't change? Google is all about relevance and once they figure out that what the market wants - topical relevance, audience relevance, or even relevance to the seller's profit motive - then they'll change.

January 19, 2005

Google's new link property

Scoble asks an interesting question about Google's new link attribute -

Oh, and, did anyone notice how Google got its competitors to do something without needing to get a standards committee involved? All within hours?
Hmm, why doesn't everything in the industry work like this?


The answers are...

  • because the industry doesn't (yet) use hypertext as the engine of application state.
  • actually, the industry that matters does work like this.


      Also - how do you know they don't follow the link? This is just a negative score value stored and used in in their weighting algorithm. Eventually you could find things people explicitly don't want found.

January 12, 2005

Blog search engine bookmarklets

Jeff has a few bookmarklets that show how search engines can use REST to enable decentralized integration - and I bet they didn't know that's what they were doing...

Jeff Sandquist - Microsoft Evangelist - Technorati, Bloglines Citations and Feedster Shootout!

This is a nice extension to the capabilities summarized in my article a while back - I wonder if they had that all along and I just missed it.

January 09, 2005

Payphone Project - a photoset on Flickr

This is bizarre - a payphone project on Flickr that has photos of payphones. These are in Hawaii.
Payphone Project - a photoset on Flickr

I can just imagine the caller-id of the future with instant access to phone number keyed photos, not just of payphones, but of people (and businesses). I bet someone out there already has a patent on that... but do they do a lookup on the public Web?

And also a next-gen email client that slides in a photo of the sender, because I'm terrible at remembering names or the opaque email addresses many people seem to have. Sure, LDAP was supposed to do that, but it didn't.

Amazon sets new record over Thanksgiving weekend

From itfacts.biz

Internet's largest retailer Amazon.com said that for the first time sales of electronics surpassed the sales of books on its Web site. During that weekend 2.8 mln units were ordered within a single day, or 1 item was sold every 32 seconds. During the busiest hour, Amazon received 700,000 unique visitors to the site.


I think they did the math wrong - 2.8M per day is 32 items ordered per second. That's a lot of orders. (The link to the news story has it right.)

E-commerce > IT Facts

From itfacts.biz - an interesting page E-commerce statistics and links to some supporting material.

Some listed items

  • $100 bil will be spent on US e-commerce in 2003
  • $12.5 bln in online payments in 2005
  • $12.8 bln will be spent on Valentine Day gifts
  • $13.2 bln will be spent online during 2004 holiday season
  • $22.34 bln in online sales in Q4 2004
  • $234 mln charged to Visa cards on Black Friday
  • $7.5 bln of apparel and accessories to sell online in 2004
  • 1.4 mln e-books sold in the US in 2003


Tapping into Tags

Jeremy points to Steven Rubel suggestion of using tags and Flickr for subscribing to product and brand awareness -
Tapping into Tags (by Jeremy Zawodny)

I've been following tag based system for a while (see my post from Nov
) but I'm not sure how they will react to spam. It looks like flickr is using voting to keep things clean, but I haven't spent much time investigating.

(Thanks for the link to folksonomies, haven't read it yet but will look soon!)

January 08, 2005

Product authority and beyond

This is a very good read on community, product authority and beyond

But ultimately, if I didn’t feel confident that I could assess empirically the good I was purchasing, I would not buy it on eBay, even if the seller had great feedback. At least for me, the risk is simply too high. That’s why there are many items I would never buy on eBay.


The full article goes beyond product authority, but the beginning does a great job of identifying the key characteristic of 'product authority' - whether two items are equivalent. When two sellers offer something, there is great value in knowing that it is the same thing. This enables several capabilities - price comparison for example. Of course, sellers don't particularly enjoy competition and you'll find subtle differences in many nearly equivalent products. Over the holidays we had four families in our house, and each of us had the same camera - which made it easy for anyone to pick up any camera and take a family shot - except the model number was different and some minor features were different. This separation based on product authority enabled a capability that is of interest to seller - differential pricing in order to capture the consumer surplus.

In the ontology (there's that word again) of selling, there are two main branches 'goods' and 'services'. Product authority so far has dealt with goods, and the ecommerce industry's expansion into services poses questions that the Corante article discusses. Check it out.

metaAmazon

My good friend Adam (name dropping in the blogosphere - how cool is that?) has an interesting piece on the vision of personal commerce, the 'long tail of ecommerce' and Amazon:

Amazon already takes me part of the way there today: by being an affiliate I can recommend books and music and get a cut of anything that Amazon the merchant sells through my referrals. But taken to the extreme, what if Amazon ceased to be a merchant and instead enabled me to be the merchant using their infrastructure? The benefit to me is that I can choose only "Rifkin-recommended" products, pricing, policies, catablogs, etc. -- and not just me.

Every person could become an Amazon.


January 01, 2005

Five things it's probably better not to do when you're kind of drunk

Good thing I saw this in time... (although that peer-to-peer based 'pay-someone-elses-bills.net' Web service embedded into the bios of every handheld computer still sounds good according to my childhood friend the Somoan lawyer...)

5ives - Five things it's probably better not to do when you're kind of drunk

  1. buy domain names
  2. hire an attorney
  3. do lots of file management from the command line
  4. sort out your finances
  5. telephone people you remember fondly from elementary school

Krzysztof Kowalczyk weblog

From the weblog of Krzysztof Kowalczyk:
Mike, who works for Amazon, defends Amazon but I don't like his spin. First, he denigrates the value of the source code and elevates the value of Amazon's web services and data they provide.


Well, I'm very sorry if what I wrote sounded like a negative judgement of the value of software itself, I just believe that the source code that we work with really wouldn't be of any value to an outside party. This is not true of all software in general, just the work that my group at Amazon does to collect and unify product information across all sellers. I think the software written at Amazon is well done, professional and very educational (to me anyway). However, when I imagine handing over this software based system to some other organization, I really don't see that it would be successful. The value come from large volumes of data. Perhaps some mundane aspects like hooking into CORBA portable object adapters for performance profiling and logging purposes might be of benefit - but who in the open source community really is using CORBA? It's too standardized and stable to attract the run of the mill coding cowboy nowadays anyway. (okay, that was a little flamebait, but it's New Years, why not have a little fun?)