Looking for moonbows in California

Keywords: searchsig , Search , E-Commerce , Edgeio , eBay , tag , Web2.0 ,listing

I went to the SD Forum Search Sig meeting on Classified Search last week.

As usual there was a pretty good panel:

• Bindu Reddy, Product Manager GoogleBase, Google
• Craig Donato, Founder/CEO, Oodle
• Keith Teare, Founder/CEO, Edgeio
• Greg Sterling, Managing Editor, The Kelsey Group (moderator)

I heard about Edgeio (Edge In Out) before and I was really looking forward the live demo.

The first question Greg asked was actually a question to the audience: what website did you use the last time to posted a classified ad? Interestingly enough the most common answers were eBay and Craigslist.

So the next question was "is there any essential characteristics that differentiate classifieds ads from other types of online advertising"? The most insightful answer was from Craig who see some specific characteristics:

• perishable "unit of one" (for example a apartment. Once it is gone, it is gone for ever (well at least for a while)) -> time aspect is important
  
• start online and finish online -> this brings some social aspect
  
• local/geography aspect (for an apartment or to complete the transaction off-line you need to be in the same area)

The reason I am mentioning this here is that it seems to me that Edgeio took all those characteristics into consideration when thinking about the user experience of their platform.

Edgeio as a seller

First let's see how Edgeio works for a seller:

1. You post a classified ad on your site and you use the listing tag to let Edgeio know that this is something they need to pick-up. Keith used his lexus as an example.
   
2. When you submit your post on your favorite blog platform, it will contact one or several ping servers to let the world know that you posted something new.
3. Edgeio back-end monitor those ping servers and will look only at the posts using the listing tag to process them.
4. Edgeio will use the trackback system supported by most of the blogging platforms to let you know that your blog has been pickup by them (so far, nobody seems to have noted that this is the ultimate viral marketing technique!!!)
5. You can then go to the Edgeio website to claim your blog (well if this is your first time). The most interesting step in the process is to GeoTag your site so that all your listings will have some location information attached to them. This cover the local characteristic.
6. You can also edit the ad directly on Edgeio website. For example you can add some new tags.
   
7. Wait and see people contact you!
   

Edgeio as a buyer

From what I understand this is all things would work for a buyer:

1. You can do a search on the listings indexed by Edgeio. The most interesting thing is that the geography parameter can be selected with a pretty cool slider (you can start from a city like Palo Alto to the whole world). As you use the slide to narrow down your search to a specific city you will see less and less results. (I think that 2006 will be the year of the slider and 2007 will be the year of the pivot, more on this on another post)
   
2. You can save a search so that you can be notified when something that meet your search query pop up. Similar to eBay Favorites Searches or in a broader context PubSub. This cover the time aspect of classified.
   

The big question I had in my mind before the meeting was how would they approach the trust challenge. I was suspecting that they would argue that since the ads are posted from a blog you just need to read the blog to learn about the seller.

Well, what they are working on is a list of profiles that you can link to your Edgeio account so that potential buyers can learn more about your reputation. In the demo, Keith's account was linked to an eBay account (e-commerce transaction history) and to a Linkedin account (who do you know). Keith also said he was a big believer in portable reputation systems and was exploring a few things with Opinity.

What is the value proposition?

The value proposition for the seller is pretty clear: Edgeio will pick-up their listings (the in in Edgeio) at no cost and it will give it more exposure (the out in Edgeio). Suppose that 20,000 people visit your blog. If Edgeio can redistribute a feed with your ad so that you get an exposure to 20,000,000 people then you are more likely to be able to find a buyer.

The funny thing is that they do not redistribute you RSS feed but their own with a link to your post (so that they can keep track of the data flow). So they basically want to be some kind of middle-ware.

The beauty of the concept is that it can work for the RSS feed of a typical blogger, the RSS feed of an online store or the rss feed of another marketplace. You get the picture...

I think that the value proposition is less clear for the buyer so it will be interesting to see how their platform evolve. One remark from the audience was: "If I am looking for a classified ad, I don't want to get results from a store". Also if they want people to search directly from their Edgeio.com site, they will have to buy mind-share in the space already crowded.

So will this get traction?

Keith mentioned that there is about 10 blog posts a day that use the listing tag. He think he got this from Technorati chart tool:
Posts tagged Listing per day for the last 30 days.

Get your own chart!
He mentioned they were using this tag on purpose. If you look at Technorati results for this tag right now, there are a lot of things not relevant to Edgeio. If you want to see how this would work with only relevant ads take a look here. So one of their challenge will definitively to filter out all the spam or non relevant posts. Just for fun, I will tag this post with the listing tag ;)

Obviously not everybody is blogging and not all the bloggers will be comfortable posting classified from their personal blog. One the beauty of eBay platform is that you are can be anonymous (at least until the checkout) and accountable (with the transparency of the feedback system) at the same time.

However keep in mind that generation Myspace is coming. More and more people like to store the information they want to share in one place such as a blog. For them, I would think it will be natural to post classified directly from their blog.

Business model

At this time, Edgeio business model has 3 components:

• Give Power Sellers an option to buy a top position for their classified ad for a specific area for 25 cents a day
  
• Adsense revenues (this is web 2.0 after all ;) )
  
• An anonymous list of location/product demand not currently fulfilled. This was not well define yet but I understand this would be something similar to eBay Want It Now.
  

Looking at the URLs from the Edgeio demo, it seems like they are using Ruby on Rails for the front end of their application (see Keith's comment below for more information about their architecture). It will be interesting to see how well this scale...

Will it work?

Well this is beauty of the web, it is very difficult to predict what will work and what will not work. The thing I know for sure from my years working at eBay is that building a "global open marketplace" is hard! So good luck to them.

Keeping track of Edgeio

Michael is keeping track of the buzz on the blogosphere here and here and he said he will be posting his thought on some of the questions being raised about Edgeio challenges. This is a very good way to start the conversation and evangelize their solution!

Edgeio blogs:
Edgeio blog
Keith Teare (co-founder/CEO)
Michael Arrington (co-founder)
Vidar Hokstad (back end engineer)
Matt Kaufman (product manager)

Google Base

It was also interesting to learn a little bit more about Google Base vision directly from its product manager.

Google Base is a new way to complement the content Google want to organize. Instead of crawling the web, why not let the content owners provide the content directly to Google.

Also why not provide the next iteration of a tag system with labels and attributes so that content owner who know the content better than anybody else can add some meta data/structure on top of the content.

The labels and attributes are used internally to route the content to the right Google property (for example a product fed to Google Base will show up in Froogle) and to experiment new ways to search (most common label/attributes will show up and you can navigate trough them to focus/narrow down your search).

This is interesting because I think that what made the web 1.0 leaders was their ability to scale their business model as well as their technology. What will make web 2.0 leaders will be their ability to leverage/scale User Generated Content.

Update #1: Interesting perspective from Will.
Update #2: Thanks to the statistics of this post, I just discovered a Swiss competitor to Edgeio. It is called Ichiba and they have a very cool cartoon on their home page to explain the concept of Edgeio and Ichiba.
Update #3: Updated my comment on their architecture based on Keith's comment.