Edgeio Update

Edgeio provides some stats on its home page which indicate the number of listings it has gathered and the number of cities associated with those listings. These have both been increasing, and it is probably too early to use this data to judge the success of the system, but there is not yet any clear rapid growth. Here is the graph for listings.

And here the one for cities.

It is interesting to note firstly the large jump in both listings and cities - a sudden increase of data in the database, perhaps associated with a bug fix and recrawling or reparsing cached data. Secondly, there is a smaller increase in cities which has no associated jump in listings. Perhaps a problem with the location identification code which had no impact on the number of listings discovered.

The Alexa reach statistics show some growth, though very spikey:

The shape of this graph seems to be mirrored in the mentions of the URL as found in BlogPulse:

What we are looking for as a measure of success is non-linear growth in the number of listings (indicating that sellers are using the system and that the system is doing a good job of gathering the data) as well as similar growth on Alexa. One of the problems with the basic idea of Edgeio is that the passive manner in which you add listings provides no confirmation - sellers would have to go to Edgeio to discover if there stuff was picked up or not. Anywhere there is such a barrier, users will be dropped.

Comments

[Keith Teare]

Hi there

Great that you have pulled this together. I can shine some light on the jumps for you.

It isn't caused by bug fixes but by single large listings owners deciding to support edgeio. This is particularly noticeable in the Jobs and Real Estate tags.

edgeio was architected for any RSS enabled publisher to upload their catalog automatically. And for updates to be uploaded in real time. We expect that alongside bloggers, ther will be many larger publishers who choose to do that. The spikes in cities and listings are when they do so.

This is a kind of non linear growth, but because it is very early it looks spikey. If it continues, then within a few months edgeio will grow to several million listings (it will continue :-) ), and be in tens of thousands of cities.

This is the power of a bottoms up publisher/submission approach. It is really hard for centralized submission based services, or scrapers, to get to this kind of scale (Craigslist is only in 100 or so cities due to its top down approach to geography).

After a week we are thrilled that we have listings from almost 1500 cities, even though each city (with some exceptions) has only a small number of listings. The approach will scale very well and quickly. Imagine the power of a system where a small town in China can have its own classified marketplace simply because people start to use the local cybercafe to come and post items for sale or wanted. and it will all be in Chinese because edgeio is publishing what the users write (both the text and the tags). It's hard for a top down system to have this power.

On your point about the need for feedback. We do confirm receipt of a listing - usually via trackback. See here - http://listings.teare.com/?p=134#comments - for an example. If trackbacks are switched off then only the publisher sees the trackback, so it is a private communication.

We will introduce some new features for less technical bloggers over the next 2 weeks, and also some tools for those without a blog, or who want a special blog for just their listings. Mike Arrington described some of that in a podcast - http://blog.edgeio.com/?p=19 - today.

Best regards
Keith Teare
founder/ceo/edgeio