What Does the "Architecture of Attention" Look Like?

From "Attention Economy" to "Attention Architecture"

Sep. 18, 2006 03:45 PM

Alex Iskold (pictured), Contibuting Editor to SYS-CON Media's Web 2.0 Journal, writes: I had an interesting chat last night with Chris Saad of Touchstone about their platform and the attention market. The conversation was prompted by the announcement on TechCrunch that one of the leaders in the attention space, RootMarkets, received funding from Chicago Board of Trades. This conversation with Chris and post on TechCrunch got me thinking: we all agree that we are heading towards the attention economy, but what does the "architecture of attention" look like?

To make the concept of attention compelling and to prove to the consumers that their attention information is important, we need to build applications that provide useful services. And to build these applications we need a platform for the attention players to plug into. In short, we need attention ecosystem, where application providers can interplay and deliver definitive value to the end users.

The Roots of Attention

It may not be widely known, but the foundations of the attention economy and architecture have been already laid out. Steve Gilmore and Seth Goldstein established AttentionTrust.org – a non-profit organization with a mission to both educate the people about the value of their attention and to establish the infrastructure for capturing individual attention.

The AttentionTrust serves as a forum for discussing and establishing principles, values and rights of consumers. The founders have outlined the following principles for its operation:

Property: You own your attention and can store it wherever you wish. You have CONTROL.
Mobility: You can securely move your attention wherever you want whenever you want to. You have the ability to TRANSFER your attention.
Economy: You can pay attention to whomever you wish and receive value in return. Your attention has WORTH.
Transparency: You can see exactly how your attention is being used. You can DECIDE who you trust.

These founding principles capture the essence of the attention economy. With every click, with every look at the computer screen you are paying attention. This attention information has a huge value, and it can be used to provide you back with valuable services.

The Foundations of Attention Architecture

AttentionTrust.org and RootMarkets, the attention company founded by Seth Goldstein, worked out an architecture for capturing and storing the user attention shown in the Figure 1 below.

The attention is captured by the browser extension, called AttentionRecorder. The recorder simply records which URL the user went to and when. With the recorder the user has an option of either storing the click stream locally or directing it to an AttentionVault.

The vault is essentially a remote database of the user click stream. Since AttentionRecorder defines HTTP-based API for communicating with the vault, their can be multiple implementations.

As shown in the Figure 2 below, the user can configure the AttentionRecorder to send the data to one of the vaults approved by AttentionTrust. This approach facilitates competition between the vault providers and, as advertised, puts the user in control. That is, the user can decide the most trustworthy, fastest, cheapest vault. Another big benefit is that consumer is explicitly part of the attention recording process. So the consumer has to understand how the system works, and so the consumer is more likely to trust the system because of that.

Beyond the basic attention

The initial version of attention architecture is simple, but it is not complete. As the attention space evolves it becomes clear that there is a need to expand it and to re-conceptualize how various applications and services fit in. Lets consider a few examples. The most obvious thing that is not being captured is whether the user liked what she saw. While it is certainly possible to build a good prediction model based on the time that user spent looking at the information, no such model will be complete or exact. In other words, current version of AttentionRecorder captures only implicit attention, but there are also rich variety of explicit attention.

Consider the popular social bookmarking sites like del.icio.us. We can think of them as vaults for the explicit individual attention. As a step up from implicit attention, when the user bookmarks a site and sends it to the del.icio.us we know for a fact that she liked the site. We still do not know how much she liked it. This information would be captured by another system - the one that also allows ratings. At adaptiveblue we are developing a higher level attention capturing service called blueorganizer. Beyond the basic URLs and ratings the blueorganizer captures the semantical information contained on the page, such as movies, books, wines and cars, creating basically a vault for semantically-rich attention.

So since there are different kinds of attention, the current architecture needs to be expanded to accommodate them. In particular, we need to recognize that:

? Attention can be captured by different sources

? Attention can come in different formats

? Attention can be stored in different ways

By focusing on these issues, we can extend the current architecture to a flexible and rich attention platform.

The key next step is to redesign the protocol to ensure that any kind of attention data can be stored. The types of attention data would need to be established and then each attention source can be paired with one or more attention vaults, again putting the user in charge of her data. Also, since attention data can be of different types, it might be beneficial to have different kinds of vaults. Some data would naturally lend to the choice of a relational database. At adaptiveblue we built the vault using Amazon S3, which I reviewed for Web 2.0 journal earlier.

Factoring in the services

So far the old and the new attention architecture has been focused on capturing and storing the user attention. These are of course important, but the least exciting aspects of the attention platform from the end user perspective. After all where are the end user benefits? The benefits must come from the plethora of services that analyze the user's attention and do something interesting with it. Personalized recommendations, personalized alerts, personalized news filters, personalized search and personalized shopping are just a few exciting services that can be build on top the attention platform.

The user will sign up for a subset of these services and point them to her AttentionVault(s). The services will then utilize the user attention information and seamlessly plugin into various aspects of on-line and off-line life to deliver huge productivity boost and time savings. For this to happen, the attention vaults need to offer the standard access API in addition to the standard input API. The actual format and the protocol for the attention data should be the same as for storing the data in the vaults.

Conclusion

The recent explosion of quantity and types of information puts us on the very fast track to the attention economy. Now more than ever before it is critical to understand and harness the value of individual attention information. AttentionTrust organization is the forum for discussing attention issues from privacy, business and technical perspectives. To truly harness the value of the user's attention, the players in attention space will need to work together and extend the existing implicit attention architecture to include wider variety of attention data and to create the standards-based infrastructure for attention services.

From attention economy to attention architecture

Published Sep. 18, 2006— Reads 13,933 — Feedback 8
Copyright © 2009 SYS-CON Media. All Rights Reserved.

About Alex Iskold
Alex Iskold is the Founder and CEO of adaptiveblue (http://www.adaptiveblue.com), where he is developing browser personalization technology. His previous startup, Information Laboratory, created innovative software analysis and visualization tool called Small Worlds. After Information Laboratory was acquired by IBM, Alex worked as the architect of IBM Rational Software Analysis tools. Before starting adaptiveblue, Alex was the Chief Architect at DataSynapse, where he developed GridServer and FabricServer virtualization platforms. He holds M.S. in Computer Science from New York University, where he taught an award-winning software engineering class for undergraduate students. He can be reached at alex.iskold@gmail.com.