Sean Park Portrait
Quote of The Day Title
I say profound things

User generated information architecture and market structure

Jesse James Garrett, founder of Adaptive Path opened this morning’s session at reboot with an interesting presentation on how tags and other user generated information will need to continue to inform and drive UI design, information architecture and optimization going forward.

He used Amazon’s algorithmic architecture as a example of leadership in this respect, although he pointed out that even Amazon has only started to scratch the surface of what will be possible. Interesting factoid: Amazon have a Chief Algorithm Officer, underlining the importance they assign to this as being core to their business model and value.

At the risk of mangling his message in an inaccurate shorthand, some of the points Jesse made are as follows:

‘How does Amazon do it?’: …in their url they create a ‘session-id’ that then allows them to track – by tagging all items on each web page, using query tags – all movements by mining their server logs…

‘What’s next?’: …need to use sophisticated statistical techniques to transform raw data into actionable items, to find patterns in consumer/user behavior is key…

…algorithmic architecture to become transparent to the user, but data needs to be better: both about the content/products (ie metadata!) and users/customers (turn every visit into a usability experiment data point)

One of the questions that followed his presentation was: ‘Is this (building powerful algorithmic reccomendation/navigational engines) Amazon’s ‘real’ business (as opposed to ‘selling books’)? Indeed with initiatives like A9 and the fact that by their own admission they are a ‘technology’ company, it would seem plausible that this might be the case. Jesse opined that it was likely that – either from Amazon or from others – these kind of tools would likely be available for sale ‘off the shelf’ in the near future. (I must admit that framing Amazon’s business in this way was one of the reasons I picked it as an avatar in my AmazonBay metaphor.)

Stowe Boyd followed up with a counterpoint presentation – The Revolution will be socialized – “all e-commerce will be socialized” using his thesis of social architecture – individuals to social groups to markets – projecting that the future of markets lay not in Jesse/Amazon’s paradigm of giant centralized algorithmic engines but “at the edge” with social, peer-to-peer transactions driven by community and social connectivity.

Funny, I agree with both of their points of view; I don’t see the conflict or any contradiction. For Stowe’s social architecture to work properly (in a scaled fashion against the background of the infinitely complex and growing backdrop of data and metadata) in my opinion you will need the algorithms. If I understood him correctly, I think Stowe agrees with this point but argues that these will all be localized at the edge. I think that both (large central and small edge) will exist together forming a powerful symbiotic partnership, allowing the ‘socially-driven’ transactions and interactions that will proliferate at the edge to flourish alongside more banal and commoditized transactions that will be more efficiently executed at a ‘super-node’ on the network.

I suppose I could be accused of sitting on the fence but I genuinely can’t see the future of markets evolving into either (to take it to its extreme logical conclusion) one super Hal9000 central hub or to a completely dispersed dust storm of bilateral virtual markets (like virtual particles that appear and disappear a trillion times a second). The imagery that springs to my mind as to what the future structure of markets may look like, with this combination of super-nodes and distributed micro-nodes is the map of the earth at night or the heavens. Not perfect visualizations but hopefully helpful nonetheless.

What would this mean for financial markets? Well, a combination of giant global exchanges (NYSE – Euronext to …AmazonBay?) and an ether of micro/virtual exchanges (which could be as small as two parties connecting bilaterally.) One consequence is the greying of the distinction between ‘exchange’-traded and OTC (over-the-counter) markets. This will pose real issues for the current regulatory paradigm in most markets and geographies, indeed it is already straining the existing received wisdom; as will the very idea of national or regional regulatory regimes in the context of global dematerialized markets. It will also continue the ‘deconstruction’ of traditional boundaries within financial markets (banking, securities, insurance, etc. etc.) as these frames are rendered more and more obsolete but the ability of the network and its users to build up any needed financial product from its base risk components and cash flows. I’ll stop now because to elaborate on these ideas would mean to write a book (or at least a long essay) and I’m not sure their is much demand for that… 😉

  1. At 8:53 pm on 04 Jun 06 JohnOS said:

    I think the exchange/OTC distinction is already blurred by the ECNs like Bloomberg and TradeWeb, and single dealer proprietary platforms.

    I for one, would like to see more elaboration on these ideas. I wrote up some of my oen ideas on whether Google could build an ECN here…

    For me the key issue is solving the trust and creditworthiness problem.

    I don’t think I see markets fragmenting and decentralising as much as you do, because traders go where the liquidity is concentrated.

Leave a comment

You must fill in an email address, but don't worry - it won't be displayed to the public.

* Required field.