Back to the day job! Disruption & Optimisation

Been to Betfair and one of Accenture’s project teams recently, and following sensible advice from colleagues, while I can’t reflect what they said, I can repeat what I said, or observed.

Betfair, fascinates me. They’re an internet age business. Their trading site is here… and they have an “about us” popup.

I occasionally bet on the horses, i.e. the Grand National, and even more rarely on politics, but I’ve always used a bookmaker, so I’ve not played with the spread bet players, nor an exchange. The Guardian’s write-ups mentioning Betfair have mainly been about either how gambling is leading an internet renaissance, or how much money the founders will make. What this misses is the rare nature of the social and commercial relationship between Betfair and its customers. The exchange model means that Betfair are enabling punters to bet with each other. Betfair thus enables a new market, and claim to offer better value to gamblers than the old bookmaker model. It means that they enable individual activity, have a very personal commercial offering and are also highly disruptive to the market and the traditional players. Its all very “Cluetrain”. They’ve created a new market and implemented it in software. They seem amazingly successful. They’re not just about channel efficiency. More an ebay, less a lastminute!

The Accenture team I met asked us to explain our IT estate modelling techniques and approach. The conversation brought home to me some changes in emphasis. When we started (two years ago) it was always about cost, now people are interested in cost, power, heat, space and/or utilisation. Occasionally, they’re interested in productivity, cost to manage and service quality. The scenario based modelling approach advocated by Sun, allows customers not only to own the data quality issues (again see my blog here…), but also to test different scenarios against the key drivers. Each customer will want to optimise the different dimensions of the answer to different degrees, deriving a unique answer.

In particular, trade-offs between space & power/cooling can be explored. I still need to work an example through. I’ve often felt that mathematical optimisation techniques might be usable to measure and/or describe the choices; as frequently the choice we have sought to offer to DC managers is to either maximise utilisation or minimise project cost, and we look to optimise on the investment profitability.

I wonder if a three dimensional graph might work, or if a radar (polar) graph could be built to help understand the optimisation decisions.


Originally posted on my sun/oracle blog, republished here in Feb 2016.

Comments are closed.