Why Did Taurus fail and Crest Succeed?

The London Stock Exchange (LSE) is one of the world’s oldest stock exchanges and can trace its history back more than 300 years. Starting life in the coffee houses of 17th century London, the Exchange quickly grew to become the City’s most important financial institution (londonstockexchange.com). The LSE has seen many milestones throughout its long history, none bigger than the “Big Bang” in 1986 when market de-regulation occurred. Interest in the market increased hugely and the old system whereby stock certificates changed hands to represent ownership was no longer practical. London realised that computerisation was the way forward and the share settlement system TAURUS was introduced with share ownership becoming an entry in to a computer database.

TAURUS was a huge project and received heavy investment. In a bid to please member firms that would all have to use the system, the TAURUS design team tried to accommodate everyone’s demands. It made an ill-fated attempt to become “all things to all men”. Inevitably the system became far too complex. As a result, TAURUS’ implementation encountered one difficulty after another – deadlines were being missed and budgets were continually being over run.

The TAURUS project is widely thought to have been such a fiasco because of the impossibility of mediating the demands of the powerful vested interests of the financial community (Flowers, 1996). In his book, Flowers reveals a system that never possessed a stable architecture and as such never had a realistic chance of achieving its objectives. The TAURUS project was terminated on March 11 2003, a failure costing £75million.

To take its place, a new system called CREST was designed. Having learned from the mistakes made in designing TAURUS, the design of CREST was co-ordinated by a core, high level design team of 4 or 5 who kept an overview of the entire project (Currie, 1997) instead of the hundreds of staff used for TAURUS. This might have been a major factor in the success of CREST.

In designing CREST, the team did not have to give in to the demands of member firms as they did for TAURUS, because CREST usage was optional. The result was a far less complex, easier to use system. The designers took advantage of this and could afford to exploit the Pareto effect, and cater only for the 10-15% of the business functions that made up 85-90% of the volume of transactions (Head, 2001).

Time and budget constraints were a major part of the failure of TAURUS and part of this could have been down to the fluid nature of supplier contracts (Currie, 1997) which would allow suppliers to prolong jobs and increase their own profit. During the design of CREST, fixed price contracts were negotiated with suppliers giving incentive to deliver their product on time and on budget.

Also during design, the CREST team decided to adopt a “follow the recipe” approach using only tried and tested technology. The disadvantage suffered by TAURUS was that the technology it used was relatively untried and untested which meant there was no support when problems arise.

Overall, it is quite clear that TAURUS was simply an over-ambitious project which in hindsight, never really had much chance of success. The problems all started at the development stage, principles used in running the overall organisation (LSE) were applied to running the project creating a legacy of “development by committee” (Flowers, 1996), a form of management totally inappropriate for systems development (Goulielmos, 2003). In my opinion, the number one reason why CREST succeeded where TAURUS has failed is that CREST was able to learn from the mistakes made by TAURUS. Each and every problem with TAURUS was addressed and eliminated by adopting a completely different method of systems development. It could be said that the LSE simply could not afford to fail with CREST such was their dwindling position in the financial world, so CREST was always going to be more likely to succeed.


Sources:

Currie, W “Computerising the Stock Exchange: A Comparison of Two Information Systems” Blackwell Publishers Ltd, 1997

Flowers, Stephen “Software Failure: Management Failure: Amazing Stories and Cautionary Tales” 1998

Goulielmos, Markos “Outlining organisational failure in information systems development” Disaster Prevention and Management Volume 12 Number 4 2003 pp. 319-327

Head, C. H “Taurus and Crest, Failure and Success in Technology Project Management”, Henley Management College, 2001

www.londonstockexchange.com