It was called the 'billion dollar challenge'. In 1996 BP, the international oil group, decided that it could save $1bn from the $15bn a year that it spent on goods and services.
The key was information. BP could only make these savings by getting an overview of all its spending decisions. Once it had the full picture, it could negotiate better deals with its main suppliers.
But two years ago, identifying those opportunities was difficult as there was no single place in which all the purchasing data was gathered. 'The whole thing tended to be put into the 'too difficult' box,' says Chris Browning, BP's former procurement manager for BP Oil (Europe) in charge of the project.
There was a need to build a computer system to collate information about who was buying what from whom. BP tackled this problem by building a data warehouse - information from many sources in analysable form.
PA Consulting was brought in to manage Project Oyster. It built a data warehouse to reside in an Oracle database running on a Unix server. Any registered user within BP could access the information using a Web browser across BP's intranet (an internal network).
Users as far afield as Australia, Alaska and Azerbaijan may interrogate the system to find if a potential supplier is doing business elsewhere in the group or if another supplier offers better terms. BP procurement managers may use this information to aggregate purchases and to negotiate better terms.
An important consequence is that BP can rationalise its supplier base. 'It could easily be seen as a big stick to beat suppliers,' says Steve Davenport of PA. But he argues that it allows suppliers to develop a more collaborative mode of working with BP.
'The real value of Oyster is in the second-tier suppliers,' says Mr Davenport. 'For heads of procurement, it is pretty obvious who are the top 20 suppliers. But look one level down, they don't actually know who are the next 20.'
The project - awarded the 1997 UK Management Consultancies Association's business improvement award, IT category - is judged a success by BP. But it has had to overcome many political and organisational obstacles.
One advantage was that it did not require much effort to input purchasing data as this could be transferred from accounting systems. The system also overcame a potential problem in the way suppliers names were described in different parts of the organisation. PA Consulting had to set up a sophisticated 'fuzzy logic' data matching system that could, for example, recognise that IBM was the same company as International Business Machines. The convenience and relatively low cost of the system has made it popular within BP. Data from 20 countries is in the system, now used by 700 people all around the group.
Within a year of the system's introduction, savings of at least $15m were made, equivalent to five times the project's cost.
