EN
Back to the archive

The encyclopedia · Engineering & Operations · Strategic decision · 1984–1986

Citgo ran its whole downstream business on one set of management-science models

Citgo deployed optimization, forecasting and expert systems from acquisitions to receivables, worth about $70 million a year in profit.

Citgo Petroleum (Southland)

the move

A petroleum company manages a fast flow of cash through buying crude, refining it, moving it, pricing it and collecting for it, and in Citgo's case each part had planned on its own.

During 1984 and 1985 Citgo invested in management science across nearly every business area, mixing mathematical programming, statistics, forecasting, expert systems and AI with organizational design. Top management backed it, many functions were involved, and the models were integrated with the firm's information systems so each amplified the others.

The result was about $70 million a year in profit improvement, and the project won the Franz Edelman Award. Its lesson was that the value came from coverage and integration as much as from any single algorithm.

why it works

  • Covering the whole cash cycle finds savings that one-function projects miss.
  • Integrating models with data multiplies what each can do.
  • Strong executive sponsorship got competing departments to adopt a shared approach.
  • Fast development kept the work relevant while the business moved.
the payoffExpose every part of the cash cycle to one model setinspired

what transfers

A few models wired together with good data and active sponsorship beat many isolated tools.

what came after

Citgo became an emblematic case of enterprise-wide management science, and the paper remains a standard reference on how to deploy models across a business rather than in a single function. The approach also fed later work in supply and distribution optimization in the oil industry.

references

spotted an error? The archive wants to know.

same kind of clever