EN
Back to the archive

The encyclopedia · Software & IT · Technical decision · 2002–2006

Agere read demand changes 1–7 months early from product 'leading indicators'

Agere Systems found products that signal demand shifts for whole groups, forecasting 3,500 semiconductors 1–7 months ahead.

Agere Systems · Lehigh University Center for Value Chain Research

the move

By 2002 Agere Systems, a semiconductor company with a global, contract-manufactured supply chain, was introducing short-life-cycle products faster than its forecasts could track. Demand for a product could shift before monthly numbers revealed it.

With Lehigh University's Center for Value Chain Research, Agere developed a leading-indicator engine: it scans thousands of products to find those whose demand patterns consistently foreshadow the demand of a product group, then uses them to forecast one to seven months ahead.

Across a data set of 3,500 semiconductor products, the leading indicators predicted group demand patterns with correlations from 0.51 to 0.95. The concept extends to financial and inventory forecasting.

why it works

  • Leading indicators trade a little precision for a lot of time, which is what supply chains actually need.
  • Group-level forecasting is stabler than product-level noise, so the signal survives.
  • The same engine applies to financial and inventory planning beyond the original chips.
the payoffForecast groups via products that lead demandclever

what transfers

Instead of forecasting every product alone, find the few products that lead the group and read the group's future through them: early signal beats precision.

what came after

Agere published the leading-indicator method in Interfaces, where it became a reference for demand sensing; the paper argues the concept extends to financial and inventory forecasting beyond semiconductors.

references

spotted an error? The archive wants to know.

same kind of clever