Oracle
Search Site  
Find
 
Crystal Ball Home
Oracle Home
Information On
   Web & Live Events
   Six Sigma & DFSS
   Education Alliance
   Training Classes
   Conferences & Forums
   Available Languages
 
Quick Links
   Shop
   Download
   Newsletters
   Contact Us
 
 
Worldwide Offices
   United States
   United Kingdom
   Germany
 
 
 
 

APPLICATION SUCCESS STORY


HP logo

Crystal Ball Allows Hewlett-Packard To Bring Printers To Market On Time

CUSTOMER OF THE MONTH (MAY, 1998)

Jeff Brown is a Project Manager in the R&D Lab of Hewlett-Packard's Work Group Color Laserjet Division. One of Jeff's main responsibilities is estimating the duration of production projects from conception to manufacturing release. An accurate production schedule provides HP the ability to effectively bring products to market on time.

In the past, Jeff has used COCOMO (Constructive Cost Model) to aid in establishing production schedules. This method alone, however, did not address the particular areas of variability that were causing delays in production. One of the difficult areas to accommodate in the planning process is the relative efficiency level of individual engineers since each one works at a different pace in various areas of expertise. Based on historical data regarding output, the engineers are grouped into "feature teams" and are responsible for the development of specific printer features. The complexity of the particular printer features is another important variable because some features require more engineers and production time to develop. Also, Jeff must consider the levels of integration between the various "feature teams" which indicates how many other teams any particular team will have to work with in order to accomplish their tasks.

Using Crystal Ball to help develop his most current production schedule, Jeff performed over 10,000 Monte-Carlo simulations using just under 100 control variables to help identify those areas which posed potential risks to the production timetable. The results from these simulations gave Jeff the information necessary to most effectively allocate production resources.

When using the COCOMO method alone, the production schedules Jeff produced were often significantly incorrect, sometimes varying as much as 9 months. After employing Crystal Ball, however, Jeff was pleased to discover that the difference between the scheduled completion date and the actual completion date had been reduced to 5 weeks. With this large of a difference between the two production schedules, it is easy to see why Jeff views Crystal Ball as being an integral tool in helping to bring HP's Colorjet Printers to market on time.

    ADDENDUM - January 5, 1999 - Response from Jeff Brown

    "I just completed another cycle of the software estimation model described in your Hewlett-Packard (2) application story, and I now have 30 managers asking all kinds of questions about things we can model, ideas we can try, and potential improvements we can make.

    They were very impressed by how quickly and easily I could change some of my control variables and begin a new simulation; the turnaround time on what-if questions was only a couple minutes. They could watch the new simulation run on our overhead screen, even as I described how the modeling techniques worked.

 
Home | Products | Services | Industries | Applications | Support | About Us | How to Buy
Privacy Policy | Trademarks | Copyright © 2007, Oracle and/or its affiliates. All rights reserved.