Home | Business News | Browse by Publication | I | Interfaces

Spreadsheet decision-support tools: lessons learned at Hewlett-Packard.

Article Excerpt
Spreadsheet tools are an inexpensive and flexible way to provide analytical support for repeated decisions. However, good modeling and analysis alone are not enough to create an effective tool. Using examples from Hewlett-Packard in forecasting, planning, procurement, and product management, we discuss lessons learned for creating reusable spreadsheet tools to support fact-based decision making. These lessons pertain to the entire life cycle of tool development: Before starting development, what role should analytics play in solving your business problem? Is a spreadsheet tool an appropriate solution? During development, are you partnering with end users to discover key insights and building a simple tool to capture those insights without unneeded complexity? Can you develop an OR champion in the process? When handing off a tool, are you providing a complete solution including roles and responsibilities, processes, and metrics? Ensuring that spreadsheet tools have maximum impact in an organization requires smart decision making before the first prototype is drafted, customer-focused management during development, and a complete handoff to users.

Key words: spreadsheets; philosophy of modeling; OR/MS implementation; inventory/production; supply contracts; forecasting; product policy; cost analysis.

History: This paper was refereed.

**********

We have collaborated on numerous engagements at Hewlett-Packard (HP) with the objective of enhancing the company's business performance through the use of analytics. As part of these engagements, we have often developed and transferred spreadsheet-based decision-support tools to analysts and managers in the company's businesses. Over a five-year period, we developed more than a dozen custom decision-support tools to support fact-based decisions in the company's product-development, marketing, operations, and finance communities. Spreadsheets have become our tool of choice for such applications, especially when the desired application is in a new problem space where solutions are not understood well. In these cases, spreadsheets have allowed us, with limited up-front investment, to rapidly build solutions that are transparent in nature, and that we can distribute easily and modify quickly as we build our understanding of how to attack the problem we are addressing.

This paper reflects our experiences and lessons learned from building and applying spreadsheet-based tools. Our focus is not on one-time strategic decisions, although we have found spreadsheets to be useful for building models to assess trade-offs in such decisions. Rather, our focus is on problems that require an analytical solution to support an ongoing series of repeated decisions. In these cases, we have frequently had to hand off a relatively complex analytical solution to users without a management science background and with little time or appreciation for detailed and complex analytics. We call such a solution a tool--which we define as a reusable, analytical solution designed to be handed off to nontechnical end users to assist them in solving a repeated business problem. While tools do not have to be spreadsheet based, we focus on those that are. Figure 1 highlights these distinctions.

In developing tools, we have had some hits; we have also developed some orphans--elegant analytical tools that solve complex problems but never find a home. We have reflected on the common themes in our successes and setbacks and created a set of guidelines for creating effective spreadsheet tools. Although good modeling and problem-solving skills are prerequisites for creating effective tools, our focus is on the roadblocks that can undermine a well-intentioned analytical effort. We draw on examples from our experiences working as consultants with HP's Strategic Planning and Modeling (SPaM) group and developing and supporting HP's Procurement Risk Management (PRM) and Design for Supply Chain (DfSC) programs.

[FIGURE 1 OMITTED]

Our approach to maximizing value from spreadsheet tools begins long before we ever conceive a tool, and extends far beyond completion of its development. Figure 2 shows a simplification of a spreadsheet tool's life cycle.

[FIGURE 2 OMITTED]

In this paper, we discuss our experiences and the lessons we learned from each of these three phases. Our suggestions may seem straightforward, but our experience tells us that they are often the key factors that differentiate tools that end users actually use from tools that they do not.

Problem Framing: Choosing the Right Path

Before building a tool, we must have a good understanding of the problem we are trying to solve to determine whether a spreadsheet tool should even be part of the solution. Our approach asks three key questions (Figure 3). We call navigating these decisions the process of choosing the right path.

[FIGURE 3 OMITTED]

Question 1: Will analytics solve the problem? Regardless of what a client may request, we believe it is critical to determine in advance if an analytical solution will make things better. Will the analytical solution enable a better, more valuable solution to the problem? Are there higher-priority or higher-impact nonanalytical solutions that should be deployed first? Are there organizational or other issues that must be resolved...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!