1. Project Tasks

1. Identify decisions that benefit from a prior analysis of past performance.

   Many examples exist: where to advertise, when to replace a machine, or whether to change a procedure. Identify the specific data that could help with decisions.

   PA-1  Selection of Business Performance Analyses  $6.50
2. Plan the electronic data collection.

   Find sources for the desired data. Determine their quality: are the data complete, correct, and up to date? List the formats in which you can obtain the information. Decide on an electronic format to use. Where possible, use a flexible standard such as XML.

   Define a process for extracting the data from their sources, removing errors, converting to electronic form, and formatting. If the business activities do not naturally yield the right data, introduce procedures that will do so.

   PA-2  Performance Analysis Input Processes  $6.50
3. Plan the exploitation of the analyses.

   Determine how to use the finished analyses in the decision-making. Who should receive the reports? How often? In what form (e.g., "dashboard" of key metrics)?

   PA-3  Performance Analysis Output Processes  $4.50
4. Design and implement the analyses, as well as the collection and exploitation activities.

   Select the algorithms for the analyses. First consider simple statistical functions (like mean and minimum), easy to program in spreadsheets. For complex procedures (such as hypothesis testing), check that you have tool support; some office applications include the required functions.

   Program the input. Example: export data from two existing spreadsheets to feed the analysis in another spreadsheet.

   Program the output. Example: chart the analysis results on a graphic and print it.

to top | to bottom

2. Project Management

• Plan.

  State the expected costs and benefits of the performance analyses. The risk exists of collecting large amounts of data but not exploiting them enough to justify the effort.

• Organize.

  Staff your team with business expertise, mathematical skills, and technical savvy. The team should be able to understand the data's meaning, master algorithms and statistics, and find technologies that move the data from their sources to the analyses and on to the decision makers.

• Execute.

  Keep track of progress against the original cost-benefit statement and make adjustments.

  Examples:

  • Eliminate potential analyses if you discover difficulties that make them impractical.
  • Accept slightly erroneous data if the cost of cleaning them is prohibitive.
  • Look for opportunities to reuse the same input data for several analyses.