Thursday, September 8, 2011

The Furlong may take a Fortnight and weigh a Firkin


A project metric is defined any data that may be mapped by a function of math.i Functions used upon these metrics include; Measurement, Comparison, Analysis, Synthesis, Estimation and Verification.

Performance metrics are defined as any metrics that are used to derive value for a given system under scrutiny, these include financial metrics, intangible projections and EVM related data.ii The process of generating metrics consists of defining requirements for measurement, and developing the measures to determine if said requirements are met as well as establishing well defined targets to measure against.

Methods used to create these metrics include project charting using earned value management, the use of performance based logistics or defining and methods to determine key performance indicatorsiii.

Key performance indicators may be classified as into the following groups:

Quantitative
Practical
Directional
Actionable
Financial

Each of these indicators can be used to develop Balanced Score Cards, HeatMaps or as inputs for SWOT or magic quadrant analysis for a given project and it's deliverables. These also include the various methods of charting such as Histogram, Candlestick, Inverted bar, and others.

The distribution of this information once it's created may be via standard methods including e-mail or the internal network (the intra-net) of an organization where proper security and accounting controls may be used to audit and control access to this information as key performance indicators on a major project can indicate corporate risk appetite and potentially damage the reputation of the company or organization should they be disclosed publicly. Eg; when Sony online's network were breached their stock dropped 15% in less than one week due to widespread fear of public reprisals including class action suits due to said disclosure. If a company or organization is spending millions on a software project to improve it's security stance the very existence of this expenditure could affect public confidence. As Tzun Tsu stated “The basics of all combat rely upon a foundation of secrecy”.

The methods cited here are by no means exhaustive such as Six Sigma or TQM as cited by Schwable; however they are used frequently enough to warrant discussion.

Another issue is the standardization of a given organization to agree upon the methods to be used beforehand to ensure consistency in the delivery of these project performance metrics as if the data cannot be read and understood than it might as not exist in the first place. Thus if the organization agrees to conform to ISO 9000 and then standardizes on ISO 12004, 10007, and 15403 for quality management they must also standardize upon the project methodology and related performance metrics to be reviewed by each of these procedures to ensure that they are both consistent and understood.

Thus we may define what performance metrics we wish to measure within our project, apply a ruler of our own design and in turn derive value from the reports generated to be used as inputs for business decisions related to our project in question. Odd units of measure that come to mind are the inverse femtobarn, attoparsec or hoppus foot, for projects we may define dollars per hour per objective feature as a standard if we so wish for efficiency, or time per line of code (t over kloc) per project.


iPocatilu, Paul (Revisita, Informatica Economia, 2007) IT Project Metrics [Online] PDF Document, Available from: http://revistaie.ase.ro/content/44/26%20pocatilu.pdf (Accessed on September 8th 2011)
iiBrown, Mark Ghram (Productivity Inc. 1996) Keeping score: using the right metrics to drive world-class performance P. 6 – 10 ISBN: 0-527-76312-8
iiiFitz-Gibbon, Carol T. (Blank House, BREA Dialouges, 1990 – 2 ) Performance Indicators ISBN: 1-85359-093-2\