The value of any information technology is wholly dependent on the quality of the data it processes. Data collection and integration in information technology projects are far more complex topics than most businesses understand. Break your data management process into discrete steps to ensure best results and the highest possible return on investment for your technology dollar.
Monitor your business processes closely. You cannot act upon business information you cannot measure, and you cannot measure business information you have not sufficiently quantified.
Collect your data, completely and unobtrusively. Complete means that you have audited your information collection techniques, and the information you are measuring does not allow some to fall between the cracks in the collection process; if there is a correlation between the data that is lost and a business operations issue, then the remaining data you rely upon will be skewed. Unobtrusively means that you do not require your employees to spend large amounts of time documenting their processes–which reduces efficiency and leads to large data errors. Collect information passively whenever possible. For example: don’t ask an employee to track newly written contracts on a manual timesheet. Have a business policy that places new contracts in the same computer directory, and use a computer script to tally these on a regular basis.
Normalize your data. Normalized data is an information technology term that means, “write everything the same way.” You can review a client list and know that William Batson and Bill Batson are the same person, but if both names are entered into the same database, Mr. Batson may be surprised to receive two invoices for the same purchase. “July 1, 2010” and “7/1/2010” are the same date to Americans, but the latter can mean Jan. 7, 2010 if it is transferred to a database in Germany. Automate your data collection whenever possible, because computer output is naturally normalized. For human-compiled and entered data, ensure that your employees understand the normalization standards, and why they are important.
Analyze your data. There is little purpose to compiling business process data which is then not acted upon, and these actions should stem from a real understanding of both the raw data, and the business logic used by your analysis models.
Review this entire process on a regular basis. Business analysis should lead to new ideas for processes you wish to monitor, which require new collection procedures and new normalization standards. Each of these steps can be improved over time to ensure that your information technology investment is maximized.