Test Data Generation

New or enhanced computer systems can change the way a business operates because of the altered way in which the data is manipulated. Orders may arrive more quickly, or at the “wrong” time of day, or mainly at weekends; forms may be too tedious for users to complete, keyed entry may be too slow for critical situations.

Behaviour like this is not easy to predict, but realistic test data can go a long way towards highlighting likely bottlenecks and changes of emphasis. Thus every effort is made to understand the business scenarios giving rise to particular test conditions, and to generate realistic, synchronised data to exercise the new system. This data may be loaded into a common-format store so that manual and / or automatic test packs can be prepared.

Confidence
Realistic test data can provide the confidence in both yourselves and your users or clients that the new or modified system is valid and fit for purpose.

Early Warning
Because the test data poses “real” questions it is able to pinpoint areas which may be challenging at pilot and /or implementation, allowing time to workshop and plan the solutions and workarounds.

Test Monitoring
The generated data is held in one database, which can then be used to allocate data for specific tests and predict and note the results, for later comparison with results obtained from the new system.

Training
A “clean”, reloadable version of the test data is an ideal foundation for training services and activities

The way we do it
We use any available existing data sources, regardless of format, to arrange the data in a single holding area. Where sources do not exist, data is generated to be as close to real experience as we can get information for – naturally the specification is the prime input for this type of data. Presentation of the data is then arranged in two ways:

as reports and queries, for discussions, workshops and input

as automatic input to the new application, so that can be exercised rapidly and repeatably