Software Testing and Quality Assurance
As a long time provider of high value, high volume, B2B data services I am keenly aware of the importance of testing and quality assurance for the services that I worked on. Most were API based data delivery and some, even recently, have been file based data delivery. Either way the common element was that our customers had to be able to trust the quality and correctness of the data being delivered. They most often integrated the data that we provided with other information and systems to create further value for their customers in critical and timely business decisions. If there are errors in B2B data services then those errors are distributed to our clients clients and a tsunami or missing or wrong information can be created. There is no room for inaccuracies, errors or missing data in a competitive data market like that of AIS.
QA Requirements in the AIS data market
When commercial AIS data services first came to market, the big market opportunity was with oil traders. This and other seaborne market analytics remains the highest value commercial sector for AIS vessel tracking data to this day.
What those clients are doing at a high level is using AIS vessel reports to assess either the current market situation with cargoes in transit, or the availability of vessels to transport cargos on particular routes. These address the supply and demand regionally for both product and transport, which together influence the prices paid in new deals. Obviously it is move complicated than that, but that is the general situation.
For data services, it means that the quality and timeliness of the data that feeds these market focussed information systems is critical.
Missing data, creates a competitive disadvantage.
Wrong data undermines product confidence and corporate reputation.
Slow delivery of data reduces the advantage of an information service.
Greenfield Testing Team
When I joined Spire Maritime, some of the priorities for a reliable, accurate, well performing, data service where not being fully achieved due to the problems of legacy software platforms that were unable to scale and perform in line with the business growth. Slowly while running the pre and post sales support functions, I enabled atleast manual testing and checking of the service quality for the first time.
When we started creating entirely new implementations of services, I initiated full testing capabilities on each product that was redeveloped. Initially I performed the testing work myself but as development scaled it was necessary to create a dedicated testing team.
What we ended up doing was having a testing lead who was an FTE and then we setup an offshore testing team through Automation Solutionz using their testing platform Zeuz.ai. With an initial Zeuz team size of 3 and eventually 5, over the period of 30 months we established automated reliable testing for all product enhancements, data flows and API products. The first year of this work focussed on product enhancements and new data flows, then the next 2 years focussed on setting up automated testing for new implementations of all existing services, which were being either migrated from physical data centre deployments to cloud for the first time, or were migrating between different cloud systems due to corporate changes.
Detailed, reliable and accurate testing of all aspects of the services meant that over 400 customers were migrated seamlessly from old to new services, with virtually no problems. That is how you protect $40M review and keep your customers happy during complicated IT redeveloments.