Put your software to the test

IT managers must use software testing to ensure their applications perform as they should.

Testing software can help you detect potential problems early on, meet business and customer requirements and save money. But many applications go live without testing.

Then there are the questions of who does the testing? What do you test and how is it funded?

It's clearly an important issue for the IT industry, which is why industry body CSSA, in conjunction with independent infrastructure consultancy SIM Group, has launched a software testing special interest group.

The group surveyed 40 testing and marketing managers about their views on software testing from an IT and business perspective. Not surprisingly, all respondents said they plan to test IT products before they are launched.

Many said that testing is critical for all types of software, especially client-based software, business critical and financial systems.

Of those surveyed, almost half said the main benefit of software testing is to ensure products work correctly.

The research shows that software testing could increase customer confidence and satisfaction. This is what you would expect, but there are questions about how much testing is actually done.

Testing can cause problems. Half of the interviewees said it's time consuming and can be difficult to finance.

"It intrudes on the development process," said one respondent. "Developers have to respond to issues raised and there isn't a very visible end result, which means it can be difficult to fund."

A similar argument was put forward by another respondent who said: "It's time consuming and the quality isn't always a tangible asset to measure."

Testing is still a largely home grown idea. Three-quarters of interviewees said they would prefer to keep testing in-house, believing they have more internal expertise and would be able to better control the process.

Just 10 per cent said they prefer to outsource testing. Some outsource because of a lack of internal resources, others only outsource particular areas, such as menial and security testing.

Fewer than half of the interviewees carry out automated software testing.

Cost, lack of experience and software limitations are some of the reasons why 45 per cent of those surveyed do not automate software.

Some 55 per cent of respondents using automated testing said they prefer it to manual testing because there are fewer errors and it is quicker.

Just 11 per cent prefer manual testing.

Almost half the interviewees said their initial software purchase was not influenced by an incident where untested software had failed. But software failure had influenced 15 per cent to purchase software.

More than half of the interviewees - 62 per cent - said they have not changed their software testing supplier. A previous track record and cost are the most important factors to consider when selecting a supplier.

Other factors include technical compatibility, references and reputation.

Some eight per cent of respondents, however, changed suppliers less than a year ago. Reasons for changing suppliers include poor performance and integration problems.

A quarter of respondents said the decision to purchase software testing rests with the IT manager. Other decision makers include the finance manager, project manager and managing director.