Introduction to Software Testing Management - Software Testing as a Service - page 4

strong that their cars are virtually defect-free. This has helped Toyota tremendously
in capturing market share and reducing their manufacturing costs. The Toyota
manufacturing model has become a de facto standard in the automobile industry,
and subsequently other industries have followed their example.

1.. Return on Investment

From another point of view, preventing defects as early as possible presents a
compelling ROI perspective. Say cost of finding and fixing a defect at prototype
stage is US$40. This defect is not found and is passed on to the design stage,
where it is trapped and fixed. In that case the cost of fixing it would be at least
fivefold as the defective prototype has gone into design and so design is also
flawed and needs to be fixed as well. At the same time this one prototype defect
introduces five design defects. After the production stage this cost will become
at least 25-fold from the prototype, as now either the product prototype along
with the machine design will also have to be fixed or these 25 defects after
the design stage have to be fixed. Finally, when it comes to customer site, the
cost will be a staggering 125 times that of the prototype stage, as now support
cost will also get involved. So the cost of fixing the defect will be US$5000
($40 × 125). So you see it definitely makes sense to fix the defect as early as pos-
sible. Of course, time and cost constraints put a limit on to what extent defects
should be traced and fixed. But continuous improvement techniques help to
reduce defect injection and improve defect prevention over a period of time once
the techniques are introduced. This is what CMM (Capability Maturity Model)
talks about.
Coming to a software product, suppose for each quarterly release on average
500 must-fix defects are found. The cost of fixing each defect at build phase is at
$100; at testing phase, $500; and at customer end, $2500. Now suppose the soft-
ware vendor has no testers. Industry standard is that 25% of errors are detected by
developers, 50% on average by skilled testers, and the remaining 25% by end users.
In our case, because no testers are involved, 75% of total defects are detected and
fixed at customer site after 25% of defects are detected and fixed by developers.
Total cost of fixing defects comes to $887,500 ($12,500 + $875,000).
Now suppose we employed three software testers at $50 per hour. So their total
salary for the quarter comes to $108,000. In this scenario the cost of fixing defects
comes to $495,500 ($12,500 + $108,000 + $375,000).
Compared to the previous scenario we can see that we are saving $392,000.
This is a saving of more than 44% over the case when no testers are deployed. So
it can be clearly seen that there is a great return on investment (ROI). The ROI
improves further when the testing function is outsourced and even further when
it is offshored.