According to the recent model-based testing user survey (see here), the respondents saw on the average a 59% reduction in escaped bugs, 17% reduction in testing costs, and 25% reduction in testing time.
What do these figures mean? Let’s say that a (software) product has 100 defects (that can manifest themselves) when it enters the testing phase, and that the present (already good) testing process finds 60% of those defects. The present process then results in 40 escaped defects. A 60% reduction in that number would result in 16 escaped defects instead of 40, so the amount of defects found would increase from 60 to 84, which is 24 / 60 = 40% increase in testing efficiency (amount of defects found per unit of work spent).
At the same time, the testing cost reduction -17% means that a testing project that took originally 100 hours takes only 83 hours, which corresponds to 20% increase in testing productivity (the testers work at “speed” of 120% compared to the original).
The compound results is that testing productivity has been increased to 140% x 120% = 168% of the original, which then means that a testing person who costs $100,000 per year now produces excess value of $68,000 per year at least. This is a simple economic argument, as if the person’s work wouldn’t be worth at least $100,000 per year it wouldn’t make sense to keep that person working in testing.
25% reduction in testing duration means that a testing phase that used to take four calendar weeks can be finished in three. The additional value of the saved calendar week in terms of time-to-market advantages is very industry-specific and attaching a concrete value tag to it is difficult.
(In our benchmark results, the productivity improvement figures have been consistently higher than in this survey, but the first model-based testing user survey spans across multiple tools and technologies and is not therefore directly representative of the Conformiq tool suite only.)