My company is an IT provider for several large websites. Before a typical release, we test all scenarios by hand. It usually takes a couple hours which are then billed to the client.<p>We're discussing the idea of moving to mostly automated testing before releases. While this is a big efficiency gain -- 3 hours turns into about 30 minutes -- this is still 2 and a half billable hours which we lose.<p>The advantages of automated testing are substantial for our clients, but don't justify a 6x rise in prices.<p>Is there an alternative approach that anyone is familiar with that utilizes automated testing for a unique revenue stream? We've been kicking around ideas, but have hit a brick wall in terms of losing billable hours.
Seems like the typical "charging for time vs. charging for projects" debate.<p>You could add automated testing as an extra feature. Pay us X up front (where X covers the average income per project from manual testing) and we'll build in automated tests to your project, which saves you money down the road. Otherwise, we continue to bill for manual testing on an hourly basis. The revenue gain or loss in this case depends on how long lasting your client work is, and thus how much time you'd spend billing them.<p>You could also just standardize your testing pricing: pay Y per release for testing. Then automate the testing internally, but continue to bill the customer for the "testing" feature at the pre-determined rate.
Could you record the tests you perform with jmeter, and then run the tests against the sites in the future?<p>I've heard selenium is good for functional testing, I've always used jmeter though.