Nowadays most major IT departments outsource at least some aspects of their operation. Although the ‘scope’ discussion is interesting for all IT function, I will focus on the development and test space, since this is the area where Trust IV specialise.
First of all, scope inevitably increases during negotiation. The picture I always have in my head is of a large tree with a substantial branch emerging horizontally. Along the branch are a number of people, each frantically sawing through the branch, (on side away from the tree of course), and saying, ”I’m the start of the retained organisation, everybody beyond me could/should be outsourced”.
With our background in testing and quality assurance, we think the key question when outsourcing development and testing is, “ ‘who’ marks ‘which’ homework? “
To answer, the ‘who’, perceived wisdom is to not have anyone mark their own homework. Yet throughout human history, artisans have always operated in that way. A medieval stone carver did not have an external QA function and yet our cathedrals are testament to the quality of the work they produced. The move to agile is a shift from the production line to the artisan model. This has had a profound impact on testing. Even with the waterfall model of development, having a different organisation do the testing isn’t going to guarantee quality.
The key thing is the appropriateness and rigour of the testing, not who does it. In fact, if not engineered effectively the creation of a commercial interface between developer and tester can generate complication and delay. So what is the answer to ‘who marks the homework? It does not matter provided that they provide metrics to show that it is being done properly.
Answering the ‘which’ is harder. The key concept is a word I used before, appropriate. It is all too easy to do inappropriate testing: testing everything because you don’t know enough to do otherwise, not testing the hard bits because they are hard, repeatedly testing the easy bits because they’re easy and so on. This results in a failure to focus your testing spend on business risk mitigation. The problem is inappropriate testing can generate as good a set of project metrics as appropriate testing.
So how do you ensure that testing is appropriate?
The process of determining what testing is ‘appropriate’ is easy to describe but much less easy to execute. Outsourcing simply makes this challenge harder; an Outsourcer would be more than happy to charge for testing everything or conversely, if operating on a fixed price budget will conspire to avoid the difficult.
If you’re struggling to decide ‘which’ or ‘what’ work can be outsourced, stay away from outsourcing until you have a better understanding of where to draw the line between retained and outsourced staff.