Wednesday, January 28, 2009

9 reasons why agile testing could fail

Agile development is hot. When there is development involved there should also be testing involved. I think it might be hard for a traditional tester to be involved in an agile project. Here some thoughts of mine of which I think Agile testing might fail:

1. The tester might be outnumbered by the developers and therefore their voice about testing will be ignored or even the tester ignores him self
2. Only "Done" Functionality is shown in demo, the undone items are also their, if the business decides to take the product in production they accept the risk that untested, undone items are also face the production line
3. In demo's only the working items are shown, won't it be good to explain also what is not yet working?
4. If Agile is new for business how can equally trust exist? Understanding is a basis for trust, if something new you cannot expect that business understands what the team is doing or trying to do. If trust is not available in the beginning delay will start and adaptation and learning will be much harder.
5. If trust from the business is based on traditional methods then part of this trust is also based on traditional testing. A traditional approach might danger the benefits from an agile testing approach. The schedule might slip and also reliability in functionality and value;
6. Often you hear that the first 2 a 3 sprints are used to define structure, process, communication and mutual understanding. On the other side you hear that an agile approach is also very suitable for small projects (not only large projects) which needs quick delivery. If already 2-3 sprints are lost nothing will be delivered within a small period of 3 months although the business needs the functionality already now. This might turn into mistrust. The project will fail after all or cost more money
7. Experience testers are focused on traditional testing approaches and traditional testing techniques and demanding traditional documentation. If this is not part of the definition "Done" they might be using time to gain that information instead of supporting the team and adapt their approach to identify the value which is delivered.
8. Agile developers and traditional testers might not be able to adapt their selves to the team goals as in human nature there is some suspicion towards new and old approaches. they are not looking how they can support each other instead they are spending time to convince each other
9. When agile approach is new for the business then the team values the proof that their approach is working above the transparency of their work. Metrics are tuned on to this proofing activity. As testing is often used of proofing the quality of products, often they are also used to proof the process. Metrics are presenting a more ideal situation of the project instead of the value for the business

I'm sure there are more reasons why Agile might fail and even more reasons why it will become a success. The points mentioned above are just some thoughts of which I can think of why it might fail after all.

No comments:

Post a Comment