2010-07-22

How to start test automation project

Just imagine: you are a test automation engineer with 3+ years of experience. You know a couple of test automation tools and one of them you know the best. You also want to grow and one day a manager comes to you and says: "I found a good test automation project for you. You will be a leader, you will have your own team, you are started from scratch and you have to start ASAP. Do you agree?"

Of course, you agree! But you have never started the test automation projects from the very beginning. So, what to do the first?

In order to help with this task, I've created the following mind map:


I explain it a little bit:
1a. Get contacts of software testers, who own and create manual test cases for automation, customer, who would like to see automation (or maybe it is initiated by QA team, so in any way get the most interested person in automation) and developers, who can help in case of object recognition issues.
1b. Understand the goal of test automation project - for what purposes and what will be a success criterion.
2. Evaluate application, tool against an application, take a couple of test cases and try to automate them in order to understand potential difficulties/pecularities.
3. Estimate work. Based on the goal of test automation project and evaluation results, do next step and estimate how much time it takes to automate testing and how many people you need (if it is an option, of course :). The article "How to estimate test automation activities" will be useful.
4a. Think about and create test automation strategy and guidelines, which your team will follow.
4b. Create Prove-of-concept in order to show everyone that your strategy is working by exact example.

From my experience, I know that for the first week-two it will be enough for the good start :).

For further reading, I recommend the article Seven Steps to Test Automation Success - was published first more than 10 years ago, but still useful.

And I'm interesting, what will YOU do in the very beginning of new test automation project?

2 comments:

  1. Sounds like a good outline to follow. Although much depends on project size, etc. Few comments:

    3. Estimation will be rough (Especially if you are only 3 years (old) experienced). Communicate it to stakeholders clearly, also insist on common understanding and sign off of clear automation goals and criteria. I personally do not believe at this stage one would have clear scope, goals, etc to be able to use estimation method you propose (see above - it all depends...). You would be assigned to this tough project to propose an automation solution and strategy, if not only it is a case when they would like to hire an automation engineer to simply automate a list of tests.

    4a. A draft vision - yes, this will evolve going further.
    4b. One might try this before 4a, and maybe even 3.

    I would brain storm on few more ideas, such as analyse Development process/lifecycle used, timelines, available skillsets, technology used and dev skillset, etc.

    Offtop, A piece of good advice out there - try to keep everything as simple as possible, plan and build only what you need right now (although plan for flexibility where necessary) - in most cases you want your work results to be useful, right now. Everything changes, you can evolve accordingly if needed.

    P.S.
    http://www.io.com/~wazmo/papers/seven_steps.html - this is really great post by Bret, and I have been using it for quite some time on real projects and as support materials for conferences.

    ReplyDelete
  2. BTW, a great thing this blog is! Will follow it :)

    ReplyDelete