How do you get automation buy-in?

I’ve seen so many cases where people don’t want to automate or don’t want their team to automate because they see the initial setup as expensive, not realising that the long term cost saving would be worth it.

How do you get automation buy-in when you’re faced with people who don’t see the value of it?

2 Likes

(as a side project) create a proof of concept; for example automating one screen / functionality of the application you’re testing, and let it run regularly. Gather results from that and show it to your team at the right time (especially when bugs are found).
Gradually keep increasing the coverage (time allowing).

2 Likes

Treat it as a proper business case. My experience comes as people tries to sell automation as it will either “Solve all problems” or “it is a thing that you must have”. Neither case is super compelling to me to invest in it. So try to break it down. We are currently investing 1000h in testing / month. Our initial aim is first automate parts of that process to make that investment 900h / month. And we expect that to take 1000h for us to make so in 10 months we will break even and from then on we will start to save money on the investment. This is a much more constructive argument as now business can say things like I need a quicker return or a larger return or a smaller investment.
Further it also forces you to stop saying “we will do test automation” and instead discuss how you can increase the efficiency or the quality of the testing and to prioritize “biggest bang for the bucks” instead of automating something that you can automate. One of the most memorable things I have read on the topic was “What not to automate”. The main point is that a test have a certain time to live, as in how many times you want to execute the test before you need to change it. You typically want to make this investment in things that have a high time to live.
A few other good candidates for automation is processes that require many steps from the tester, complex execution patterns like timing based testing.
Also I would in general suggest to start with a productivity / efficiency angle first. Aim to save money.

4 Likes