Failure Traps of Automation Project

90% of Automation Projects Fail to Give Promised Value. They keep getting trapped in the same old mistakes.

Here are common traps to avoid if you want to succeed in automation: 

1. Don’t underestimate the cost of automation.

  • Underestimating the true time needed to do automation will eventually give a sense of automation taking too much time to your stakeholders.
  • Setting correct estimates and processes is important. 

2. Don’t underestimate the need for training / learnings.

  • Maybe you know the tools and languages in your toolchain but is your team on board too? If not, better account for the time required for training / learning.
  • The team may need training on topics like clean code, etc. 

3. Don’t expect to be more productive over the short term.

  • Don’t be over-optimistic with your productivity.
  • Account for team meetings, org meetings, fests, and other things that will reduce your actual work time. 

4. Don’t spend all your time and effort on regression testing.

  • Regression is one facet of delivering value via automation.
  • What about adding value by automating test data generation, environment setup, ci scripts, etc?
  • In short: Go beyond regression testing. 

5. Don’t use instability of code as an excuse.

  • Unstable code is a shock for anyone coming from a dev background.
  • Building a stable code is part of your responsibility as a tester.
  • Ask for testability/automatability at the start. Don’t wait till the end and then give an excuse. 

6. Don’t put off finding bugs in order to write test cases.

  • Automation is important but stakeholders came to you for “testing” first.
  • Stop prioritizing test scripting over raising bugs in the product.
  • Give them feedback first and then work on your scripts. 

7. Don’t write simplistic test cases.

  • Don’t waste your time validating that a field should be empty by default or that the text of the submit button is “Submit”.
  • Focus your time and effort on tests that actually matter.
  • Don’t just automate anything for the sake of it. 

8. Don’t shoot for 100% automation.

  • 100% is a dream number in testing world.
  • Don’t use it for automation too.
  • Be realistic, Don’t set wrong expectations. 

9. Don’t use record / playback tools.

  • I know they find it easy to create scripts.
  • Easy is a disaster in the long run.
  • Don’t cut corners. Record and playback is a failure recipe (Time Tested) 

10. Don’t create test scripts that won’t be easy to maintain in the long term.

  • Think of modularity, reusability, and design principles.
  • Don’t create a short-term mess for rotting in the long run. 

11. Don’t make the code machine / user-specific.

  • No hard codes or reckless assumptions in the scripts.
  • Keep things configurable

12. Don’t fail to treat this as a genuine programming project.

  • Automation = Development.
  • Whoever doesn’t believe it is living in a fairy world.
  • It’s a separate project on its own. 

13. Don’t “forget” to document your work.

  • The next new engineer in your team doesn’t know the “obvious” things that you do.
  • Document the work or wait for them to recreate and re-engineer the same thing again and again.
  • Document, PLEASE!!! 

14. Don’t deal unthinkingly with legacy project code.

  • Old codes can be messy.
  • Don’t assume.
  • Don’t make minor modifications assuming no side effects. Better test it before making a change in production. 

15. Don’t insist that all your testers be programmers.

  • Not everyone is meant for everything.
  • Let people work on their strengths.
  • Otherwise, you will have a team of “JUGADU” automators. 

16. Don’t ignore library updates and their effect on your automation.

  • Sometimes, you may face breaking changes.
  • Sometimes, you will face library bugs.
  • Be open to realities 🙂 

17. Don’t forget to clear the fantasies that might have been spoonfed to your management by some vendor.

  • Marketers can create fake illusions for management disrupting their expectations.
  • Clear any such fantasies. Be realistic! 

Hoilllaa! That’s all!

Credits: Cem Kaner, Software Test Automation Paper. 

Leave a Comment

Your email address will not be published. Required fields are marked *