I am a bit worried by your question. If your main goal is test automation, you are doing it wrong! Automation cannot be the mission in any test strategy. I am not saying you shouldn’t do automation. I think any project could benefit from automation. But it needs to solve a problem. Automation in testing is never a solution in itself.
I would suggest a couple of heuristics for sustainable automation:
- make automation part of your daily work and a team responsibility
- tread the test code as production code
- have people working on it who have to right skills (automation = programming, what to automate = testing skill)
- think what feedback loops you need (what information you need to collect)
Also have a look at this:
- TRIMS - a mnemonic for valuable automation in testing by Richard Bradshaw
- Which Test Cases Should I Automate? by Michael Bolton
- Discussion here on the Club on “what to automate”.