- This topic has 2 replies, 3 voices, and was last updated 5 years, 4 months ago by
.
-
Posts
-
Do you really need metrics for that?
When clients request some feature, they don’t have to provide metrics or even to think about them. They need it because it may bring them X or save time on Y.
Testing is about learning and finding information. Automation results are like indicators informing you. Gas tank indicator does not add value to the car, you can travel the same distance without it, but it informs you how much gas do you have and may help you to not get in trouble. Performance testing informs you how much customers your e-shop can serve simultaneously, but it won’t bring you a single customer on its own.So my take is that if you don’t need indicator on something or that particular indication does not justify the cost of acquiring it, than you need not automation there. It’s not about value. It’s about which things you need to keep indicators on.
Some measures that can help to know whether we are getting continuing value from automation are-
- How many bugs were found by automation in a given timeframe. If you are using tools like JIRA you can query this data
- How much manual effort has automated tests saved for testers. For example – If we need to fill up a 50 field form manually everytime we need to start testing, it will take more time than doing it in an automated way. The amount of time saved here with automation is good feedback
- How fast you can get feedback about the critical functionalities of the system
- If you have automated deployments, then the number of deployments that happened in a period of time is a good metric
And much more…
So, it boils down to what metrics are meaningful to you based on the context of your project
-Raj
- You must be logged in to reply to this topic.