“Correctly structured” could be very open to interpretation. And it relies on what model of the check pyramid you need to use as there are a number of interpretations of it. Generally, it may be tough to quantify in numbers one thing that’s qualitative by nature.
If we go along with the traditional pyramid of “Unit, Integration, UI,” then it comes all the way down to discovering out numbers. In idea, the check pyramid actually factors to the quantity of check automation there may be in a venture. You’ll be able to lengthen that to think about different types of quantitative measures.
Listed here are some questions to think about to determine this out:
- What number of unit checks are there? e.g. 78 unit checks
- What’s the unit check code protection? Contemplate differing types like assertion protection, department protection, strains of code (loc) lined. e.g. 57% assertion protection, 25% department protection, 45% loc
- What number of integration checks are there? (And by integration, do you imply API stage or one other type of integration?) e.g. 25 integration checks, 0% automated.
- What number of UI checks are there? e.g. 200 handbook checks, 0% automated.
And if there may be a couple of code repository, you will have to run the numbers on each to get an correct evaluation.
Now, in case you are writing a report in your workforce/management on the present state of testing and check automation, I would lengthen this to incorporate another metrics. These will aid you decide if testing has been a precedence. Not all of those could also be related, just a few concepts to think about. Possibly it will aid you consider extra areas to evaluate and quantify.
- What is the go/fail price of any automated checks? In the event that they fail typically, are they catching bugs or are they false positives/negatives?
- What number of open bugs are within the backlog? e.g. 150 open bugs
- How previous is the oldest bug within the backlog? e.g. 3 years 5 months or # of days.
- Common age of open bugs within the backlog? e.g. 1.7 years
- What number of bugs are fastened/closed per dash?
Should you’re utilizing Jira, you will seemingly have to make use of a customized JQL command to seek out these numbers.
“leaving prospects annoyed”
- Is there a solution to quantify this in any respect? This one could also be tougher to collect details about. You could want to speak to customer support managers/results in assist get this information.
- Are you able to correlate the variety of open buyer assist tickets there are?
- Variety of calls to the decision middle?
- If prospects are canceling service, what number of over what time period?
- Can you discover out if the variety of consumer complaints go up/down after a software program launch to manufacturing?
