Prompt literature about software evaluation using statistics from technical support

Question

Helpdesk

I can’t find any articles or literature, maybe I’m looking for the wrong keywords ...

I want to extract some metrics from technical support that will allow me to evaluate some qualities of a software product and its components. Obviously, technical support can give some information about the quality of testing or the quality of documentation. But how to judge this quality (after all, everything is relative), what kind of metrics need to be included in the technical support system in order to accumulate useful statistics, there are probably not at all obvious metrics. Will technical support help you understand which module of the software product is most in demand, and which one is the most crooked? (although if the module is in demand, then there is more feedback on it, and not because it is crooked) ...

In general, I would be grateful if someone publishes an article on Habré or gives a link to existing Russian or English-language materials on the topic “technical support statistics in the service of a software development manager”

Answer 1

You can take into account:
1. How many calls to technical support were there (after the introduction of features, they began to write more, write less, did not change?)
2. How many questions are asked for each of the features?
3. How many bugs were found in each feature?
...
On the first question, in principle, it is easy to calculate. But it’s not clear whether the change in the number of emails is related to changes in the product, or is it seasonal? Or is it related to something else (prices were raised/lowered, discounts were given, a banner was hung on a popular site)? We keep such statistics, but they can show that the company as a whole is moving in the right direction.
But according to paragraphs 2-3, everything is more interesting: someone must calculate these values.
Will technical support mark somewhere what part of the product was the question? New, old? What type of question was - did not understand how to do it? Thank you? Asked for a new feature in addition to the existing ones? Such calculations will take time from technical support and reduce its effectiveness. If the information you receive is justified by the increased costs, then you can implement. Otherwise, consider whether it is necessary.
How many bugs have customers found? Everything is pretty simple here. We keep such statistics.
Everything happens as follows:
1. If a letter with a new problem has arrived, the technical support employee informs the tester about it.
2. The tester studies the problem, finds out whether it is a bug or not. If this is not a bug / not our bug (browser, java, flash player, dotnet, etc.), then we simply unsubscribe to the client with advice on what to do and how to get rid of it.
3. If this turns out to be our bug, then the tester opens a ticket in the bug tracker and indicates in a special field that the problem was found by the client.
4. At the end of the month, statistics are summarized: how many% of the total number of bugs were found by clients. I think 5% is a good value, but not all projects manage to achieve this. Somewhere and 1% happens, and somewhere 15%. Accordingly, it is clear on which project the developers wrote well, and the testers tested thoroughly, and where the team did not work very well. It practically does not require implementation costs, and the information is useful.