I have taken over a product which is close to 10 years old, it has a lot of interesting features which catered to the on premise needs. But I feel some of these features are not used by customers anymore. Currently I want to do a couple of things to improve the product's usability and also move it into the cloud. For this process, I need to collect data on feature usage by our customers. I would like to capture information like 1. What is the most used feature grouped by customer 2. What is the least used feature of the product a. Can I depreciate this feature so that I can reduce the maintenance overhead 3. What is the path(workflow) a user takes to use a particular feature. a. What is the most used entry point for using a feature b. Is there any default values or other aspects which will make things easy for the user.
Are there any tools which would help me collect this information? Are there any articles/references for understanding how people do this and any information on their experiences ?
Rgds, Siva.
You can also use the client to create custom events.
When you move things to "the cloud". Take a look at Sentry[1] for error monitoring. It'll show you errors and exceptions aggregated instead of you looking into logs for exceptions and errors. It has integrations with GitLab, Slack, etc.
Additional resources you might be interested in are Segment[2] and RudderStack[3].
One good way to get things going is to have a way to interact with your users. Having a button that says "Send us an issue" for when they have problems with an `href` that opens a mail client with a prefilled template and email addresses can be useful.
Complain!
Another useful thing is to have either Slack channels so people can ask for help there, or help each other out to solve their problems. Essentially building a community.(I'm biased against chatbots, but you may not have an allergic reaction).
Essentially, all these help focus on issues based on impact (positive or negative) and frequency.
When you get the 20th user complains about something, you know the next thing is to get that fixed if its impact is big.
I wrote a bit about this here[4], here[5], and generally about stuff here[6]
- [0]: https://posthog.com
- [1]: https://www.sentry.com
- [2]: https://segment.com
- [3]: https://rudderstack.com
- [4]: https://news.ycombinator.com/item?id=22827841