UX: Determine how we should recommend actions to users in the moment (usage PQLs)
There are many paid features of GitLab that can make it easier for an individual user or their team to do their jobs but it's not always clear to the user that these features exist.
Example areas where recommendations around paid features could improve the users experience:
- A free namespace is utilizing SAST scanning, at a certain point in-app we should recommend a trial to test our advanced paid scanning options we have like FUZZ testing. At the right moment we should recommend to them they test out Ultimate to see how they can make their code even more secure.
- A free namespace with multiple users is actively submitting MRs but they have not setup code owners. At the right moment, we should recommend a trial and setting up code owners and required code owner review for MRs to increase code quality.
A bad example of this type of recommendation engine is Microsoft Words Clippy, which was trying to be helpful in making recommendations but almost always seemed to make them at the wrong moment and distracted the end-user away from the task they were trying to complete.
How can we potentially recommend these types of actions to users while they are using the GitLab product in a way that's not annoying and doesn't take away from the task they're trying to complete?
What are some good industry examples of recommendation engines that are actually helpful?
- Slacks slackbot that gives me a recap of my messages and which channels I have been inactive in for an extended period of time with the option to leave the channel.
- Searching for an Uber to/from an airport, if you don't book it they ask if you'd like to schedule a pickup for a particular date/time in the future