From collect to action items. A single Insight Jira board where every signal — internal, external, observed, given, requested — is logged, qualified, grouped, weighed, and routed back to the people who raised it once it has been addressed.
The product backlog is fed by signal from many surfaces. All of them must converge into the same Insight board so they can be compared, grouped and weighed against each other.
- Market research, competitors, market watch
- C-level feedback
- Internal test-fest feedback, including developers' suggestions
- UI team's discovery process
- Features from other markets that trend to become the new norm
- Online final users' data analysis
- Customer service feedback
- Sales team feedback and request on what prospects are asking for
With automation and AI tools.
This feedback is logged in an Insight Jira board. Every time someone sees feedback or has a suggestion, it gets logged. The board is organized to make the volume actionable rather than overwhelming:
- Qualify each ticket with a label system (tags). Set up rules with "automation for Jira". Study the semantics of each ticket with an NLP tailored dev (or an existing tool purchased on the Jira marketplace).
- Group related tickets automatically and assess the frequency — hence the weight — of each. The mechanic: AI-generated parent tickets that aggregate manually entered tickets as child tickets. The key to making the tool usable is this grouping.
- Monitoring & alarms. Set up an alarm system when something gets too many requests: an email, a notification, an extra tag like Important, or the automatic edition of the priority level (e.g. P1). PMs run a weekly process that includes checking the related dashboard. The KPIs: number of new tickets logged, number of group tickets existing, the count of each by priority level, and the average age of the tickets (in weeks or months).
- Trackability back to the source. The people who raised or discussed an item get notified when the issue is fixed, or at least addressed. The loop is only closed when the originator hears back.
A filter on the Insight board.
There are three types of feedback from customers: Given, Observed, Requested. Taking the Given into account is a must. Taking the Observed into account is a should. Requested is a nice to have.
Given — must
Feedback the customer chose to send us, unprompted.
- Form in the Contact page, or another explicit open box for messages to the team
- Customer Care, After Sales services, Account managers
- Social networks, online comments and direct messages (a watch must be set up)
Observed — should
Feedback inferred from how customers actually behave.
- Full recordings of sessions
- Qualitative data analysis
Requested — nice to have
Feedback we proactively asked for.
- Experience review, online survey in a pop-in
- Product and order review post purchase
- From a mailing