Build Process
How We Estimate and Prioritize Issues
Status meanings in Linear
To do = issue that needs to be completed but that we haven’t started working on
In Progress = being worked on
In Review = dev says it is ready for strategist validation
Done = strategist validated it meets the client need and acceptance criteria
Backlog = issue that was originally part of the scope but that has been put aside to focus on higher priority work
Definitions
Before starting to work on any issues, the dev assigned to the project should:
1. Review the priority defined by the AI strategist
2. Ask clarifying questions
3. Estimate the number of points it will take to complete the task
An issue can’t be moved to “In Progress” until issue priority, and estimate points have been set.
When interpreting the meaning of each priority and when setting estimate points, we should follow the definitions below.
Priorities
At both client and issue level, priorities are defined by the strategist based on importance of the work (bugs, fixes) or time to deadline.
If multiple issues are considered of 'High' priority, we'll use the daily syncs to assess resources and/or prioritize a specific set of issues.
| Priority | Client impact | Aim to finish within |
| Urgent | Client is blocked, tool live but doesn't work, or hard deadline < 48h | 24 hours |
| High | Big impact for current project, not fully blocked but time sensitive | 2 working days |
| Medium | Normal planned work for this cycle, not blocking | 1 week |
| Low | Nice to have, polish, internal improvement | 2+ weeks, flexible |
Estimates
Estimates are (in most cases) entered by the specific dev that has been assigned to the project.
An issue can’t be moved to “In Progress” until the estimate has been entered for the specific issue.
| Points | Rough time | Description |
| 1 | up to 1 hour | Tiny change, no risk |
| 2 | 1 to 2 hours | Small, straightforward |
| 4 | 2 to 4 hours | Medium, likely same day |
| 8 | 0.5 to 1.5 days | Biggish task, may spill over to next day |
| 16 | 1.5 to 3 days | Big, should consider if it can be split into sub-issues |