By: WKFGo Editorial · ·
Story points vs hours: two questions, two units
Story points answer "how big relative to our baseline?" Hours answer "how long did we actually spend?" Conflating them produces fake velocity, angry finance, and sprint commitments that ignore reality. Pick a primary planning unit; use the other deliberately.
Symptoms of mixed units
- Velocity "32 points" but CFO asks hours — PM converts with made-up ratio.
- Tasks estimated in hours labeled as points — trends meaningless.
- Capacity uses points; contractor billed in hours — double booking.
- AI estimates time from points without baseline — wrong every sprint.
Seven-step estimation policy
- Choose planning unit — points for product teams; hours for fixed-bid.
- Document on wiki — reference "1 point" task example.
- Store estimate consistently on task fields.
- log_time for actuals — time_summary and CHRO capacity.
- Never convert points→hours for execs — show both honestly or pick audience.
- Retro calibration — estimate vs log_time trends.
- estimate_task MCP — human review required.
Decision guide
| Context | Plan | Actual |
|---|---|---|
| Scrum product | Points | Optional hours |
| Fixed-price | Hours | Hours required |
| Agency hybrid | Points internal | Hours billing |
WKFGo supports estimated time, log_time, reports.
Anti-patterns
- Changing unit mid-release without history rename.
- Points because "devs hate timesheets" while finance needs hours.
- AI points without reference story.
How WKFGo helps
- Task estimate fields, log_time, time_summary
- Reports, workload_heatmap
- estimate_task MCP
FAQ
Both on one task? — Yes if roles clear.
Velocity for execs? — Trend only, stable team.
Jira import? — Map fields explicitly.
Kanban no sprints? — Hours or flow_aging often clearer.
Checklist before you close this topic
- One pilot project selected and named owner assigned
- Relevant WKFGo view or MCP tool run once with real data (not demo-only for decisions)
- One anti-pattern explicitly banned this sprint (written in team wiki)
- Follow-up calendar invite for calibration or metric review
- Soft links shared: /features for capability overview, /api-docs for automation, /blog for adjacent playbooks
Estimate honestly in WKFGo
Primary references and verification
This list supports further verification; inline links remain the source for a specific claim.