A client asks how many hours an eight-point ticket will take. Someone proposes a convenient rule: one point equals four hours.
That shortcut feels useful because it connects the backlog to a budget. It also makes a relative estimate look like a time commitment. Once the conversion becomes a promise, the team can spend more energy defending the number than learning from delivery.
Story points and hours can coexist. Give them different jobs and keep the relationship visible without pretending it is fixed.
What story points tell you
Story points express relative effort for a team. They can account for complexity, uncertainty and the amount of work involved. An eight-point item is a larger estimate than a three-point item within that team’s scale; it is not a universal unit of labor.
Atlassian’s estimation guide explains the relative nature of points. That is why comparing point totals across teams or using them to rank individuals is unreliable. Different teams can use different scales and definitions of done.
Historical throughput or velocity can inform planning when the team and working conditions are sufficiently comparable. It still describes a range of likely delivery outcomes, not a guarantee for every ticket.
What hours tell you
Recorded hours show elapsed labor time attributed to work. They can support delivery costing, capacity review and invoicing where the agreement charges for time.
Time estimates answer a prospective question: how much labor might this require? Time records answer a retrospective one: how much was spent? Neither should silently replace the other.
Jira supports estimation and time-tracking concepts, with available fields and behavior depending on configuration. Atlassian’s issue-estimation documentation is a useful reference when deciding which fields your team will use.
Why a fixed conversion breaks down
Suppose two tickets are both estimated at five points. One involves a familiar component with substantial implementation work. The other is a small change with an uncertain external integration.
The first takes 10 hours. The second takes 18 after an unexpected dependency issue. The point estimates may still have reflected comparable perceived effort and risk before work began. The actual hours explain what happened afterward.
You could calculate an average of 14 hours for these two tickets. That tiny sample does not justify turning every future five-point item into a 14-hour commitment. Even a larger historical relationship should be treated as contextual evidence, with a range and clear assumptions.
Mike Cohn makes the underlying distinction explicit in Don’t Equate Story Points to Hours.
Use a workflow with separate planning and billing records
Start by estimating backlog work in the unit your team finds useful. Separately, create the commercial estimate or budget using the expected delivery approach, roles, risk and scope. See how to estimate project hours for that part of the process.
During delivery, record actual time against the relevant project and issue reference. Include review, testing and coordination under clear categories. At the billing stage, apply the agreement: approved eligible hours for time-based work, or the agreed milestone and fee for fixed-price work.
| Record | Main purpose | Avoid using it as |
|---|---|---|
| Story points | Relative backlog estimation | An individual performance target |
| Time estimate | Labor planning and budget assumptions | Proof of time actually spent |
| Actual hours | Delivery evidence and eligible time billing | A substitute for completed outcomes |
| Invoice | Agreed commercial charge | An automatic translation of points |
Choose one authoritative time record
If people log time in Jira and a separate time platform, define which record is authoritative and how the connection works. Check whether entries are linked, copied or synchronized in your actual setup. Do not assume every integration behaves the same way.
Verify issue references, edits, duplicates, billable status and rate application before relying on the records for an invoice. A ticket being closed does not necessarily mean its time has been approved for billing.
Time Trakkr provides Jira-linked time workflows alongside approvals and reporting. Review the current features and configure the process so the team records each block of work once.
Common questions
Can we estimate directly in hours?
Yes, if that suits the work and team. The important distinction is between an estimate, actual effort and a commercial promise. Changing the estimation unit does not remove uncertainty.
Can historical points help with a delivery forecast?
They can, when interpreted within the same team and comparable conditions. Use an evidence-based range and revisit assumptions when scope, staffing or the definition of done changes.

