Preparing a Jira Export to Analyze Software Delivery Flow
How to extract a clean CSV file from Jira Cloud or Data Center, select the right columns, and configure your delivery flow analysis in Cauba.
This guide helps you extract a clean CSV from Jira Cloud or Data Center without plugins or admin permissions, then configure your delivery flow analysis in Cauba in minutes. Only the columns you explicitly map are imported and retained, with built-in guardrails against personal data and free-text fields.
The essentials in 30 seconds
- Define the scope: rolling window (90d) for continuous flow or sprint list (Sprint in (...)) for Scrum, without mixing parent and child items.
- Export "current fields" as CSV: keep only Key, Issue Type, Status, Created, Resolved, Sprint (plus Started and Story Points if used) so descriptions, comments, and personal names stay out.
- Map the 2 mandatory fields and classify every status: Issue ID + Status (at least one status classified as "Done", and cancelled tickets classified as "Discarded").
- Check dates and sprints: local timestamps without time-zone offsets (Z / +01:00 unsupported), populated Resolved dates on Done tickets, and confirmed sprint order/dates when multiple Sprint columns exist.
1. Define the Export Scope and Protect Sensitive Data
Before running a query, verify three points: your export should match a single team’s scope (an entire Jira project or your board’s saved filter via filter = ...), focus on one consistent unit of work (never exporting parent items and their children together — such as Epic + Story or Story + Sub-task — to avoid double-counting), and cover the relevant timeframe (continuous flow window or sequence of sprints).
2. Use and Adapt the JQL Query
Pick the query matching how your team operates. The default query does not exclude any issue type automatically: add an issuetype filter only when needed to avoid mixing parents and children.
Example A — Continuous flow / recent activity (90 days): targets tickets resolved recently, created recently, or currently in Jira’s "In Progress" category. Limitation: an older unresolved ticket (> 90d) still in Jira’s "To Do" category is excluded even if it belongs to a recent sprint.
project = "MY_PROJECT" AND (resolved >= -90d OR created >= -90d OR statusCategory = "In Progress") ORDER BY created ASCproject = "MY_PROJECT" AND (resolutiondate >= -90d OR created >= -90d OR statusCategory = "In Progress") ORDER BY created ASCExample B — Sprint-based analysis: export every issue from the studied sprints
When analyzing a known sequence of sprints, filter directly on those sprints without adding status or creation-date conditions so older unstarted sprint items and carried-over tickets stay included:
project = "MY_PROJECT" AND Sprint in (101, 102, 103, 104) ORDER BY created ASCJira’s JQL autocomplete converts sprint names to unique IDs (quoted sprint names or closedSprints() / openSprints() also work). If your team does not equal one entire Jira project, replace project = "MY_PROJECT" with your board filter (e.g., filter = 12345) or a team criterion (component, label). Limitation: sprint membership shows which iterations a ticket belonged to, but does not by itself reconstruct initial sprint commitment or timestamped mid-sprint additions/removals.
| JQL Clause | Explanation & Purpose |
|---|---|
| project = ... or filter = ... | Targets a single team or board without mixing multiple teams sharing one Jira project. |
| issuetype (optional) | E.g. AND issuetype in (Story, Bug, Task) to keep a single work unit without mixing parents and sub-tasks. |
| resolved >= -90d (or resolutiondate) | Includes tickets finished over the last 90 days even if created earlier (Lead Time, Cycle Time, P50/P85, Throughput). |
| OR created >= -90d | Includes recently created tickets regardless of status. In Cauba, Completion Rate covers tickets created within the analysis window; this clause does not cover older tickets pulled into a recent sprint. |
| OR statusCategory = "In Progress" | Includes older tickets in Jira’s system "In Progress" category. Note that this category depends on Jira’s workflow setup and may differ from statuses you classify as "Active" in Cauba. |
| Sprint in (101, 102, 103, 104) | Exports all tickets attached to the listed sprints without creation or status filters, preserving repeated Sprint columns on rollovers. |
3. Export the CSV from Jira (Cloud & Data Center)
On Jira Cloud
- Open Filters > View all issues (or Search work items) and switch to "JQL" mode.
- Run your query and switch the display to "List view".
- Open "Columns" and select only the relevant fields: Key, Issue Type, Status, Created, Resolved, Sprint, plus your start-date field and Story Points if used.
- Click "Export" > "Export CSV (current fields)". If a needed custom field cannot be added to the view, use "Export CSV (all fields)" and leave unneeded columns unmapped in Cauba.
On Jira Data Center & Server
- Open Issues > Search for issues in advanced JQL mode.
- Switch to "List View" and select only the relevant columns via "Columns".
- Click "Export" > "CSV (current fields)" and save the .csv directly without re-saving it from a spreadsheet app (which may alter encoding or dates).
4. Verify Fields and Classify Statuses in Cauba
Cauba pre-fills recognized header matches and remembers for your account the column associations validated in your previous analyses:
| Typical Jira column | Cauba field | Requirement | Role in calculations | If missing or partial |
|---|---|---|---|---|
| Issue key / Key | Issue ID (issue_id) | Mandatory | Uniquely identifies each ticket, merges duplicate rows, and updates the team dataset across sequential imports. | Import rejected without unique Issue ID; rows with blank Issue ID are rejected. |
| Status | Status (status) | Mandatory | Maps every status into Cauba’s 4 categories (at least one "Done" status required). | Mandatory to run the analysis. |
| Created | Created Date (created_at) | Recommended | Start boundary of Lead Time (whole calendar days) and entry window for Completion Rate. | If unmapped: Lead Time unavailable. Blank rows excluded from Lead Time. |
| Started / In Progress Date (dedicated field) | Started Date (started_at) | Recommended | Start of active work: Cycle Time, P50/P85 delivery times, spread (P85 − P50), and upstream wait share. | Never falls back to Created Date: if missing, Cycle Time and P50/P85 are unavailable. P50/P85 require at least 10 complete Done items. |
| Resolved / Resolution Date | Completed Date (resolved_at) | Recommended | End boundary for completed items: Lead Time, Cycle Time, P50/P85, Kanban Throughput, and completion sprint attribution. | If unmapped: durations, P50/P85, and sprint rollovers are unavailable (Throughput then needs dated sprints). |
| Issue Type | Issue Type (issue_type) | Optional | Descriptive breakdown of delivered work by type (Story, Bug, Task…). | Overall flow metrics still computed without work-type breakdown. |
| Sprint (single or repeated columns) | Sprint Name (sprint_name) | Optional | Velocity and iteration rollover distribution (0, 1, 2+ rollovers) once sprint dates are known. | Without Sprint: continuous weekly flow analysis (Kanban). |
| Sprint ID, Sprint Start Date, Sprint End Date | Sprint ID and Sprint Dates | Optional | Provides sprint boundaries directly when Sprint Name maps to a single column. | Not mappable with multiple Sprint columns: a dedicated step lets you order and date the sprints instead. |
| Story Points / Story point estimate | Story Points (estimate) | Conditional | Used only for Velocity in "Sum estimates" mode. All other metrics count items. | Hidden by default ("Count completed items"); becomes mandatory in "Sum estimates" mode. |
Classify Every Status into Cauba’s 4 Categories
Cauba never guesses a status from its name and imposes no workflow order. Every detected value must be assigned to one of 4 categories (saved for future team imports):
Mandatory rule: 100% of present statuses must be classified and at least one status must be classified as "Done".
| Cauba category | Meaning | Jira examples | Impact on metrics |
|---|---|---|---|
Not started NOT_STARTED | Ticket created or planned, active work not started. | Backlog, Open, To Do, Selected for Development | Excluded from WIP and delivered counts; included in total created tickets for Completion Rate. |
Active ACTIVE | Work actively underway (dev, review, QA, verification). | In Progress, In Development, Code Review, In QA, Testing | Feeds WIP (tickets currently "Active" during the period). |
Done DONE | Work fully delivered per the team’s definition. | Done, Closed, Resolved, Released, Deployed | Feeds Throughput, Lead Time, Cycle Time, P50/P85, Velocity, Backlog Forecast, and Completion Rate numerator. |
Discarded DISCARDED | Cancelled, rejected, or duplicate closed without delivery. | Cancelled, Won’t Do, Rejected, Duplicate, Abandoned | Excluded from delivered work and WIP (never distorts Throughput or durations), kept in Completion Rate denominator. |
5. Check the CSV Before Import: Illustrative Structure Sample
This 5-row fictional snippet illustrates the expected CSV structure (headers, status categories, and repeated "Sprint" columns) — it is not a large enough sample for a full statistical diagnosis:
| Issue key | Issue Type | Status | Created | Started | Resolved | Sprint | Sprint |
|---|---|---|---|---|---|---|---|
| PAY-101 | Story | Done | 2025-01-06 09:15:00 | 2025-01-07 10:00:00 | 2025-01-14 17:20:00 | Sprint 24 | - |
| PAY-102 | Bug | Done | 2025-01-08 11:30:00 | 2025-01-08 14:00:00 | 2025-01-10 16:45:00 | Sprint 24 | - |
| PAY-103 | Story | Done | 2025-01-07 14:00:00 | 2025-01-10 09:00:00 | 2025-01-23 18:00:00 | Sprint 24 | Sprint 25 |
| PAY-104 | Story | Cancelled | 2025-01-09 10:00:00 | - | 2025-01-15 11:00:00 | Sprint 24 | - |
| PAY-105 | Story | In Progress | 2025-01-18 08:30:00 | 2025-01-20 09:00:00 | - | Sprint 25 | - |
How Cauba reads it: both "Sprint" columns map to "Sprint Name", ISO timestamps (YYYY-MM-DD HH:mm:ss) are auto-detected, "Done" → Done, "In Progress" → Active (PAY-105 in WIP), and "Cancelled" → Discarded (PAY-104 excluded from Throughput and durations). Once Sprint 25 dates (e.g. Jan 20–Feb 2) are entered and Sprint 24 dates (Jan 7–Jan 20) confirmed, PAY-103 is attributed to Sprint 25 via its Completed Date (Jan 23) and counted as 1 rollover. Sample thresholds: P50/P85 need at least 10 completed items with start and end dates; throughput variability and Monte Carlo need at least 3 complete periods.
6. Troubleshooting Common Issues
If rows are rejected or indicators appear as "Unavailable", check these five items:
Deepen Your Indicator Interpretation
Once your Jira export is imported into Cauba, explore our methodology guides to interpret your measurements:
Cycle Time vs Lead Time: Definitions, Formulas, and Distributions
Learn how to interpret the gap between Lead Time and Cycle Time, why arithmetic averages are skewed by outliers, and how to read your P50 and P85 percentiles.
Sprint Spillover & Delivery Predictability
Understand what the carried-over ticket distribution (0, 1, 2+ rollovers) actually measures, its boundaries, and how to improve flow stability.
