You can feel when delivery is going well, but a feeling is hard to act on. Analytics turns the work in your workspace into a picture you can read in a few seconds: what the work is made of right now, how it has been flowing over time, and how each project is actually performing. It reads across every project you can see, and it changes nothing, so you can open it any time without touching a single issue.
Everything lives on one screen at /{workspace}/analytics, split into three tabs. Composition answers "what does the work look like right now". Flow answers "how is it moving". And Per-project lets you drill into a single project's velocity and timing. The views that used to live under Insights are folded in here as those same tabs, so there is one place to look.
The whole surface, at a glance
Here is the Per-project tab, the deepest of the three. A project picker sits up top (it defaults to your first project), a strip of KPIs summarises it, and two charts carry the detail: completed issues over time stacked per epic, and a histogram of how long issues took to finish. Switch the project in the picker to compare one against another.
32
Velocity / cycle
+3 vs prior
2.6d
Cycle p50
9.1d
Cycle p90
41
Completed
Throughput by epic
Cycle-time distribution
Three tabs, three questions
The tabs are not just filters over one report. Each answers a different question, so you can move from the current-state picture to trends to a single project without losing your place. Switch between them below (use the arrow keys once a tab has focus), and watch the chart change with the question being asked.
Each tab shows a representative chart: Composition opens on the overview tiles and the state-group donut, Flow shows completion throughput with cycle and lead time, and Per-project shows velocity across recent cycles. The figures are illustrative.
291
Work Items
152
Open Issues
128
Completed
14
Members
6
Projects
8
Cycles
Issue distribution by state group
- Completed128
- Started34
- Unstarted46
- Backlog72
- Cancelled11
Composition: what the work is made of
The default tab, and the fastest read. Overview stat cards count the essentials (Work Items, Open Issues, Completed, Members, Projects, Cycles). An issue-distribution donut breaks every issue down by state group, completed, started, unstarted, backlog, and cancelled, each with its own count. A "Completed (last 12 months)" area chart shows the shape of your finishing over the year, and "Most Issues Created" and "Most Issues Completed" leader lists surface top contributors when there is data to show.
The piece that earns its keep is Custom Analysis. Pick a dimension to group issue counts by, then read the result as a bar chart and an exact-number table underneath. You can group by:
- State Group, State, Priority, for how work is distributed across your workflow.
- Assignee, Label, Project, Cycle, Module, for how it is spread across people and structure.
- Created, Completed, Start, and Target Date, for how it lands across time.
Scope the analysis to All projects or a single project, then read the table below the chart for the exact figures. An "Issues by Project" stacked bar rounds out the tab, breaking each project's issues down by state.
Flow: how work is moving
Flow is about throughput and timing across the whole workspace. A KPI row reads Open issues, In progress, and Completed for the period. A completion throughput area chart covers the last 30 days. Distribution shows up two ways, "By state" as a donut and "By priority" as a bar chart, and a top assignees list shows who is carrying the work. Timing is summarised as Cycle time and Lead time, each reported at p50 and p90, so you see both the typical case and the long tail rather than a single average that hides it.
Per-project: one project in detail
Pick a project (the picker defaults to your first) and Per-project shows its detail. Velocity runs across the last six cycles. Throughput by epic stacks completed issues over time per epic, with anything that has no epic grouped under "No epic" so nothing is dropped. And a cycle-time histogram buckets how long issues took to finish, with p50, p90, and the average summarised beneath it. Change the project in the picker to compare one team's cadence against another's.
A single average hides the tail. A p90 next to a p50 tells you both the common case and the worst common case, which is the number that actually hurts.
When there is nothing to report yet
Empty states are treated as direction, not as zeroed dashboards. Analytics tells you what to do next instead of showing a wall of noughts:
No projects in the workspace
Analytics shows a "No project" panel with a Create a project button, rather than zeroed cards. There is nothing to report on until there is a project to report on.A chart with no matching data
Instead of an empty axis, the chart shows a small "No data" placeholder, so a blank region reads as "nothing here yet" and not as a broken chart.Custom Analysis returns nothing
You get "No data available for the selected filters". Widen the scope (switch the project picker back to All projects) or pick a dimension your issues actually carry: issues with no cycle contribute nothing to a Cycle grouping, for instance.
Reading it well
- Analytics only counts issues in projects you can see. If cards read zero but you know there is work, confirm you are in the right workspace and that you have access to the projects holding it.
- Charts read live data on load. If a contributor or trend looks stale, reload the page to refetch.
- Start on Composition for the current-state picture, move to Flow for trends and timing, and open Per-project only when you need to drill into one team's cadence.
None of this asks you to instrument anything. The issues, cycles, epics, and estimates your team already keeps are the input; Analytics just reads them back to you as composition, flow, and per-project detail, so a team lead can see how delivery is going without chasing a status update.