Analytics
Analytics holds the organization's numbers, on boards you build yourself. A board is a grid of widgets, and each widget charts one or more data sources over a time period.
Which project a board counts
A board created while a project is open counts that project only, whatever the project picker says later. A board created with no project open counts whichever project is open when you read it, or every project. The line under the board's name says which: Counting Storefront only or Counting every project.
Starting a board
Create New Board — or the + at the end of the board tabs, once there are boards — asks for a name, what to start with, and who sees the board: Shared with everyone in the organization, or Private to you. It starts as an empty board, or as a starter board of five widgets for one role. A starter board left unnamed takes the name in the table.
| Starter board | Named | What is on it |
|---|---|---|
| Tester | My testing | My open bugs, my untested cases, blocked cases, my bugs marked Done, hunt progress |
| QA lead | Quality | Automated pass rate, flaky automated tests, open bugs by severity, requirement coverage, bugs reported vs closed |
| Release | Release readiness | Milestone readiness, open bugs by severity, bugs reported vs closed, hunt progress, overdue milestones |
| Leadership | Leadership | Five numbers: automated pass rate, bugs caught, automation coverage, hours explored, manual time saved |
The widgets are ordinary ones: edit or remove any of them afterwards. The Starter plan allows three boards of five widgets each; if a board has room for fewer, it says how many were added. Standard and Business have no limit on either.
Creating a widget
Open a board and choose Create (Create New Widget on a board with none yet). The dialog has the form and a live preview — beside it on a wide screen, under it on a narrow one: the widget drawn from your real data, exactly as it will appear on the board. It updates as you pick sources, chart types and periods.
-
Name: the heading on the widget.
-
Chart type: Line, Bar, Pie, List or Number. A type that cannot draw what you picked is greyed out, with the reason written under the buttons.
-
Data sources: up to six, with a search box above the list. A source that cannot join what you have already picked is greyed out, with the reason on its own row.
-
Width on the board: Small — a quarter of the row on a laptop or any wider screen — or Medium, twice that. This changes the board only. The preview is always drawn at one width, so you can compare as you edit.
:::note Recently changed There used to be three widths. The old Small was too narrow to read on a laptop, so it is gone: every widget kept its width, under a new name — the old Medium is now Small and the old Large is now Medium — except the old Small ones, which grew to the new Small. :::
-
Time period: see below. A source that is usually read as it stands today, such as My open bugs, starts the widget on Right now.
-
Milestone (optional, shown when the project has milestones): counts only the bug hunts, sessions and automation runs of one milestone instead of Every milestone. Sources that have no way to reach a milestone — test cases, unstable tests, requirements, errors, Failing automated tests and the milestone sources themselves — are greyed while one is chosen. A scoped widget names its milestone in its header.
Create widget (Save changes when editing one) stays off until the widget can be saved, and the footer says what is still missing.
Time periods
Days are cut in your own timezone: "today" is the day on your clock, not the server's.
| Period | Covers | Drawn in |
|---|---|---|
| Right now | Everything, as it stands | One figure per source |
| Today | Midnight to now | Hours |
| This week | Monday to today | Days |
| Last week | The whole of last week, Monday to Sunday | Days |
| Last 30 days | The last 30 days, today included | Days |
| Last 3 months | The same date three months ago, to today | Weeks, Monday to Sunday |
| This year | 1 January to today | Months |
| Last year | The previous calendar year | Months |
| Custom range… | The two dates you pick, both included | Days up to 31 days, weeks up to 120, months beyond |
When a period starts partway through a week or month, the first bar or point covers only the part inside the period. Its label is the day the period starts.
Right now has no time axis. A count becomes its total so far, a breakdown shows everything in its current state (Test cases by state for the whole library, not only the cases written this month), and a line chart is not offered.
Three kinds of data
Every source is one of three kinds, and the kind decides how it can be drawn and what it can share a widget with.
- Counts over time: how many things happened per day, week or month, such as Automation runs or Bugs found in sessions.
- Breakdowns: how many things fall in each category, with no time axis, such as Test cases by state or Session bugs by severity.
- Rankings: the top rows by a number, such as Slowest runs — five, or eight for Milestone readiness and Hunt progress. A ranking of runs or tests shows when each row happened, because runs of the same suite share a name. A list that was cut short says so: Top 5 of 12.
What each chart type draws
| Counts over time | Breakdowns | Rankings | |
|---|---|---|---|
| Line | One line per source | — | — |
| Bar | One set of bars per source | One bar per category, grouped by source | Horizontal bars |
| Pie | Two or more sources, not percentages: one slice per source, its total over the period | One breakdown: one slice per category | — |
| List | One row per source, its total over the period | One breakdown: one row per category | The top rows, with their dates where they have one; a row opens the run, hunt, milestone or error it names |
| Number | One source: its total, with its change on the period before and a small trend line | One breakdown: all its categories added up | — |
What can share one widget
- Counts over time combine freely, up to six, as long as they are measured the same way: percentages and counts never share an axis.
- Breakdowns combine only when they count the same categories. For example, Test cases by automation state and Unstable tests by automation state can share bars. Test cases by state and Session bugs by severity cannot, because their bars would not line up. A pie or a list shows one breakdown; use Bar to compare two.
- Rankings stand alone: one per widget.
- A Number shows one source.
- The three kinds never mix in one widget.
A widget saved before these rules (or created through the AI connector) may hold a combination the editor now refuses. It draws what it can and names the rest under the chart (Not drawn: …). Edit the widget to split them.
The change on a Number
A Number compares the period with the one just before it, of the same length:
| Period | Compared with |
|---|---|
| Today | Yesterday |
| This week | The same days last week |
| Last week | The week before |
| Last 30 days | The 30 days before |
| Last 3 months | The 3 months before |
| This year | The same months last year |
| Last year | The year before |
| Custom range… | The same number of days just before it |
The change is green when it moves the good way for that source (more tests passing, fewer open bugs), red when it moves the bad way, and grey for a source with no good direction. A percentage changes by points, not by a percentage of itself. Right now has nothing to compare with.
Data sources
Some sources are dated by when something was created. Others are dated by when it was last updated, because Sekee stores no separate "finished at" time: Sessions completed, Bug hunts completed, Hunt test cases executed and Hunt results. Most sources have a line of description under their name in the dialog.
Test cases
- Test cases added: every test case written, automated or not.
- Automated tests added: test cases created in the period that are now marked Automation: Done.
- Awaiting automation: test cases created in the period that are now marked Can be done or Partially.
- Outdated test cases: test cases created in the period that are now Outdated.
- Test cases by automation state and Test cases by state: breakdowns of the test cases created in the period.
Unstable tests: counted by the day the flag was raised, and only while it is still raised. Clearing a flag removes it from every period.
- Tests flagged unstable
- Unstable automated tests
- Unstable tests with no assignee
- Unstable tests by automation state
- Unstable tests by state
Sessions
- Sessions held: by the date each session was scheduled for.
- Sessions completed: by the session's last update.
- Sessions by status
- Bugs found in sessions
- Session bugs by severity and Session bugs by status
- Session comments
Bug hunts
- Bug hunts started: by the hunt's planned start.
- Bug hunts by status
- Hunt test cases executed: cases with a result other than Untested, by their last update.
- Hunt results
- Bugs found in hunts
Test cases removed from a hunt are not counted.
Automation
- Automation runs
- Automation runs by result: Passed, Failed (any failed test or step) or Running.
- Automated tests executed, passed, failed and skipped: summed over the runs in the period.
- Runs that fail the most: ranking.
- Slowest runs: ranking.
- Automated pass rate: passed ÷ (passed + failed), as a percentage; skipped tests are left out.
- Flaky automated tests: tests that both passed and failed in the period, most failures first.
- Failing automated tests: tests whose latest result is a failure.
- Slowest automated tests: single test results by run time.
- New errors: distinct failures first seen in the period.
- Top errors: distinct failures seen in the period, ranked by how often they have happened in all. A row opens Error Tracking on that one error, with its details open; See all in Errors under the list opens the page over the same period where the Errors page has it (Today, Last week, Last 30 days, Last 3 months, Last year), and over all time otherwise.
- Errors by defect type: distinct failures seen in the period, split into product defects, test defects and the ones nobody has classified yet.
- Product defects found and Test defects found: distinct failures first seen in the period that are the application's fault, or the automation's — broken locators, test code, timeouts.
- The five error sources need error tracking, which is on the Business plan; on other plans they are greyed in the dialog.
- CI time and CI time on failing runs: total run time.
- Manual time saved: each passed or failed run of an automated test, times the estimate of the manual test cases it is linked to (1–5 min counts as 3, 5–10 as 7.5, 10–20 as 15). An estimate: its description in the dialog says how it is worked out.
My work: read as the person viewing, so one shared board shows each person their own.
- My open bugs: bugs you reported that are still open or in progress, by severity.
- My untested cases: cases assigned to you in active hunts with no result yet, by hunt.
- My bugs marked Done: bugs you reported that were marked Done in the period, worth a retest.
Bugs: hunts and sessions together unless the name says otherwise.
- Bugs reported and Bugs closed (moved to Done or Won't fix): put them on one line chart to see whether bugs arrive faster than they close.
- Open bugs by severity, All bugs by severity, Hunt bugs by severity, Hunt bugs by status.
- Time to resolve bugs: median days from reported to Done.
- A bug closed before Sekee recorded when a status changed is dated by its last update in Bugs closed, My bugs marked Done and Time to resolve bugs.
- Hunt bugs per folder: by the folder of the test case each was found against.
- Bugs per milestone: a bug counts under every Milestone its bug hunt or session is in, so the bars can add up to more than the bugs.
Bug hunts also has Blocked cases (in active hunts, by hunt), Hunt progress (share of each active hunt's cases with a result, least done first) and Bug hunts completed. No source breaks numbers down by person: boards measure the work, not the people doing it.
Sessions also has Hours explored and Bugs per session hour.
Test cases also has Test library size (a running total), Automation coverage (marked Automation: Done ÷ every case not archived, as the library grew), Test cases by estimate, Awaiting review (written or changed by an AI assistant and not approved), Test cases approved (with an approver, as the review queue counts them), Share written by AI and AI actions (changes made through an AI connector, from the activity log — Standard and Business; it reads zero on Starter, which keeps no log).
Requirements: Requirement coverage (passing, failing, blocked, untested or uncovered, as on the Requirements page), Uncovered requirements and Requirements added. A requirement filed for the whole organization rather than a project drops out of a project's count, as it does on the Requirements page.
Milestones: Milestone readiness (open milestones by the share of their readiness checks that pass, least ready first) and Overdue milestones.
A custom range reaches back ten years at most.