# Review campaign analytics and logs

## Purpose
Find out how a campaign is going, question by question and person by person, and see
what the system did on its own.

## Who this is for
Account administrators, campaign managers, and anyone holding the campaign log or
report permissions.

## Where the numbers live

There are four different views and they answer different questions.

| I want to know… | Go to |
| --- | --- |
| How is this campaign doing overall? | Campaign **Overview** tab |
| How did people answer one particular assessment? | The chart icon on a deployment row |
| What is one person's result? | Participant row **⋮ → View scorecards** |
| What is the scored, weighted picture? | Campaign header **→ Scorecards** |
| What has been happening in the account? | Sidebar **Activity Logs** |

## Campaign overview

Open the campaign. The **Campaign Overview** tab carries five cards: **Active
assessments**, **Unique participants**, **Total submissions**, **Completion rate
(cumulative)** and **Avg completion time**. When there is too little data to average
honestly, the last card says **Insufficient data** rather than showing a misleading
number.

The "(cumulative)" in the fourth label is not decoration: it is the one card that does
NOT follow the filter bar.

Four controls above the cards re-scope everything below them:

- a date range: last 7 / 30 / 90 days, or a custom from–to pair;
- participant status: Invited, Active, Suspended, Completed;
- one assessment, or all of them;
- (in custom mode) explicit from and to dates.

The filter state is written into the page URL, so a filtered view can be bookmarked or
sent to a colleague.

## Per-assessment analytics

The small chart icon at the end of each row under **Active Deployments** opens
**Telemetry & Analytics** for that assessment. It shows **Respondents**, **Tracked
variables** and **Avg completion time**, then a **Question distribution** panel: one bar
chart per question, in the questionnaire's own option order (options nobody chose are
drawn at zero rather than dropped), with counts and percentages. Values the
questionnaire never declared are appended after them, largest first.

It measures **Avg completion time** for that assessment alone, the same way
as the campaign card: from when each submission was first opened to when it was
submitted. Submissions with no usable pair of timestamps are skipped, and if none
remain the card says **Insufficient data**.

## One participant's results

From the participants list, **⋮ → View scorecards** opens that person's report page,
which runs every participant-scoped report definition for them. **⋮ → Download XLSX
Report** produces a spreadsheet of their submissions instead.

## Campaign activity

**Logs** in the campaign header opens **Campaign Audit Trail**: this campaign's audit
trail, one sentence per row (**When**, **Who**, **What happened** and **About** whom),
with a search and a filter by kind of event. **Who** is the person's name where it can be
resolved; a respondent's own change shows their name from the roster; anyone else is
named for what they are: *A team member*, *A platform administrator*, *An API
integration*, *The platform team* (a platform script or set-up kit) or *Someone outside
this account*. **About** names the person by their roster name wherever the campaign
knows it (an invitation or an application is matched to the roster by address), with
the address in the tooltip; it shows the address itself only for somebody the campaign
does not know by name. **What happened** says which details changed where the record says
so (a segment such as *province or territory*, a *participant class*). **Export XLSX** in
the header downloads it.

Platform administrators also see a **Message delivery** section under it: counts for
**Waiting to send**, **Failed**, **Retried events**, **Total retries** and
**Dead-lettered**, a **Campaign notification events** table (time, what happened,
delivery), and a **Dead-letter recovery** panel for re-queuing events that gave up. Each
event reads as a sentence ("The results were published"); its event code and record id
are under **Technical details**. Account members do not see this section.

The account-level activity log is the same for every campaign at once.

## Account activity

Sidebar **Activity Logs** covers the whole account: the audit trail of member actions
(and, for platform administrators, message delivery). It has a CSV export (see
[Read the activity log and export it](/operate/tenant-admin/activity-log-and-csv-export)).

## Exporting

- **Export** in the campaign header opens a menu of three formats. Each builds its file
  server-side from the campaign's submissions, honouring the assessment / status / date
  filters currently set on the overview, and downloads it directly rather than assembling
  it in the browser, so a large wave never holds two copies of itself in the tab. Below
  1280px of width the three formats are in the header's **More** menu.
  - **Export as Excel (.xlsx)** downloads `campaign_<id>_report.xlsx`: the Summary,
    Participants, Submissions, Notes, Codebook and ShapeReport sheets, then, when the
    campaign's assessments carry Specialized set labels, one sheet per set with that
    set's answers for the responses that answered it.
  - **Export as CSV files (.zip)** downloads `campaign_<id>_report.zip`: the same sheets
    as one CSV file each (`summary.csv`, `submissions.csv`, `codebook.csv` and so on), so
    the codebook that explains every column travels with the answers. Each file opens in
    Excel with accented text intact.
  - **Export as JSON (.json)** downloads the same content as one JSON file, for scripts.
- Where an auditor recorded a correction, the exported answer is **theirs** (the value
  the scorecards use), with the respondent's original in a sibling column. Auditor notes
  travel too, unless the auditor unticked *include in export*.
- Where a question asked for an **uploaded artefact**, the export cell names the file and
  its reference together: `policy.pdf (asr-file:8f2c…)`. Two attachments on one question
  are separated by `|`. The reference is what identifies *which* `policy.pdf` out of a
  wave that may hold dozens; paste it into the campaign's **Respondent Evidence** list,
  or hand it to an auditor, to find the artefact. The file's contents are never in an
  export, and neither is a download link: a download link expires in about two minutes
  and an export outlives it by months, so a link in a file would be a link that looks
  like it should work and does not.
- **Export XLSX** on the campaign's **Logs** page downloads that campaign's audit trail.
- **Export all** on the Scorecards page downloads every visible scorecard as one XLSX,
  CSV or JSON file.
- Auditors can export their own caseload from the **Auditor Workspace**.

## Limits and gotchas
- Draft (in-progress) submissions are deliberately excluded from scoring, so a
  participant who has started but not finished does not move the scorecards.
- **Avg completion time** follows the filter bar, so widen the date range before
  concluding that a campaign is empty. **Completion rate (cumulative)** deliberately does
  not: it is the whole wave's rate whatever the filters say, which is why its label
  carries the word.

## Respondent evidence

The campaign's **Files** tab carries two lists. **Campaign Files** is what your team
uploaded. **Respondent Evidence** is what the wave sent back: every artefact respondents
attached to their answers, across the whole campaign (identified respondents and
walk-ups alike), newest first.

Each row names its **Respondent**, but only where there is one to name. A
respondent who arrived through an invitation or a roster entry is shown by the
same name and address the participants tab shows; a **walk-up**, who took the
assessment through the campaign's open link without identifying themselves,
reads *Walk-up*. That is a decision, not a gap: naming a walk-up would defeat
the point of the walk-up door, and the platform holds no name to print.

The list is read-only. There is no upload, because a member cannot author a respondent's
file; and there is no delete, because these artefacts are cited by answers that may be
under review, and the platform has no automatic sweep for them yet: a manual delete
button would be the only way one could disappear, and it would break a citation an
auditor is relying on.

Seeing the list needs the same **file viewing** permission as the rest of the files
browser. Auditors, who normally do not hold it, reach the same artefacts a different way:
through the submissions they are assigned to review (see
[Review answers and share feedback](/operate/tenant-admin/auditor-reviews-and-feedback)).

## Related pages
- [Reports and scorecards](/operate/tenant-admin/scorecards)
- [Read the activity log and export it](/operate/tenant-admin/activity-log-and-csv-export)
- [Publish results to a public link](/operate/tenant-admin/publish-results-to-a-public-link)
