The Figma activity audit: how to export and read Figma Admin activity logs

Summary

Why audit Figma activity

For most design organisations, Figma is where the work actually happens, and its activity log is the most complete record of how that work flows. A structured audit answers questions that otherwise rely on anecdote:

• Where is attention fragmented across too many files or projects?
• Which projects are moving steadily, and which have stalled or gone dormant?
• Who is building, and who is only watching?
• Which paid seats have had no productive activity for months?

Our 43-project benchmark found that file-switching measured from these logs was the strongest predictor of project outcomes among the factors examined, and that the signal is visible in the first three weeks. The audit below lets you compute the same measures yourself.

Before you start

You will also want your members list (exported from Admin) if you plan to audit seat use, and a simple mapping of Figma teams to your org structure.

Method 1: export a CSV from Figma Admin

1. 1

   Sign in as an organisation admin

   Activity logs are only visible to Figma organisation admins.

2. 2

   Open Admin settings

   From your organisation in Figma's file browser, open Admin (Admin settings).

3. 3

   Open the Activity tab

   Select the Activity (activity log) tab to see organisation events.

4. 4

   Set the date range and filters

   Choose a window (90 days is a good first audit) and, optionally, filter by event type or user.

5. 5

   Export to CSV

   Export the filtered log to CSV. Large ranges may need to be exported in monthly chunks.

6. 6

   Store the file securely

   The export contains personal data such as email addresses. Keep it in an access-controlled location and delete raw copies once processed.

Method 2: use the Activity Logs API

On Enterprise, the same events are available programmatically. Create an OAuth app owned by your organisation, request the read-only org:activity_log_read scope, and have an organisation admin authorise it.

import time, requests

TOKEN = "..."  # OAuth access token with org:activity_log_read
URL = "https://api.figma.com/v1/activity_logs"

def fetch_activity(start_ts: int, end_ts: int):
    params = {"start_time": start_ts, "end_time": end_ts, "limit": 1000}
    while True:
        r = requests.get(URL, params=params,
                         headers={"Authorization": f"Bearer {TOKEN}"})
        if r.status_code == 429:          # rate limited: back off and retry
            time.sleep(30); continue
        r.raise_for_status()
        meta = r.json().get("meta", {})
        yield from meta.get("activity_logs", [])
        if not meta.get("next_page") or not meta.get("cursor"):
            break
        params["cursor"] = meta["cursor"]
        time.sleep(1)

Anatomy of an activity log record

The CSV export and the API describe the same event in slightly different shapes.

The events that matter for DesignOps

Activity-log exports use event names such as fig_file_view. For flow analysis, file events matter most. Administrative events (membership, permissions, seat changes) matter for licence and security reviews.

The seven-step Figma activity audit

1. 1

   Define the question and window

   Start with one or two leadership questions, such as “Where are we most fragmented?” or “Which seats are idle before renewal?”. Use a 90-day window: long enough to see patterns, short enough to act on.

2. 2

   Clean the data

   Remove service accounts and integrations, drop exact duplicates, convert timestamps to UTC and always key files by Acted On ID or Key, because names change.

3. 3

   Map files to work

   Attribute each file to a project and team (from API context or your naming conventions), then to divisions. Every finding should be readable at project and team level.

4. 4

   Separate builders from watchers

   Anyone with create, rename, move, duplicate or export actions is a builder; view-only people are watchers. The benchmark found viewer-heavy teams scored lower on outcomes. Visibility is not contribution.

5. 5

   Look for fragmented attention

   Follow each designer's sequence of file events. Are they sustaining attention on a few files, or hopping between many? In the benchmark, file-switching was the strongest predictor of project outcomes. See the methodology for why it matters.

6. 6

   Look for stalls and rework

   Plot activity per project over time. Steady activity signals momentum. Long quiet spells, projects that never regain pace, and repeated bursts of rework that never settle are the visible forms of friction.

7. 7

   Check seats and report

   Join to the members list and list paid seats with no productive actions in the window. Report findings by team and project, with a short list of interventions, not a league table of people.

Interpreting what you find

Look at new projects' first three weeks first. In the benchmark, early switching, activity and momentum correlated 0.97–0.98 with the final picture, which makes them the most useful early warning you have.

Common pitfalls

• Keying on file names. Renames split one file into several. Always use the file key.
• Library and template files. Design system libraries attract views from everyone and inflate switching. Analyse them separately.
• Bulk moves during reorganisations. A burst of fig_file_move events is housekeeping, not work.
• Time zones. Mixed local times break session and active-day calculations. Convert to UTC first.
• Service accounts and guests. Integrations and external collaborators can dominate counts. Tag and exclude them.
• Viewer-heavy organisations. Large numbers of view-only users lower apparent switching for reasons unrelated to focus.
• Single snapshots. One export tells you where you are; a series tells you whether changes are working.

Privacy and governance

Activity logs identify people, so UK and EU GDPR apply. A defensible audit usually has:

• A documented purpose (operational health, capacity, licence management) and lawful basis, typically legitimate interest supported by an assessment.
• Restricted access to raw exports, with raw files deleted once aggregated.
• Reporting at team, project and division level, not individual rankings.
• Clear communication to staff, and consultation with works councils where they exist.
• Data minimisation: you only need actor, timestamp, event and file. Drop IP addresses if you don't need them.

FlowStatus follows the same approach. It stores activity metadata only, never file contents, and its leadership views are built around teams and projects. See how FlowStatus works.

Frequently asked questions