Metric

What is workspace activity rate, and how do you measure it in Metabase?

Workspace activity rate is the share of eligible workspaces with enough recent human activity to count as active. A workspace can be a board, channel, site, space, folder, project, base, or document container.

TL;DR — Active workspaces ÷ eligible workspaces. Require a minimum number of meaningful events and contributors so bot noise does not inflate the rate.

Definition

Choose a window, qualifying event types, and a threshold. The example treats a workspace as active when at least two people generate five qualifying events in 30 days.

Data needed

  • Workspace inventory with archived/test flags
  • Activity events with workspace, actor, type, and timestamp
  • Human/bot or automated-event classification
  • Workspace owner, team, purpose, and lifecycle stage

SQL pattern

Share of workspaces active in the last 30 daysPostgreSQL
WITH activity AS (
  SELECT
    workspace_id,
    COUNT(*) AS qualifying_events,
    COUNT(DISTINCT actor_id) AS contributors
  FROM collaboration_events
  WHERE occurred_at >= CURRENT_DATE - INTERVAL '30 days'
    AND is_automated = FALSE
  GROUP BY workspace_id
)
SELECT
  ROUND(
    100.0 * COUNT(*) FILTER (
      WHERE COALESCE(a.qualifying_events, 0) >= 5
        AND COALESCE(a.contributors, 0) >= 2
    ) / NULLIF(COUNT(*), 0),
    2
  ) AS workspace_activity_rate
FROM workspaces w
LEFT JOIN activity a ON a.workspace_id = w.workspace_id
WHERE w.archived_at IS NULL;

Pitfalls

Counting automated events as engagement.→ Imports, bots, syncs, and bulk updates can dominate event volume. Exclude them or report them separately.
Assuming quiet always means abandoned.→ Reference libraries and stable policy sites can be healthy with low activity. Segment by workspace purpose before setting targets.

FAQ

Should I compare activity rates across tools?
Only after normalizing what counts as a workspace and a qualifying event. A Teams channel, Miro board, and SharePoint site have very different natural activity patterns.