Skip to main content
Data sources refresh automatically on a schedule that varies by data source — most cached sources update once or twice a day at fixed UTC times. Two refresh actions exist in the product, and they do different things:
  • Source refresh pulls fresh data from the source platform into the cache. Use it when the platform has new data the cache hasn’t picked up yet — for example after a manual correction in the source UI, or when conversion data has just been posted.
  • Dashboard refresh re-renders widgets against the existing cache. Use it when widgets look stale relative to the cache — for example after editing a widget or changing a setting.
Confusing the two is the most common refresh issue. The sections below cover each in turn. For background on why data sources refresh at different rates, see Refresh strategies.

Source refresh

A source refresh re-runs the data source’s fetch against the source platform and writes the result to the cache. Widgets then pick up the new values on their next render. There are two places to trigger a source refresh:
  1. The Refresh history entry point at the bottom of the Data sources page — the workspace-wide hub for refresh activity, and the only manual-refresh path for most data sources.
  2. The per-source Refresh tab inside data source settings — available on some data sources only.

Trigger a source refresh from the Data sources page

The Refresh history entry point at the bottom of the Data sources page is the workspace-wide hub for refresh activity. It is both the place to trigger a source refresh and the place to watch every refresh attempt across every connected data source.
1

Open the Data sources page

Navigate to Data sources in the top navigation.
2

Open Refresh history

Scroll to the bottom of the page and click Refresh history. The Source Refresh History page opens and lists every refresh attempt across all connected data sources, with status, start and end times, and any error message.
3

Trigger a source refresh

Use the manual refresh control on the page to pick a data source and a date range, then submit. The refresh runs in the background and appears in the history list with status pending, then processing, then completed or failed.
Because triggering a source refresh and watching its history happen on the same page, refresh history is built into the source-refresh flow — not a separate concept to navigate to.

Trigger a source refresh from the per-source Refresh tab

When already inside a specific data source’s settings — for example mid-investigation of one account’s numbers — the per-source Refresh tab keeps the refresh action and a focused history feed for that source in one place. The tab is available on some data sources only.
1

Open data source settings

Navigate to Data sources, find the source in the list, and click the settings icon on the row.
2

Open the Refresh tab

The Refresh tab appears in the settings drawer only if the data source supports it. If the tab isn’t present, use the Data sources page path above.
3

Choose what to refresh

Two refresh modes exist, depending on the source type:
  • File-style sources (for example File from Email, Google Sheets) — Select which file or collection to refresh.
  • Date-partitioned sources — Pick a date range and the collection or report to re-pull.
4

Submit and watch progress

Click Refresh. The history feed below the form updates every 15 seconds. A green badge with a job count shows while the refresh runs; the entry becomes completed or failed when finished.

Data sources where the per-source Refresh tab is available

The per-source Refresh tab is present on these data source families:
  • Organic social — for example Facebook Organic, Instagram Organic, TikTok Organic, LinkedIn Organic, X Organic, Pinterest Organic.
  • Public social — for example YouTube Public, Instagram Public, TikTok Public, X Public.
  • File from Email.
  • Adform Custom (not the standard Adform data source).
  • Google Sheets.
For all other data sources — including most major ad platforms (for example Meta Ads, Google Ads, TikTok Ads, LinkedIn Ads) — the per-source Refresh tab is not present. Use the Data sources page Refresh history path instead.

What refresh history shows

Every entry in Refresh history — whether from the workspace-wide page or the per-source tab — includes:
  • The refresh mode (file or date range).
  • Start and end timestamps.
  • Status — pending, processing, completed, or failed.
  • Origin — manual (user-triggered) or scheduled (automatic cycle).
  • Any error message returned by the source platform.
The per-source Refresh tab shows a feed scoped to that single source; the workspace-wide Refresh history page shows every entry across every source.

Quotas and rate limits

Most data sources that support manual refresh allow 1 manual refresh per day per data source. The cap exists because manual refresh hits the source platform’s API and the platform itself enforces rate limits. When the cap is reached, submitting again returns a rate-limit error. The next refresh becomes available at the start of the next UTC day, or whenever the scheduled refresh runs.

Data sources where manual refresh is not available

A small number of data sources do not expose any manual refresh control — they neither surface on the Data sources page Refresh history list nor expose a per-source Refresh tab.
  • Database data sources that query live on every dashboard load (for example BigQuery, PostgreSQL, Redshift, MySQL) — there’s no cache to refresh.
  • File pickup data sources that pull on a fixed schedule (for example Amazon S3, Custom URL Import) — schedule-only.
  • A handful of cache-only platforms where API rate limits make on-demand refresh impractical (for example Taboola).

Dashboard refresh

Customers trigger a dashboard refresh when the widgets on screen look out of date relative to the underlying cache — for example after editing a widget definition, after changing a setting on a data source, after a blending tweak, or returning to a dashboard tab that was open in the background. Refresh data in the dashboard header re-runs widget queries against the existing cache. It does not pull fresh data from the source platform. To trigger it, click Refresh data in the dashboard header. Widgets re-query the cache and redraw. Because dashboard refresh only re-renders against existing cached data, it does not surface numbers that are newer than the most recent source refresh. For fresher source data, trigger a source refresh first (using one of the paths above), then refresh the dashboard.

Troubleshooting

Three common causes: (1) the source platform hasn’t finalized new data yet — most platforms close out a day shortly after midnight in the account’s time zone, and a source refresh before that returns the same totals; (2) the data was already current as of the last scheduled refresh; (3) the dashboard widget is showing cached query results — click Refresh data in the dashboard header to re-render widgets against the new cache.
Manual refresh has a daily quota — typically 1 per source. After the cap is hit, retrying returns the rate-limit error. The next manual refresh becomes available on the next UTC day. Scheduled refresh runs are unaffected.
The per-source Refresh tab is only available on certain data source families — see the list above. For data sources without the tab, trigger source refresh from the Data sources page Refresh history path instead. Manual refresh is unavailable entirely for live-query database data sources (no cache exists) and for a few data sources where it’s disabled by design.
Widgets keep serving the most recent successful refresh. A failed refresh never corrupts cached data. The next scheduled cycle will re-attempt; trigger a source refresh sooner if the data is needed before then.
Source platforms can restate completed days inside their attribution window (commonly 7 to 30 days). Daily-refresh data sources re-pull a rolling window on every scheduled run to capture restatements automatically. See Refresh strategies for the full mechanics.
Related: Refresh strategies · Time zones & data freshness · Data history & retention