Skip down to the research record

Forecast data refresh limits: 9 figures to read each week

Quota Forecast Editorial Team · Published · 10 min read

Power BI allows eight scheduled refreshes per day, and up to 48 on Premium, PPU, or Fabric capacity (Microsoft Learn). The daily quota resets at 12:01 AM local time.

Forecast data refresh limits: the short answer

Your weekly forecast dashboard relies on specific technical constraints that determine how fresh your data is and what it can display. The nine core figures governing these systems include two Power BI refresh quotas, one scheduling example, three row and time limits for queries and models, one CRM stage cap, one export row cap, and one change search example. Understanding these published limits helps you interpret why certain data points appear or disappear during your forecast call.

The primary constraint for most users is the daily refresh quota. This limit applies to standard environments where the semantic model does not reside on premium infrastructure. This distinction is critical when planning your data extraction timeline, as the number of available refresh slots directly impacts how frequently your pipeline numbers update throughout the week.

Key takeaways

  • 48 refreshes on Premium capacity: If your semantic model resides on a Premium, PPU, or Fabric capacity, you can schedule up to 48 refreshes per day in the semantic model settings. Check which capacity hosts your dashboard to know which count applies. Microsoft Learn
  • One million-row DirectQuery limit: DirectQuery and live connection mode have a one million-row limit for returning data. If your forecast view exceeds this, the query will not return the full dataset. Microsoft Learn
  • 225-second query response limit: Queries in DirectQuery or live connection mode have a 225-seconds response time limit. Slow queries that exceed this threshold will fail or time out during the weekly call. Microsoft Learn
  • 1 GB incremental refresh suggestion: Consider using incremental refresh for semantic models larger than 1 GB or those that take several hours to refresh. Microsoft Learn
  • 100 stages for deal pipelines: Deal, ticket, and custom object pipelines can have up to 100 stages. Ensure your sales team’s stage definitions fit within this cap to avoid configuration errors. HubSpot

What each Power BI refresh figure measures

Power BI requires that data in reports and dashboards remains recent to support quick movement from data to insight, as noted by Microsoft Learn. Knowing how to refresh this data is often critical for delivering accurate results. The platform allows you to configure scheduled refreshes based on when your source files update. For instance, if a service updates a source file in OneDrive or SharePoint Online every night at 1:00 AM, you can configure the scheduled refresh for 2:30 AM. This timing gives Power BI enough time to complete the OneDrive refresh before starting the data refresh.

The documentation illustrates a refresh schedule operating on a 12-hour interval. When refreshing semantic models frequently, such as 48 times per day, query caching might not be as beneficial for associated reports and dashboards. Therefore, if your capacity supports high-frequency refreshes, expect less reliance on cached queries for immediate performance gains.

To apply these figures to your weekly forecast call, check the specific timing of your CRM extract. If your extract lands at 1:00 AM, set the Power BI refresh for 2:30 AM. Verify whether your semantic model refreshes 48 times a day; if so, note that query caching might not be as beneficial. Check the refresh interval in the semantic model settings. These checks ensure the dashboard reflects the most recent available data.

Row, query time and model size limits behind dashboards

DirectQuery and live connection mode in Power BI carry specific constraints that can affect how quickly your weekly forecast data appears. According to Microsoft Learn, this mode has a one million-row limit for returning data. Check whether any single query returns more rows than this.

The same source notes a 225-seconds response time limit for running queries in DirectQuery/live connection mode. This means any query taking longer than 225 seconds will fail or timeout. For a revenue leader checking a dashboard during a live call, this is a hard ceiling on how complex a single view can be. If your forecast view aggregates large historical datasets without optimization, it may hit this wall. Keep individual visual queries lightweight to stay under this threshold.

For larger models, Microsoft Learn suggests considering incremental refresh. This applies specifically to semantic models that are larger than 1 GB or that take several hours to refresh.

These three figures—row count, query time, and model size—define the technical boundaries of your data layer. If your dashboard relies on DirectQuery, verify that your most complex views return fewer than one million rows and respond within 225 seconds. If your model exceeds 1 GB, evaluate whether incremental refresh is feasible for your data source. These checks ensure your weekly forecast numbers reflect the latest available data without hitting engine limits.

CRM stage caps and export row caps

Your forecast dashboard pulls data from CRMs that enforce structural limits on how deals are organized.

Pipeline stage limits in HubSpot HubSpot distinguishes between standard deal pipelines and other object types when setting stage maximums. Deal, ticket, and custom object pipelines can have up to 100 stages according to HubSpot. This higher cap allows for granular progression tracking in complex sales cycles. However, appointment, course, listing, lead, order, and service pipelines can have up to 30 stages. If your weekly forecast relies on lead or service objects rather than primary deals, verify that your stage count does not exceed this lower threshold.

Export row caps in Stripe reports When you generate financial context for your forecast call, Stripe’s reporting tools impose hard limits on itemized output. An itemized PDF includes the same complete summary data and the first 5,000 rows of itemized Balance change from activity data, as stated by Stripe. If your itemized balance change data exceeds 5,000 rows, the PDF omits the rest. The PDF still includes the complete summary data, so use the summary for totals and do not treat its itemized rows as a full transaction list in high-volume months. For other export checks, see weekly pipeline export checks.

Archived items in Pipedrive Data hygiene directly affects how you interpret active pipeline volume. In Pipedrive, archived items count toward the archived item limit, not active usage, according to Pipedrive. Before the weekly call, review your archive logs to confirm that removed opportunities were correctly categorized, ensuring your active pipeline number reflects only live deals and not residual data clutter.

Change search examples in Close Monitoring recent activity helps validate data freshness. The Close help center illustrates filtering by searching for changes made within the past 5 hours. This is presented as an example of time-based filtering rather than a system-wide restriction on query duration. Use this timeframe as a benchmark for checking last-minute updates to deal stages or amounts before the meeting starts, allowing you to catch edits made by reps after their final status update. For a related check, see Close date change log: why the date moved.

Forecast data limits reference table with sources

The table below lists the specific refresh, row, and stage figures documented by each publisher. Each row pairs the limit with what it measures so you can verify your dashboard settings against the published numbers during your weekly review.

PublisherLimit FigureWhat It Measures
Microsoft Learn8 scheduled refreshes dailyDaily quota that resets at 12:01 AM local time
Microsoft Learn48 refreshes per daySemantic models on Premium, PPU, or Fabric capacity
Microsoft Learn1 million rowsReturned data in DirectQuery/live connection mode
Microsoft Learn225 secondsResponse time limit for running queries
Microsoft Learn1 GBSemantic model size threshold for incremental refresh
Stripe5,000 rowsFirst rows of itemized Balance change in PDFs
HubSpot100 stagesDeal, ticket, and custom object pipelines
Close5 hoursExample timeframe for searching recent changes

Note that the HubSpot documentation also specifies a cap of 30 stages for appointment, course, listing, lead, order, and service pipelines. The Close entry describes an example search window rather than a hard system limit. Query caching might not be as beneficial for semantic models refreshed often, such as 48 times per day, according to Microsoft Learn.

How to compare your own dashboard against each figure

Start by opening the settings for your semantic model. Microsoft Learn notes that while Power BI enables you to go from data to insight to action quickly, you must make sure the data in your reports and dashboards is recent. Check the scheduled refresh times listed in your workspace configuration. Write down the exact time your CRM extract lands on the server. Compare these two timestamps directly. Ensure the scheduled refresh runs after the source update finishes. If the refresh starts before the data arrives, the dashboard will not reflect the latest records. Adjust the schedule if necessary so the sequence is correct. For a fuller walk-through, see forecast refresh troubleshooting steps.

Next, review the row counts in your DirectQuery views. Look at the query logs to see how many rows are returned during a typical refresh. Keep this number well below the one million-row limit mentioned in other sections. Also check the query response times. If any query approaches the 225-second threshold, investigate the underlying logic or indexing. These checks help you stay within the operational boundaries of the platform.

Finally, examine your CRM pipeline configurations. Count the number of stages in your deal pipelines. Verify that deal, ticket, and custom object pipelines do not exceed the 100-stage cap. For appointment or course pipelines, ensure they stay under the 30-stage limit. This alignment prevents unexpected behavior when moving deals through different phases. By matching your internal setup against these published figures, you maintain a clear understanding of where your data sits relative to the system's defined limits.

Forecast data refresh limits: frequently asked questions

What are the forecast data refresh limits in Power BI?

According to Microsoft Learn, the scheduled refresh quota is eight times per day. The same source defines a one million-row limit for returned data in DirectQuery and live connection modes. These figures establish the baseline constraints for refreshing semantic models within the platform.

Does the 48 refreshes per day figure apply to my Power BI capacity?

The higher frequency applies specifically when the semantic model resides on Premium, PPU, or Fabric capacity. In those environments, you can schedule up to 48 refreshes per day in the semantic model settings. This expanded availability is distinct from the standard daily quota mentioned in general refresh documentation.

When does the daily scheduled refresh quota reset?

The quota of eight scheduled refreshes resets daily at 12:01 AM local time. This specific timestamp marks the boundary where the count returns to zero for the new day. Tracking this reset time helps align your scheduling with the available refresh slots.

How many stages can a HubSpot deal pipeline have?

A HubSpot deal pipeline can have up to 100 stages, the same cap as ticket and custom object pipelines (HubSpot).

Do archived Pipedrive deals count toward active usage limits?

No. Archived items count toward the archived item limit, not active usage, as stated by Pipedrive. This distinction separates archived records from those currently consuming active workspace capacity. Understanding this separation clarifies how archival actions affect your overall item counts.

Sources