Power My Analytics serves your reports from two places: a fast, pre-collected store of data called the cache, and real-time "live" requests sent straight to a platform's API. This article explains how each one works, when PMA uses which, and what that means for the numbers you see in your reports.

The Two Ways PMA Delivers Data

Every time you open a report, chart, or export, Power My Analytics has to answer one question: should it pull the numbers from data it already collected and stored, or should it go ask the source platform for fresh numbers right now? There are two answers:

1. Cached data: Data that PMA has already collected from your platforms and stored in its secure data warehouse. Because it is already on hand, it loads quickly.
2. Live data: Data that PMA requests directly from the source platform's API at the moment you run the report. It is the most current data available, but it takes longer to retrieve.

Most of the time you do not have to choose; PMA decides automatically based on what your report is asking for. The rest of this article explains how that decision is made.

How Cached Data Works

When you connect a data source, PMA does not wait until you open a report to gather your data. Instead, it collects your data ahead of time and keeps it ready in a data warehouse hosted on Google Cloud. (You choose a US or EU region during onboarding. For details, see Where Is My Data Warehoused?)

Cached data is built and kept current in two ways:

• Initial backfill: When you first connect a source, PMA automatically pulls in 2 years of historical data for most connectors.
• Daily sync: Every day at midnight in your selected time zone, PMA refreshes the most recent 7 days of data so your warehouse stays up to date.

Because this data is already stored and organized, reports that rely on cached data load almost instantly, your historical trends are preserved even if a platform stops serving older data, and your source platforms are not pinged hundreds of times a day.

How Live Data Works

A live request skips the warehouse and asks the source platform's API for the numbers in real time, right when you run the report. This gives you the freshest possible data, which is useful for fast-changing metrics or for seeing how today's campaigns are performing before the nightly sync runs.

The trade-off is speed: because PMA has to wait for the platform to respond, reports that use live requests take longer to load, and heavy use of live data can run into a platform's API rate limits.

Some fields are "live fields," meaning their values are never stored in the cache and are always fetched in real time. In the Schema Explorer, live fields are marked with a yellow lightning bolt icon (⚡) next to the field name.

When PMA Uses Cached Data vs. a Live Request

By default, PMA serves your reports from cached data, because it is fast and covers the history you have already collected. PMA switches to a live request only when something in your report requires real-time data. There are two main triggers.

1. A Live Field Is in the Report

If your report includes even one live field (⚡), PMA generates the entire report with a live API request, no matter what date range you select. A single live field overrides cached data for the whole report.

2. The Date Range Includes "Today"

For advertising (PPC) connectors, today's numbers have not been collected by the nightly sync yet, so PMA fetches them live. How this works depends on your date range:

• Date range excludes today (for example, last month): PMA serves cached data only.
• Date range is today only: PMA sends a live request only and bypasses the cache.
• Date range includes today and past dates (for example, last 30 days): PMA uses a hybrid request, fetching today's data live and joining it with cached data for the earlier days.

Amazon Ads is an exception. Because the Amazon Ads API is slower, including "Today" in your date range does not trigger a live request for that connector.

Quick Reference

A Special Case: Unique Metrics Paired with the Date Dimension

There is one situation that works the opposite of what you might expect. Some fields, such as Facebook Ads "unique" metrics, are normally live fields. But when you pair a unique metric with the Date dimension, PMA serves the report from cached data instead of sending a live request.

When a unique metric is paired with the date dimension, this will result in a request for cached data instead of a request for live data. For this reason, only cached data will appear in a report or chart that pairs a unique field with the date dimension.

The practical effect is that a report combining a unique metric with a daily breakdown shows the most recent cached values, not real-time numbers. For a full explanation, see Cached Data Requests in Facebook Ads with Unique Metrics and Date Dimension.

What This Means for Your Reports

Understanding which kind of request your report uses helps explain both its speed and its numbers:

• Performance: Cached reports load quickly. Reports that include live fields or today's data take longer because PMA waits on the platform's API. Heavy use of live fields is a common cause of slow loads and timeout errors in large exports. (See What to Do When a Report Is Loading Slowly.)
• Data freshness: Cached data reflects the last completed sync, so it may lag the platform by a few hours until the next nightly refresh. Live data is always current to the moment.
• Historical depth: Live fields cannot be backfilled, because PMA does not keep a stored history of their values. Cached fields preserve history even when a platform's API no longer offers it.

How to Pull Live Data on Demand

If you want to refresh a report with the latest numbers before the next daily sync, the method depends on your connector type:

• Advertising (PPC) connectors: Include today's date in the report's date range, then refresh the report.
• Other (non-advertising) connectors: Run a manual backfill from the hub, then refresh the report.
• Data Builder AI Insights: In the analysis date-range picker, enable Include today to incorporate the most recent live data.

Tips and Best Practices

• Use live fields only where you truly need real-time values; for most historical reporting, cached data is faster and more reliable.
• For large date ranges or big exports, removing unnecessary live fields will speed up loading and reduce the chance of timeouts.
• If a report seems to be "missing" today's data, check whether your connector pulls today's numbers live and whether today is included in your date range.
• Remember that pairing a unique metric with the Date dimension returns cached data, so use a non-date breakdown if you need the most current unique-metric values.

For additional assistance, please contact our support team.