Why your GA4 reports lie: data sampling explained

Google Analytics data sampling means GA4 builds a report from part of your data.

It then scales that part up to an estimate for all your traffic.

The verdict: sampled numbers look exactly like real ones, so check the data quality icon before you trust any report. Green means full data. Yellow means a subset you shouldn’t base a revenue decision on.

At Customer Impact we steer on customers and revenue, not on vanity numbers in a dashboard.

A sampled report is exactly that kind of number: convincing, and sometimes pointing the wrong way.

What is data sampling in Google Analytics?

It’s a shortcut: GA4 calculates a report from a representative sample instead of counting every event.

For a giant dataset that’s a logical choice, because it keeps your reports fast.

The downside is that the outcome becomes an estimate, not an exact count.

Why Google samples at all

Every report is a query on your raw events.

Big queries, such as two years of data with three segments, cost time and computing power.

So above a certain size, Google answers the question on a sample and multiplies the result. Google’s own wording: sampling may occur “when the number of events used to create a report, exploration, or request exceeds the quota limit” (Google’s data sampling help).

Where you notice it in practice

You mostly notice sampling when you dig deeper:

• an ad hoc analysis with segments
• a long period in one go
• complex filters or comparisons

A standard report without filters usually runs on full data.

Same data, processed differently, with different numbers as a result.

Making the numbers you steer on correct is exactly what our data analytics services cover.

When does GA4 switch to sampled data?

GA4 samples when an ad hoc query crosses an event threshold within the selected period.

It never happens at random.

And unlike the old Universal Analytics, GA4 counts in events, not in sessions. Our comparison of GA4 vs Universal Analytics explains that shift.

The thresholds in numbers

According to the documentation on data sampling in GA4, the limits per query are:

• Standard (free): 10 million events
• Analytics 360 default: 100 million events
• Analytics 360 maximum: up to 1 billion events, for more detail

Each step up is ten times the last one:

It’s about the query, not your total traffic

The limit counts the events inside the date range and query you request.

It’s not about all the traffic you’ve ever had.

A busy webshop hits 10 million events in no time. But a B2B site that pulls a year of data in one go, with several segments, can cross it too.

Where sampling shows up

Sampling mainly shows up in:

• Explorations with many dimensions and segments over a long period
• Standard reports with a comparison or filter that makes the query complex
• Long date ranges combined with detailed breakdowns

The standard overview reports you check daily are usually safe.

The risk sits in the moment you ask an interesting question, because that’s when you run the heavy queries.

Example: how far a small B2B property sits from the limit

Here’s a standard Traffic acquisition report from our own property, filtered on one source:

GA4 Traffic acquisition in our own customerimpact.be property, last 90 days, October 2026. 309 events are 0.35% of the total.

That puts the whole property at roughly 88,000 events in 90 days.

A light report is nowhere near 10 million.

So for a site our size, the danger isn’t the standard report.

It’s the heavy exploration across years and segments.

How do you spot sampled data in GA4?

Look at the data quality icon, the small shield at the top of your report.

Google’s data quality help puts it next to the title of a detail report, and in the top right of explorations.

What the colours mean

Google’s help for Analytics 360 says the icon “is red or yellow if the results are currently sampled, and green if unsampled” (unsampled explorations help). In practice:

• Green: the report uses 100% of the available data
• Yellow or red: GA4 applied sampling, so you see a subset

Read the percentage

Click the icon and Google shows what share of the data it used.

Google’s own example reads: “This report is based on 8.88% of available data.”

If a report runs on, for example, 40% of the data, the margin on your numbers is substantial.

The lower the percentage, the more your conclusion can wobble.

Here’s where to look in an exploration:

Sampling isn’t the only warning

The same icon can also flag thresholding.

Then GA4 hides rows that don’t meet minimum aggregation thresholds, so data appears only above that minimum.

That’s a different problem with the same lesson: read the icon before you read the numbers.

Make the check a habit

Glancing at that icon is one of the simplest and most valuable reflexes you can pick up.

It also matters if you want to know which channels deliver your best B2B customers. Those are exactly the reports where you want full data.

In a well designed marketing dashboard setup, that check is built in.

How much can sampled data distort your conclusions?

Far more than you’d think: sampling can turn growth into decline.

In an anonymised client example, the same regex on organic sessions told two opposite stories, depending on whether the data was sampled.

The same regex, two answers

The two views showed:

• Segment, sampled: +13% organic sessions year over year (6,754 versus 5,986)
• Filter, unsampled: a 9% decline year over year (6,012 versus 6,600)

One view says your organic traffic is growing. The other says it’s shrinking.

Same period, same site, same regex.

Side by side, the reversal is painfully visible:

What that does to a budget decision

Imagine raising your SEO budget because of that +13%, while reality is a 9% decline.

You’d be steering on a fiction.

That’s why we insist on reliable data before a single euro gets reallocated.

A dashboard that grows on paper but shrinks in reality costs you real customers.

Watch out: the sampled view was the optimistic one here.

A wrong answer that flatters you is the hardest to question.

If organic results drive the decision, our guide to GA4 for SEO shows which reports to trust.

How do you prevent sampling in your reports?

You can’t switch sampling off in the free version. But you can make it rare.

Work through these steps in order, from the cheapest fix to the most robust one.

Step 1: Shrink your date range

Request a shorter period and the events stay under the threshold, so GA4 doesn’t need to sample.

Rule of thumb: pulling three quarters separately is often more reliable than pulling a full year at once.

Step 2: Simplify your query

Fewer segments mean a lighter calculation. So do fewer dimensions.

Build your analysis step by step instead of cramming everything into one exploration.

Our walkthrough of cohort analysis in Google Analytics shows that approach, one question at a time.

Step 3: Use filters instead of segments where you can

As the client example showed, filters in standard reports sometimes run on full data while segments sample.

Test both. Keep the one that gives you the green icon.

Step 4: Export the raw data

For anyone who wants it truly exact, the BigQuery export of GA4 is the gold standard.

There you get every event row unsampled and calculate your own numbers.

Google’s BigQuery Export help calls the daily export “all the raw, unsampled event data”. Two limits matter for a Standard property:

• Daily export: capped at 1 million events per day
• Streaming export: no volume limit

This is the same layer a solid conversion tracking setup leans on.

Many teams then visualise it in Looker Studio.

Step 5: On Analytics 360, ask for more detail

Analytics 360 lets you pick “more detailed results” through the data quality icon in Explore.

That raises the sampling limit for that exploration.

Unsampled explorations exist too, but only “if the query would return more than 1 billion events”.

Most B2B companies never need 360 for that.

Which fix fits which situation?

Pick the fix by the question you’re asking:

The common thread: the lighter and more focused your question, the better your odds of full data.

Why does this matter especially for B2B?

Because in B2B, a handful of sessions can decide a deal worth thousands of euros.

You’d think sampling is only a problem for sites with enormous volumes.

For B2B it’s subtler, and therefore more treacherous.

Small numbers, big decisions

A B2B site often has relatively little traffic, but traffic that counts heavily.

Every session can be a potential customer worth thousands of euros.

That’s why small distortions make a big difference.

A gap between +13% and -9% across a handful of channels can decide whether you scale a campaign up or shut it down.

At a webshop with millions of sessions, such an error often averages out.

For you, there’s a concrete decision attached that directly touches your roas.

Three precise reports beat one broad one

We don’t work for webshops or e-commerce, but for B2B companies that steer on lead quality.

So we’d rather run three precise reports than one broad report that leans on a 40% sample.

A small, fast team that truly understands the data finds the truth quicker than a big agency that blindly trusts a dashboard.

Sampling is one of several data leaks

Sampling rarely travels alone. Missing consent, duplicate tags and spam hits, for example, distort the same reports.

Our checklists for GA4 tracking errors and data quality cover those leaks.

Want it checked for you? That’s what a tracking audit does.

Not sure your setup is sound? Start with what is tracking.

Frequently asked questions about data sampling in GA4

Can I switch off data sampling completely in GA4?

Not fully in the free version. You can avoid sampling by using shorter periods and simpler queries, or by choosing the least sampled option within an exploration. For guaranteed unsampled data you need the BigQuery export or Analytics 360.

What does the yellow shield icon in my report mean?

The yellow shield indicates that GA4 has applied sampling and that your report is based on a subset of the data. Click it to see what percentage of the data was used. A green shield means you’re looking at 100% of the data.

From how many events does GA4 sample?

According to Google, GA4 counts in events, not in sessions. The threshold for ad hoc queries sits at 10 million events per query on the free Standard version, and at 100 million events by default on Analytics 360 (expandable to 1 billion). It’s about the number of events within the requested period and query, not about your total traffic.

Are standard reports in GA4 sampled too?

Most standard overview reports run on full data. The risk of sampling arises as soon as you add comparisons, segments or complex filters, or build an exploration over a long period.

Is sampled data always wrong?

Not necessarily, a sample can sit close to reality. But you don’t know in advance how large the deviation is, and as the client example showed, that deviation can be big enough to flip your conclusion entirely. For decisions with budget attached, you’re better off working with full data.

Can a free GA4 property use the BigQuery export?

Yes. Google lists the daily and streaming exports for Standard properties, with a daily export limit of 1 million events. BigQuery itself is a Google Cloud product, so check its pricing before you start.

Stop steering on numbers you cannot trust

Sampled data looks just as convincing as real data, and that’s the danger.

Base a marketing decision on a 40% sample and you’re gambling with your budget.

The fix isn’t a complicated tool, but three reflexes: check the icon, simplify the query, pull raw data where needed.

We set up tracking and reporting for B2B companies so every decision rests on full data. Contact us and we’ll check your most important report first.