Data discrepancies

Data discrepancies are differences in numbers that happen when the same metrics or events are measured in multiple analytics tools but the numbers don’t match exactly. For example, page views, conversions, or audience sizes might appear higher in one tool and lower in another, even though both are tracking the same website or campaign.

Data discrepancies are common when comparing personalization and analytics tools. They usually result from differences in definitions, identity models, session logic, consent handling, or event timing, not from missing or incorrect data.

The key is to compare equivalent metrics, over the same time range, using finalized reports.

NOTE: Small variations between numbers across tools can be expected. But if you are seeing unexpectedly large differences, please follow the troubleshooting checklist under the Quick checks section before contacting support.

It’s common to see different numbers when comparing our tool with other analytics tools, such as Google Analytics (GA4). Below are the most frequent scenarios that can occur between our tool and, for the purpose of this article, GA4. 

Visitor or user counts don’t match

You may notice that visitor or user numbers differ across tools, especially over longer time ranges.

• We report Total Visitors as the sum of unique visitors per day. A visitor returning on multiple days is counted once per day.

• GA4 reports Total Users as unique users across the entire selected date range.

Why this happens:
Each tool defines and aggregates users differently. Over 7, 30, or 90 days, Contentful numbers often appear higher. This is expected and does not indicate missing data.

Session counts are significantly different

Session counts are one of the most common sources of discrepancies.

• Contentful Analytics sessions always end after 30 minutes of inactivity.

• GA4 sessions default to 30 minutes but can be extended up to 7 hours and 55 minutes.

Why this happens:
If GA4’s session timeout has been increased, it will report fewer sessions than our tool for the same user behavior.

Events appear in one tool but not another

Some events, such as clicks or page views, may appear in one tool but be missing in another.

Common reasons include:

• Events firing at slightly different times.

• One tool loading earlier or later than the other.

• Users navigate away before both tools capture the event.

• Ad blockers or browser privacy settings block client-side tracking.

Why this happens:
Client-side tracking depends on timing and page lifecycle. Fast navigation or early exits can cause one system to miss events.

User identity and profile counts differ

Different tools identify and group users in different ways.

• Contentful Analytics assigns each visitor a persistent stable_id to build a profile across sessions.

• GA4 relies on device-based identifiers and consent-dependent signals.

Why this happens:
Identity resolution works differently in each system, which directly affects user and profile counts.

One tool consistently reports higher numbers

In some cases, one tool may always show more traffic or events.

This can happen when:

• A Consent Management Platform (CMP) allows one tool but restricts another.

• Ad blockers or browser privacy settings block client-side tracking.

• One tool runs server-side or at the edge, while another runs client-side.

Why this happens:
Tools that are blocked less often or run server-side typically capture more data.

Use the following troubleshooting checklist to narrow down the most common causes of data discrepancies, before reaching out to support:

1. Confirm you’re comparing the same time range.

   1. Make sure both tools use the same date range, and check whether time zones match.

   2. Avoid comparing real-time data with historical or processed reports.

2. Verify you’re comparing equivalent metrics.

   1. Confirm the metrics mean the same thing in each tool (for example, visitors vs users).

   2. Check how sessions are defined and whether session timeouts differ.

   3. Ensure event names and definitions match across tools.

3. Review consent and ad-blocking behavior.

   1. Check whether a Consent Management Platform (CMP) is in place.

   2. Confirm both tools are treated the same under consent rules.

4. Look for implementation issues.

   1. Verify that tracking requests are visible in the browser’s Network tab.

   2. Check whether requests consistently include required identifiers.

   3. Watch for missing events, incomplete payloads, or errors.

When to reach out to support

If discrepancies are large, sudden, or growing over time, and the checks above don’t explain them, please reach out to support.

Why do I see discrepancies between Contentful Analytics and tools like Google Analytics?

The most common reason is because of ad blockers and how Google Analytics is treated compared to our tool. Google Analytics is often blocked by most ad blockers, while other analytics tools are not always blocked.

Other reasons include different ways of counting unique visitors or users. Contentful Analytics assigns each visitor a persistent stable_id to build a profile across sessions and be correctly counted as one visitor. Some analytics tools do not allow identifying users. As a result, a visitor could be incorrectly counted as multiple users.

How can I verify that a personalization or experiment is actually set up?

1. Check if the baseline <> experiment and variant <> experiment references exist in their respective content entry.

2. Check that holdout and traffic allocation is what you intended.

3. Check that audience membership for experiments is what you intended.

4. The system is event driven, so check the Live Events dashboard to see if component events and track events are coming through.

5. Check the privacy plugin configuration.

6. Open the website where the personalization or experiment is deployed, go to the network tools and check if you are sending events on viewing, or sending track events.

What should I do if my experiment results are inconclusive or show no clear winner?

Start a new experiment. All experiments start with a hypothesis. If the data does not prove or disprove your hypothesis, then the experiment is inconclusive and it needs a rethink of the parameters. We recommend using different content, a different part in the website, maybe boosting more traffic. Tweak the variables you have at your disposal and conduct a new experiment.

What are the most common mistakes that can affect experiment or targeting results?

The root causes are almost always:

• Events (component events, track events, page events, identify events) are not sent at all, or they are incorrectly sent. If that is the case, then the output is incorrect.

• Issues with the experiment setup. For example, incorrectly using data buckets, or bugs that come from using multiple environments.