Two analytics platforms tracking the same site will almost never report the same number of visits, sessions or pageviews. So, do traffic metrics vary between different analytics providers? Yes, routinely, and a gap between them is not by itself evidence that either one is broken. Each platform counts according to its own collection technology, metric definitions, configuration and implementation, which means both can be accurate on their own terms. The only thing you can say with certainty when the numbers don’t tally is that the numbers don’t match.
The question matters more in a large organization than on a single site. An enterprise estate often runs a corporate Google Analytics 4 (GA4) property, a second analytics platform, campaign and ad platforms, Search Console and a third-party traffic estimator side by side, with different teams owning each. When leadership asks which number is right, the honest answer depends on why the numbers differ. Below are the six most common causes, how analytics data differs from traffic estimates and search data, how to reconcile two platforms, and how to choose which source answers which question.
1. Different collection technology and data processing
Every analytics provider runs its own script. Most are JavaScript, but the scripts are not identical and do not operate identically. They differ in how data is collected when a visitor arrives on a page, how it is routed from the browser to the provider’s endpoints, and how it is processed and stored. No provider documents every detail of how another’s pipeline works, so some discrepancies can’t be traced to a single cause. They aren’t “trouble” in the sense of an error. The technologies are simply different.
2. Different metric definitions and settings
Metrics that share a name are often defined differently from platform to platform, and many definitions are configurable. Even with default settings on both sides, the counts won’t necessarily agree. If the criteria behind a metric differ, the data will too.
Session timeout. In GA4, a session ends after 30 minutes of inactivity by default, but admins can adjust the session timeout, up to a maximum of 7 hours and 55 minutes. A shorter timeout produces more sessions, and a longer one produces fewer. Two platforms with different timeouts will report different session totals for the same traffic.
Campaign changes mid-session. Platforms also differ on whether a new campaign source starts a new session. Universal Analytics started a new session when the campaign source changed mid-session. GA4 does not, which is one reason GA4 session counts can run lower than Universal Analytics did. Any comparison between GA4 and another platform has to account for how each one handles that case.
Filters and exclusions. Sometimes a metric is defined the same way in both platforms, but a filter in one leaves out an IP range such as internal users, or excludes visitors from a particular location. Filters aren’t visible in the data itself, so unless you know one has been applied in the settings, a gap in total visits looks like a fault.
Access levels. Users can also have different levels of access in different platforms, and some platforms restrict data more granularly than others. Someone who sees all data in one platform and only part of it in another will see contrasting numbers.
3. Different property and domain scope
Platforms decide differently what counts as one site. GA4’s page reports show the page path by default, so when several subdomains are tracked in the same property, site.com and sub.site.com can both appear under “/” as one homepage. A platform that reports the full URL, or that treats each subdomain as its own site, will show them separately. Siteimprove’s help center describes this as one of the most common reasons page-level numbers diverge. Before comparing pages, confirm that both platforms are tracking the same domains and grouping them the same way.
4. Implementation can vary
How a script is implemented on a website also changes what it collects. A script placed high in the HTML usually loads earlier than one placed near the bottom, especially on a heavy page with a long load time. If a visitor clicks through to another page before the first one finishes loading, the earlier script may record the visit while the later one doesn’t.
Vendors recommend different placements. Google’s guidance is to paste the Google tag immediately after the <head> element on every page. Siteimprove’s guidance is to place its analytics script just above the closing </body> tag. Both scripts load asynchronously, but the one in the head loads first, which can produce small discrepancies by itself.
Tag managers add another layer. They make it easy to fire different scripts at different times or under different conditions, and if the triggers differ between two analytics scripts, the data will too. Coverage gaps cause larger differences still. A script missing from some templates undercounts, and a script loaded twice on the same page can count one pageview twice.
5. Privacy choices and consent influence the numbers
Some visitors install ad blockers that stop an analytics script from running. If a blocker allows one provider’s script but not another’s, or a visitor allowlists one provider and not the other, the two platforms will see different audiences. Browsers differ too, since some apply stricter tracking protection by default than others.
Consent banners widen the gap. A script that waits for consent before loading never sees visitors who decline. GA4 can go further: with an advanced consent mode implementation and enough data, it applies behavioral modeling to estimate users and sessions from visitors who declined. A GA4 report may therefore include modeled figures that another platform, counting only observed visits, will never show. Teams that have moved toward privacy-first analytics should expect consent rates, not just tracking technology, to shape what each platform reports.
6. Bots and crawlers might be counted (or not counted) differently
Many analytics platforms exclude known bots and crawlers, but they use different exclusion lists, and some use none at all. Depending on how much automated traffic a site receives, those lists can change the reported totals substantially, and two platforms filtering against different lists will disagree even when everything else matches.
Analytics, traffic estimates and search data measure different things
Some of the sharpest disagreements aren’t between two analytics platforms at all. They come from comparing analytics with a source that counts something else, which is a common way metrics mislead teams.
| Source | What it counts | Why it won’t match your analytics |
|---|---|---|
| Tag-based analytics platforms, such as GA4 or Siteimprove.ai Analytics | Visits and pageviews on your own site, where the tag loaded and was allowed to run | These are your first-party numbers, but each platform applies its own definitions, filters and consent handling |
| Third-party traffic estimators | Modeled estimates of a site’s traffic, built from panels, clickstream data and other external signals | They never see your site’s own visitor data, so they are useful for relative comparison between sites, not for reconciling totals |
| Google Search Console | Impressions and clicks on Google search results, recorded by Google before the visitor reaches the page | A click can fail to become a session if the page doesn’t load, the visitor leaves first, or the tag is blocked |
| Ad platforms | Ad clicks, recorded at the moment of the click | The same click-to-session losses apply, plus differences in attribution windows |
Referrals from AI answer engines add another gap. An answer engine may cite a page, and an estimator may project AI referral volume, but analytics only sees the visitors who click through, and only attributes them correctly when the referrer reaches the page intact.
How to reconcile two analytics platforms
Reconciling two platforms doesn’t mean making them agree. It means ruling out the differences you can control so that the remaining gap is explained. Work through these checks in order:
| Check | What to compare |
|---|---|
| Date range and time zone | The same period, reported in the same time zone in both platforms |
| Scope | The same domains and subdomains, and the same pages carrying both tags |
| Filters and exclusions | Internal traffic, IP ranges, locations and bot filtering |
| Metric definitions | Session timeout, campaign handling, and what each platform counts as a visit or pageview |
| Consent and modeling | Whether either script waits for consent, and whether either platform reports modeled data |
| Implementation | Script placement, tag manager triggers, and missing or duplicate tags |
Then compare trends, not totals. If both platforms show the same weeks rising and falling, they agree on what matters for most decisions. A stable gap is usually a configuration difference. A gap that changes suddenly usually means something changed on the site, such as a template losing a tag, a new consent banner, or a replatform that broke measurement continuity.
Choose a system of record for each question
In an enterprise estate, the practical answer to “which number is right?” is a decision, not a calculation. Instead of forcing every platform to agree, decide which source answers which question and document it.
- Site traffic trends for leadership reporting: one designated analytics platform, used consistently over time.
- Search demand and search clicks: Google Search Console.
- Campaign performance: the ad platform for clicks and spend, and the analytics platform for what visitors did after they arrived.
- Competitive benchmarks: third-party estimates, read as relative indicators rather than counts.
Write down the definitions behind each metric, including session timeout, filters and consent handling, in a shared glossary. When marketing, IT, compliance and leadership read the same number the same way, a mismatch between platforms becomes a known quantity rather than a recurring argument.
Frequently asked questions
Do traffic metrics vary between different analytics providers?
Yes. Providers use different collection scripts, metric definitions, filters, consent handling and bot exclusion lists, and sites often implement each script differently. Some variation is expected even when both platforms are set up correctly.
Which analytics number should I trust?
Trust the source that was designed to answer the question you’re asking, and use it consistently. For trends on your own site, pick one analytics platform as the system of record. Use Search Console for search clicks and third-party estimates only for comparison between sites.
Why doesn’t GA4 match my other analytics platform?
Common GA4-specific reasons include its adjustable session timeout, the fact that a campaign change doesn’t start a new session, modeled data from consent mode, and a recommended tag placement at the top of the page head. Check those alongside the general causes above.
From matching numbers to trusting them
Analytics platforms look alike and often show metrics with the same names, but comparing totals across them is not reliable for the reasons above. Two platforms that disagree can both be correct, each tracking data according to its own implementation and setup. What matters is that the person analyzing the data knows exactly what each platform is counting and which one answers the question at hand. For platform-specific troubleshooting between Siteimprove.ai Analytics and Google Analytics, the Siteimprove help center keeps a step-by-step checklist.
