Guides
Why GA4 and your CRM's web analytics show different numbers
Published November 3, 2026 · 8 min read
You open Google Analytics 4 and see one number of visits. You open the web analytics in your CRM and see another. The usual reaction is to assume one of the tools is broken. It almost never is: each one counts with different rules, at different moments and over an audience that is not quite the same.
This guide explains where the differences come from, which ones are normal and which ones point to a setup problem. It is not an article against GA4: plenty of businesses run both tools side by side without trouble.
Key takeaways
- Two analytics tools never give identical numbers. A stable gap between them is normal; a gap that changes suddenly usually means something is broken.
- The main cause is usually consent: each tool only measures people who accept its cookie category, and every banner is set up differently.
- Definitions differ: what a session is, what a bounce is and in which time zone the day starts.
- GA4 is better for aggregate traffic and Google campaigns; analytics inside the CRM is better for knowing what each contact did and whether they bought.
- Compare trends, not absolute numbers.
Why do two analytics tools never match?
Every analytics tool is a script that runs in the visitor's browser and sends data to a server. Between the real visit and the number on your dashboard there are several decisions: whether the script loaded, whether the visitor consented, whether the server accepted the data, how it was grouped into sessions and on which day it was counted. Change any one of them and the numbers drift apart.
So the useful question is not "which number is right?" but "is the gap stable?". If your CRM always counts somewhat fewer visits than GA4 and both curves rise and fall together, both tools work. If one of them drops to zero one day, something is broken.
The most common causes of the gap
| Cause | What happens | How to check |
|---|---|---|
| Consent | Each tool only measures people who accept its cookie category | Check which category each script has in your banner |
| Cookieless mode | Without identifiers there are no sessions or unique visitors | Check which mode your site uses in the CRM |
| Blockers | Some ad blockers stop well-known third-party scripts from loading | Test the site with and without a blocker |
| Do Not Track and GPC | Some tools switch off when the browser asks them to | Check each tool's settings |
| Bots | Each tool filters automated traffic with its own list | Compare odd traffic spikes in both |
| Session definition | How much inactivity closes a session | Read each tool's documentation |
| Bounce definition | GA4 derives it from "engaged" sessions; other tools from page views | Compare the definition, not the percentage |
| Time zone | A day can start at local midnight or at UTC midnight | Compare weeks, not single days |
| Registered domains | Some tools drop visits from domains you have not added | Check the list of domains on your site |
Consent explains most of it
In Europe, most analytics tools need the visitor's consent before using cookies. If GA4 sits in one category of your banner and the CRM's analytics in another, each will measure a different group of people. Even with the same category, the moment each script reads consent can vary: one records the page where the visitor clicked "Accept" and another starts counting on the next one.
If you use Google Consent Mode, keep in mind that GA4 can also fill its reports with estimates (modelling) when there is enough data. A tool that only counts real, consented visits will always show lower numbers than one that estimates the missing ones.
Sessions and bounces do not mean the same everywhere
In Flowgration's web analytics, a session closes after 30 minutes of inactivity and a bounce is a session with a single page view. GA4 uses a different concept: an "engaged" session depends on time on page, conversions and the number of pages, and a bounce is a session that does not meet that bar. Both definitions are reasonable, but they are not comparable. A 60% bounce rate in one tool and 40% in another can describe exactly the same behaviour.
Days do not start at the same time
Flowgration's daily roll-ups use UTC days. GA4 uses the time zone set on the property. If you are not on UTC, late-evening visits can land on a different day in each tool. You notice it on a single day; over a week or a month it almost disappears.
What gets dropped on the server
Before storing anything, the CRM's analytics applies several filters: it drops known bots, rejects events from domains you have not registered on your site, keeps only the URL path (no query string or fragment) and stores nothing in consent mode if the visitor has not consented. On top of that, if the browser sends Do Not Track or Global Privacy Control, the Flowgration script does not run. Each of these filters removes visits that another tool may still count.
When does a gap point to an error?
Some signs cannot be explained by definitions:
- One tool drops to zero and the other does not. Usually someone removed or broke the snippet, or changed the cookie banner.
- The gap changes suddenly. If the ratio was stable for months and then jumps, look for a recent change: a new banner, a new template, a new domain.
- Specific pages are missing. If a section of the site does not appear in one tool, check that the snippet is on every template.
- A subdomain does not appear. Check that the domain is registered on the site and, if visitors move from one domain to another, that cross-domain tracking is set up.
- Visits arrive late. In Flowgration data is processed every 5 minutes, so a visit takes 5 to 10 minutes to show up. That is not a fault.
Which numbers should you compare, and which not?
Compare trends: if the week's visits go up 20% in one tool, they should go up by a similar amount in the other. Also compare the ranking of top pages and traffic sources. If your pricing page is the third most visited page and Google is the top source in both tools, both are telling the same story.
Do not compare bounce rates, average duration or unique users across tools. The definitions differ and the result tells you nothing useful.
Does it make sense to use both?
Yes, if each one answers a different question:
- GA4 is useful for aggregate traffic across the whole site, for the connection with Google Ads and Search Console, and for the reports your agency already knows.
- Analytics inside the CRM is useful for seeing what a specific contact did before getting in touch, which pages the leads who end up buying visit, and for starting automations from a visit.
If you use both, load them through the same cookie banner and list them in your cookie policy. And decide up front which one is the reference for each report: for example, GA4 for total monthly traffic and the CRM for conversions to lead.
Frequently asked questions
Which number is right, GA4 or the CRM?
Neither is "the right one". Each counts by its own rules. What matters is that the gap is stable and that you use each tool for the questions it answers best.
Why does my CRM count fewer visits than GA4?
Consent is the most common reason: the CRM's analytics in consent mode records nothing until the visitor accepts. Bot filters, Do Not Track and the estimates GA4 can add with Consent Mode also play a part.
Can I drop GA4 if I already have analytics in the CRM?
It depends on what you use. If you rely on Google Ads, Search Console or agency reports built on GA4, keep it. If you only look at visits per page and sources, the CRM's analytics may be enough.
Does cookieless mode count the same visits?
Not exactly. Cookieless mode has no identifiers, so there are no sessions, bounces or unique visitors. It counts page views and sources anonymously, so its numbers cannot be compared with a session-based tool.
How Flowgration does it
Flowgration's web analytics runs on its own first-party script that does nothing until the visitor accepts, unless you choose cookieless mode. It detects Cookiebot, CookieYes, OneTrust, Complianz and Google Consent Mode v2, stores URLs without query strings, does not geolocate by IP (the country comes from the browser language) and respects Do Not Track and GPC. Sessions close after 30 minutes of inactivity and daily roll-ups use UTC days.
What it does not do: it does not model visits without consent, it does not replace GA4's link with Google Ads and it does not record sessions on video. Anonymous page analytics starts on the Essential plan; visitor identification and heatmaps on Growth. There is a free plan and a 14-day trial with no card; details are on the pricing page. For the legal side, read the guide to web analytics with consent and GDPR.
Build your WooCommerce CRM in 60 seconds
Flowgration turns your store's orders into a living CRM and runs email, SMS and WhatsApp on autopilot. Free up to 500 contacts.
Start free