Say GA4 shows three times more visits than usual one day, but no extra enquiries come in. A spike like this rarely means a sudden surge of interest. More often it is bots, spam or your own visits. Here is how to spot fake visits, what GA4 already filters out and what you can do yourself.
What does GA4 do about bots on its own?
GA4 automatically excludes visits from known bots and search engine crawlers. It identifies them using Google’s own research and the bot list maintained by the IAB (Interactive Advertising Bureau). You can’t switch this off, and GA4 doesn’t show how many visits were left out.
So what reaches your data is bots that aren’t on the list, and spam sent straight to GA4 with your Measurement ID without anyone visiting your site. On top of that come visits from you, your staff and your developer, unless they are filtered out. These aren’t bots, but they skew the numbers in the same way.
In client reports I separate bot visits from real ones. Otherwise ad results would be judged on fake visits.
How do you recognise bot visits?
One sign on its own proves nothing. But when several of them show up at the same time, you are most likely looking at bots or spam.
- A sudden spike with no reason. Visits jump several times over in a day, but no campaign started and no newsletter or post went out that would explain it.
- Engagement close to zero. GA4 counts a visit as an engaged session if it lasts longer than 10 seconds, has a key event or has at least two page views. If the engagement rate of the extra visits is close to zero, there are no real readers behind them.
- Everything from one source. The whole spike is in direct visits (Direct) or comes from one unknown site.
- An unexpected location. Visits come from a city or country where you have no customers and no ads.
- An unfamiliar hostname. The hostname shows on which domain’s page the event happened. If it lists a domain that isn’t yours, it is spam.
The quickest way to see these signs is in a GA4 exploration (Explore). Choose a Free form exploration, add date, city, channel (Session default channel group) and Hostname as dimensions, and Sessions and engagement rate as metrics. If on the day of the spike all the extra visits come from one city with zero engagement, the cause is most likely a bot or spam.
When I monitor a client’s ad accounts, I regularly review GA4 and the website too, so changes like this don’t go unnoticed.
Where do fake visits come from and what can you do?
Go through the causes in this order. You can filter out the first two in GA4 settings before the data reaches your reports. The last two you can only deal with in your reports.
1. Your own and your developer’s visits
How to recognise it: the visits line up with your own or your developer’s work, for example the day the site was updated.
How to check: in GA4 Admin, open “Data collection and modification” and then “Data filters”. Check whether an Internal traffic filter exists and which state it is in. I always check this in a new client’s GA4.
What to do: mark your IP addresses as internal in the Google tag settings (“Define internal traffic”) and add the filter. A filter has three states. In Testing, GA4 only labels matching data with the “Test data filter name” dimension. In Active, GA4 excludes it permanently. It won’t reach BigQuery either, and you can’t get it back later. In Inactive, the filter does nothing. A filter takes 24–36 hours to apply, so leave it in Testing for a few days before activating it and check that it catches the right visits.
2. Spam from another domain
How to recognise it: the Hostname dimension shows domains that aren’t yours, or the hostname is empty. Someone sends this data straight to GA4 with your Measurement ID, and nobody actually visits your site.
How to check: add the Hostname dimension to an exploration and see which domains are sending data. When I check tracking, I usually start where the event happens. The hostname shows exactly that: whether the event happened on your site at all.
What to do: since 21 September 2026, GA4 lets you accept data only from your own domains. Open “Data filters”, choose “Create filter” and “Web hostname traffic filter”. As the filter operation, choose “Include only” and add all your domains. There are four match types: “exactly matches”, “begins with”, “ends with” and “contains”.
Before activating it, think about where else your data comes from: subdomains, a domain for another language or a separate shop domain. If one of them is missing from the list, its data will be excluded permanently. An “Include only” filter also excludes events with an empty hostname. It doesn’t apply to events sent through the Measurement Protocol (from a server). Leave this filter in Testing for a few days before activating it, too.
3. Spam posing as a referral
How to recognise it: an unknown site appears under referrals (Referral), sending short visits with no engagement.
What to do: the unwanted referrals list (List unwanted referrals) won’t fix this. It doesn’t remove any visits. GA4 simply stops showing that site as the source of the visit. The list is meant for cases such as a payment page that buyers return from to your shop. Check the hostname of the visits from that site. If it isn’t yours, step 2 helps. If it is yours, it is a bot that actually visited your site (step 4).
4. A bot that visits your site like a person
How to recognise it: the hostname is correct, but the visits arrive in bulk from one city and last a couple of seconds.
What to do: GA4 data filters let you exclude internal traffic, developer traffic and data from other hostnames. A bot that actually opens your page doesn’t fall under those filters and reaches your data. In an exploration you can use a segment to view such visits separately or set them aside, but they stay in the collected data. If there are a lot of them, it is worth checking whether your web host or CDN offers bot protection.
Judge your ads and your site by key events, meaning enquiries and purchases, rather than by the number of visits. Bot visits also lower your conversion rate, because the same number of enquiries is divided by a larger number of visits.
Practical tip
If GA4 shows a sudden spike, first check whether enquiries went up at the same time. If visits tripled and enquiries stayed the same, look for bots and spam before you start changing your ads.
What to do next
If you seem to have fewer visits instead, the cause may be cookie consent. When a consent banner is added to a site, I tell the client in advance that Analytics numbers may drop and that this doesn’t mean fewer visitors. The article “Cookie Consent and Consent Mode” explains why this happens.
Open a GA4 exploration and look at the last 90 days of visits by hostname and city. If you find a domain that isn’t yours, add a hostname filter in Testing. You can go through the rest of your setup with the GA4 and GTM checklist. If you would like me to review your GA4, see the GA4 setup and audit and Web analytics pages.
Sources
- Google Analytics Help: Known bot-traffic exclusion
- Google Analytics Help: Engagement rate and bounce rate
- Google Analytics Help: Filter out internal traffic
- Google Analytics Help: Filter web hostname traffic
- Google Analytics Help: What’s new in Google Analytics (21.09.2026)
- Google Analytics Help: Identify unwanted referrals
Facts checked against official documentation on 11 November 2026.
Richard Pikmets