How ghost spam works
Analytics platforms accept data over an open collection endpoint. A tag on your site sends a request to it whenever a page loads, and the platform records what arrives. Ghost spam skips the site entirely: a script sends requests straight to the endpoint, carrying a property identifier and whatever referrer, page path or campaign name the sender chose. The visit never happened; the record of it is real.
The motive is usually advertising. The sender wants a site owner to look at their reports, notice an unfamiliar referring domain and visit it out of curiosity — the report itself is the delivery channel. That is why the fake referrer is often a memorable name or a slogan rather than a plausible website, and why the same domain shows up across many unrelated properties at once.
Why ghost spam matters
It corrupts exactly the numbers people act on. Sessions inflate, conversion rate falls because none of the fake visits convert, and channel reports suggest referral partnerships that do not exist. On a small site, a burst of it can outweigh genuine traffic and make a month unreadable.
It is also contagious across reports. Once it is in the property it flows into every dashboard, every client summary and every year-on-year comparison built from that period. And because it distorts engagement and conversion rate, any decision benchmarked on those figures — pausing a campaign, redesigning a page — is being made against a number that was partly manufactured.
Common mistakes with ghost spam
The first is treating it as a security problem. Nothing has been hacked, no server has been breached and the website is untouched; only the analytics property received junk. Calling in a developer to secure the site fixes nothing.
The second is blocking it at the server with firewall rules or blocked addresses. Ghost traffic never reaches your server, so there is nothing there to block. The third is confusing it with bot traffic — crawlers and scripts that genuinely load your pages. Those do hit the server, and they need a different remedy.
How to act on it
Judge it by the hostname. Genuine hits carry the address of the page that sent them, so a hit claiming a hostname that is not your site, or no hostname at all, did not come from your website. Filtering reports to your own domains removes most of it in one step, and doing that as an included filter is safer than chasing each new fake referrer with an exclusion rule you will be updating forever.
Leave your unfiltered data intact somewhere so you can always see what arrived before the rules were applied, and keep the platform’s own known-bot exclusion switched on for the other kind of junk. If a spike has already landed, note the affected period in your reporting rather than pretending the trend line is clean — anyone reviewing your analytics and tracking later will need to know why that month looks odd.
