Back to blog
Engineering

Link clicks vs Google Analytics: why the numbers never match

Zayd
Nov 19, 2026·5 min read

It's one of the first questions anyone asks after running a campaign with short links. My link says 1,240 clicks. Google Analytics says 870 visits from that campaign. Which one is lying?

Usually neither. They count different things at different moments, and a handful of forces push them apart in predictable ways. Once you know what those forces are, the gap stops being worrying and starts being useful.

A click and a visit happen at different moments

A link platform counts a click when someone taps the link. That happens on our server, before their browser has loaded anything. Google Analytics counts a visit only once the destination page loads and its tracking script runs inside the visitor's browser.

Anything that goes wrong between those two moments shows up as a gap:

  • People leave before the page loads. On a slow connection, a few seconds of white screen is enough for someone to give up. We counted the tap; Analytics never saw them.
  • Ad blockers and privacy tools stop the analytics script. The person reads your page, but Analytics isn't allowed to record it.
  • Consent banners. If a visitor doesn't accept analytics cookies, many sites don't load Analytics at all.
  • Redirect chains drop tags. If the destination bounces through another redirect that strips the UTM parameters, the visit arrives without a source.

None of this means the click count is inflated. It means the click count sees people your website's analytics can't.

Link previews: the clicks nobody made

Paste a link into WhatsApp, Slack, LinkedIn or Facebook and a preview card appears with a title and an image. To build that card, the app's own servers visit the link first. Some do it once; some do it again every time the message turns up somewhere new.

Those visits come from software, not people, and a careless tool counts every one as a click. TRIMS recognises the crawlers that build previews for the big platforms, including Facebook, Slack, WhatsApp, Discord, Telegram and LinkedIn, along with search engine bots, and labels them as bots rather than people.

Bots that look like people

Preview crawlers are polite: they say who they are. Not every bot does. Email security systems are the big one. Many company mail servers open every link in an incoming email to check it for malware before the recipient sees the message. Send a campaign to a list full of business addresses and you can see a burst of clicks within seconds of hitting send, from machines in data centres, long before any human has opened the email.

Some of these scanners identify themselves and are easy to filter. Others deliberately act like a normal browser so that malicious sites can't hide from them, which makes them hard to tell apart from people. The giveaways: clicks arriving within a minute of sending, lots of clicks from the same data-centre network, and every link in the email being clicked at once.

On Pro, TRIMS reports bot and VPN traffic separately from human clicks, so you can judge a campaign by the people it actually reached.

Dark social: why so much traffic says direct

The other half of the puzzle runs the opposite way. When someone shares your link in a private chat, an email or a text message, the visit that follows often reaches your website with no referrer at all, because messaging apps and email clients frequently strip it. Analytics has nothing to go on, so it files the visit as direct, as though the person had typed your address from memory.

This is usually called dark social, and it's bigger than most people expect. A lot of what looks like loyal visitors typing in your URL is really people forwarding links to each other.

Short links help here because the click is counted at our end, where we can often tell which app opened the link even when your website can't. TRIMS also estimates how much of a link's direct traffic is likely to be dark social, so you can see how much sharing is happening out of view.

Smaller things that add up

  • Time zones. If your link platform reports in one time zone and Analytics in another, each day's numbers shift by hours, and daily totals disagree even when the weekly totals match.
  • Repeat taps. Someone taps a link, goes back, taps again. That's two clicks and, depending on timing, a single session.
  • Different ideas of unique. Unique visitors in a link tool and users in Analytics are estimated in different ways, so they'll rarely line up exactly.

Which number should you trust?

Both, for different questions.

  • Use link clicks to judge reach: how many people responded to the place where you shared the link.
  • Use website analytics to judge what happened next: which pages they read and how long they stayed.
  • Use conversions to judge money. A conversion reported by your own server and tied back to the click that started it is the one number here that ad blockers and consent banners can't touch.

A gap of 10 to 20 percent between clicks and visits is common and mostly explained by the things above. If it suddenly grows much wider, look for a broken redirect, a slow landing page or a burst of bot traffic. It's nearly always one of those three.

Two honest numbers that disagree tell you more than one number you can't check.
Zayd
Founder & Engineer, TRIMS

Building TRIMS — a fast, honest link platform with analytics that actually help you decide.

Next up
Introducing ML insights: forecasts, spikes & segments