Link Analytics and UTM Tracking: A Practical Guide
What click data can and cannot tell you, how UTM parameters work, and which numbers are worth acting on.
Link analytics has an unusual failure mode. The numbers are easy to collect, easy to display, and easy to look at every morning — and remarkably easy to learn nothing from. A dashboard showing four thousand clicks tells you almost exactly as much as one showing four hundred, unless you know what a click is, where it came from, and what you would do differently at either number.
This guide covers what click data genuinely measures, how UTM parameters attach a source to a visit, and which of the numbers in front of you are worth acting on.
What a click actually is
A click, as recorded by any redirect-based system, is a request to the redirect. That is the whole of it. It is not a person deciding to visit you.
The distinction is not pedantry. Several things generate that request without a human being involved:
- Link previews. Paste a link into a messaging app and it fetches the destination to build a preview card. That is a request.
- Security scanners. Corporate email gateways follow every link in an incoming message to check it. Sometimes more than once.
- Bots and crawlers. Some announce themselves honestly. Others do not.
- Prefetching. Browsers occasionally fetch a link before the user has decided to click it.
A reasonable analytics system filters the obvious cases, and no system catches all of them. The correct posture is to treat click counts as a strong relative signal — this link outperformed that link, this week beat last week — rather than as an exact headcount. Comparisons within your own data are reliable. Absolute numbers, quoted to a third party, are softer than they look.
How UTM parameters work
Analytics platforms cannot see where a visitor came from with any reliability. The referrer header is missing, stripped, or misleading often enough that it cannot be the basis of a report. UTM parameters solve this by having the link carry its own origin story.
They are ordinary query parameters appended to the destination URL. Your analytics reads them on arrival and files the visit accordingly.
| Parameter | Answers | Example | Needed? |
|---|---|---|---|
utm_source |
Where specifically did this come from? | instagram |
Yes |
utm_medium |
What kind of channel is that? | social |
Yes |
utm_campaign |
Which push does it belong to? | spring_sale_2026 |
Yes |
utm_term |
Which paid keyword? | running+shoes |
Paid search only |
utm_content |
Which creative variant? | hero_banner |
A/B testing only |
Three of the five do nearly all the work. The other two exist for paid search and creative testing, and adding them without a reason simply makes your URLs longer.
The mistake that silently halves your numbers
Analytics is case-sensitive. Instagram and instagram are two different sources. So are Email and email, and Spring_Sale and spring_sale.
Nothing warns you. The campaign simply appears twice in your reports, each with half the traffic, and the totals you compare against last month are quietly wrong. Worse, this is nearly impossible to repair retrospectively — historical data cannot be re-tagged.
The prevention is a convention you decide once and never deviate from: lowercase, underscores instead of spaces, no abbreviations you will not recognise in six months. Write it down. Our UTM campaign builder lowercases values by default for precisely this reason.
Two rules that are not obvious
Never tag internal links. A UTM parameter on a link between two pages of your own site restarts the session attribution, overwriting the original source. A visitor who arrived from a search engine and then clicked a tagged internal banner is reattributed to that banner, and the search engine disappears from your report. Reserve UTM tags for traffic arriving from outside.
Tag the destination, then shorten. Parameters live on the destination URL. The redirect passes them through intact. Building the tagged URL first and shortening it afterwards works; shortening first and trying to append parameters to the short link generally does not, because the parameters are dropped at the redirect. If the values contain spaces or ampersands, encode them with our URL encoder before assembling the link.
Which numbers deserve action
Most link dashboards present a dozen metrics. Perhaps four of them ever change a decision.
Read them together, because individually each is misleading.
A page with high views and a low click ratio has a mismatch: whatever brought people there is not what the page offers. The fix is on the page — the labels, the order, the first link.
A page with a healthy ratio but few views has a distribution problem, not a page problem. Editing the page is wasted effort; the fix is upstream, wherever the link is published.
One link taking eighty percent of clicks is not a success story about that link. It is a message that the other links are decoration. Remove them or replace them; every unclicked link taxes the ones that work.
And a daily number that moves is usually noise. Weekly and monthly direction is signal. Reacting to a single day's dip is how people make their pages worse.
What analytics cannot tell you
Worth stating plainly, since dashboards imply otherwise.
Why. Click data records what happened, never the reason. It cannot distinguish a link that was compelling from a link that was merely first, or a campaign that resonated from one that simply ran during a holiday. Causation has to come from somewhere else — a change you made deliberately, with the rest held still.
Who. Aggregate geography and device type are not identity, and any tool claiming to tell you which individual clicked is either wrong or is doing something you should not be comfortable with.
Whether it was worth it. Clicks are not revenue, and a link that drives many clicks and no outcomes is worse than one that drives few of both, because it costs attention to sustain. Attribution ends at the click; the rest of the answer lives in whatever the click led to.
A workable routine
Tag every external link with three parameters and a convention you never break. Shorten afterwards. Look at per-link clicks and the click ratio weekly, not daily. Change one thing at a time, and give it long enough that the change is distinguishable from the week it happened to fall in.
That is close to the whole discipline. The tooling — including ours — mostly exists to make those four steps take minutes instead of hours. It does not make the interpretation easier, and the interpretation is where the value is.
Related reading: how redirects and short links work, and why the domain your links live on matters.
Frequently asked questions
What does a "click" actually measure?
It measures a request to the redirect, not a human deciding to visit. Bots, link previews in messaging apps, and security scanners all generate requests. A reasonable analytics system filters obvious automated traffic, but no system can be certain, so treat click counts as a strong signal rather than an exact headcount.
What is the difference between clicks and unique clicks?
Clicks count every request. Unique clicks attempt to count each visitor once, usually within a time window. Unique clicks are the more honest number for reach; total clicks are more useful for spotting a link that people return to repeatedly.
Which UTM parameters do I actually need?
Three: utm_source, utm_medium and utm_campaign. Source names where the traffic came from, medium names the type of channel, campaign ties it to a specific push. utm_term and utm_content exist for paid keywords and creative variants, and most people never need them.
Why is my traffic split between two sources with the same name?
Almost always capitalisation. Analytics treats Facebook and facebook as different sources, so a single campaign quietly reports as two. Choosing lowercase and sticking to it prevents the problem, which is difficult to repair after the fact.
Do UTM parameters survive a short link?
Yes, when the parameters are on the destination URL. The redirect passes them through intact, and they arrive in your analytics as normal. Build the tagged URL first, then shorten it — appending parameters to a short link after the fact usually does not work.
Should I use UTM tags on internal links?
No. Tagging links between pages of your own site overwrites the original source of the session, so a visitor who arrived from a search engine is reattributed to your own newsletter. Reserve UTM tags for traffic arriving from outside your site.
Start with a free Linkly account
A branded bio page, short links and dynamic QR codes — with analytics on every click. No card required.
Get Started Free