On this page
Three signals, read together: the page they landed on, the referring domain, and the timing against when you published. None is proof on its own, and together they are usually enough to decide what to do next week. Which is the actual job — not attribution, decision-making.
This matters more than usual if your analytics sits on a free tier, because the standard advice — tag every link with campaign parameters — quietly assumes a feature you may not have.
Does link tagging actually work on a free plan?
Often not, and this is worth checking before you tag anything.
On Vercel Web Analytics, UTM parameter breakdown is part of the Web Analytics Plus add-on at $10 a month per team, on top of Pro. It is not on Hobby and not on plain Pro. The tags still travel in the URL; you just do not get the report that makes them useful.
So before building a tagging scheme, open your analytics and confirm there is somewhere for the tags to appear. Plausible, Umami or Vercel Analytics has what each tier includes.
What works when tagging does not?
A different landing page per channel.
If the link in your TikTok bio goes to one URL and the one in your YouTube descriptions goes to another, then the landing page is the attribution. It shows up in the most basic report any tool offers, it survives a change of analytics provider, and no app can strip it.
| Channel | Lands on |
|---|---|
| Short-form bio link | the free thing |
| Long-form descriptions | the article the video is about |
| Newsletter | the product page |
Three URLs, zero configuration, and it keeps working on a free tier. The cost is that you cannot separate two posts on the same channel — which, at the volume most people are at, is a distinction they would not have acted on anyway.
Why does so much traffic show as "direct"?
Because phones lie, and not on purpose.
Social apps open links in their own in-app browser, and those frequently omit or strip the referring page. The result is that traffic from an app arrives looking like someone typed your address from memory.
So "direct" is the least informative row in any report, and a jump in direct traffic an hour after you posted is not a mystery — it is the post. This is also why the timing signal matters as much as the referrer: it is often the only one you get.
The whole list
67 tools, what each costs and when you actually need it. Updated twice a year.
How do you read the three signals together?
| What you see | What it probably means |
|---|---|
| Spike on one page, plausible referrer, within an hour | that post worked |
| Spike in direct, right after posting | that post worked, through an app |
| Steady trickle to one article, no referrer pattern | search is starting to find it |
| Views high, site flat | the hook landed, the next step did not |
| Site traffic, no subscribers | people arrived at the wrong page for the offer |
The fourth row is the common one and the most useful. A video doing well while the site stays flat is not a distribution failure — distribution worked. It means the thing you said and the thing you linked to were about different subjects, or the reason to click was never made.
What should you not conclude?
That one week means anything about an article.
Social content decays in days; search traffic arrives in months. An article judged at seven days looks like a failure even when it is going to be your best page by March. The two need different review intervals, and mixing them is how people delete the thing that was about to work.
| Format | Judge after |
|---|---|
| Short-form post | 3 days |
| Long-form video | 2 weeks |
| Article aimed at search | 3 months |
What is the number that actually settles it?
Subscribers, if you have a free thing worth taking.
Views are a platform's business. Clicks are yours but anonymous. An email address is the first point at which someone has said, in their own words, that the next thing is worth receiving — and it is the only one of the three you keep if the platform changes its mind about you.
So the honest version of "did it do anything" is usually: did anyone join the list this week, and can I tell which post they came from? If the second half is unanswerable, give each channel its own landing page and it stops being unanswerable next month.
What about asking?
It is undervalued, it is free, and it takes one line.
The reply-to on a delivery email, answered by a person, produces better information in ten exchanges than a dashboard does in a quarter. "What were you trying to do when you found this?" is the whole research programme at this stage.
Analytics tells you what happened. It is remarkably bad at telling you why, and at small numbers the why is the only thing that changes what you build. The five numbers worth checking every week covers keeping the what small enough that you still have attention left for the why.
Where these numbers came from
Vercel's Web Analytics limits and pricing documentation for the UTM and add-on details, checked 12 October 2026. The rest is method rather than figures, and the method is deliberately cheap: it is designed to work on a free tier and on small numbers, which is where it is needed.
The Affiliate Playbook is the one that covers turning attention into something that is yours rather than a platform's.
Tools in this piece
| Tool | What it does | Price |
|---|---|---|
| Plausible | Simple, privacy-friendly visitor stats with no cookie banner. | ~€9/mo |
| Umami | Same idea, free if you host it yourself. | Free self-hosted |
| Vercel | Hosting. Puts a project live on a real address with one command. | Free tier |
| Kit | Email platform for creators. The list is the only thing you actually own. | Free under 1,000 subs |
Every tool has its own page with the price, who should skip it and what to check before paying. Some links are affiliate links.
Questions people ask
How do you measure whether a social post sent anyone to your site?
Three signals together: the landing page they arrived on, the referring domain, and the timing against when you published. No one of the three is proof, but a spike on a specific page from a plausible referrer within an hour of posting is usually enough to act on.
Do UTM tags work in Vercel Web Analytics?
Only with the Web Analytics Plus add-on, which is $10 a month per team on top of Pro. On Hobby and on plain Pro the breakdown is not available, so tagging links there does not give you the report you expected.
What is the best free way to attribute traffic?
A different landing page per channel. If your TikTok bio points at one URL and your YouTube description at another, the landing page itself is the attribution and no tooling is required.
Why does social traffic often show as direct?
Because apps open links in their own in-app browser and frequently strip or omit the referrer. Traffic from a phone app is routinely reported as direct, which makes direct the least trustworthy row in any report.
What does it mean when a video does well but nothing happens on the site?
Usually that the content and the destination are about different things. Views measure whether the hook worked; clicks measure whether the next step was worth taking. A big gap between them is a message problem, not a traffic problem.
How long should you wait before judging a piece?
For social, a few days — the decay is fast. For an article aimed at search, months, because the traffic arrives long after you published and will not show up in a weekly check for a while.