How to Track Website Traffic: Your Own Site and Anyone Else's

"How to track website traffic" is two questions wearing one phrase, and mixing them up is why most guides on this leave you no further forward.

The first question is how to measure your own site, where you can install code and get exact numbers. The second is how to estimate someone else's, where you cannot install anything and every number you see is a model's guess. The tools are different, the accuracy is wildly different, and the mistakes are different. This covers both, in that order, plus how far off the estimates actually are.

Which job are you doing?

If you are doing both, do your own site first. You need a real dataset before competitor estimates mean anything, because the only way to sanity-check an estimator is to run it against a site whose true numbers you already know.

Part one: tracking your own website's traffic

Step 1. Install analytics, not three analytics

Google Analytics 4 is free, it is what most guides and most contractors assume you have, and it is the default for good reason. Create a property, copy the measurement ID, and put the tag on every page. On WordPress a plugin does it. On Shopify, Webflow, Squarespace and Wix there is a field in settings for exactly this.

The failure mode here is installing the tag twice, usually once through a plugin and once through a theme setting. This doubles your pageviews and it is the single most common reason a site's traffic "jumps" overnight. Check with a tag assistant extension before you trust the first week of data.

If GA4's interface defeats you, that is a reasonable reaction rather than a personal failing. The privacy-first alternatives below are far simpler and they answer 90 percent of the questions a small site actually has.

Step 2. Connect Google Search Console

Analytics tells you what visitors did once they arrived. Search Console tells you how they found you, and it is the only place you get impressions, which is the count of times you appeared in results whether or not anyone clicked.

That impressions number is the one most people never look at, and it is the most useful early diagnostic on the list. A page with 4,000 impressions and 3 clicks does not have a traffic problem. It has a title and snippet problem, and rewriting the page is the wrong fix.

Verify your site in Search Console, submit your sitemap, and wait. It backfills nothing, so the day you connect it is the day your history starts.

Step 3. Add behaviour, once the numbers exist

Numbers tell you that 70 percent of people leave a page. They do not tell you why. Microsoft Clarity is free with no volume cap, records real sessions, and produces heatmaps and rage-click reports.

Watch ten sessions on your most important page. It is more informative than a month of dashboards and it takes twenty minutes. Do not do this in week one, though. Wait until you have enough traffic that ten sessions is a sample rather than an accident.

The privacy-first alternatives

There is a real case for skipping GA4 entirely. Ad blockers refuse Google's tag far more often than they refuse these, so on a technical audience the smaller tools frequently report more traffic, not less, despite measuring the same visits.

Plausible. One page, no cookie banner needed in most jurisdictions, and a script small enough not to matter. Paid, cheap, and the easiest of these to hand to someone non-technical.

Fathom. The same idea with slightly more emphasis on compliance paperwork. Choose between the two on interface preference, because the underlying job is identical.

Matomo. Free if you self-host, and the only option here that gives you full GA-style depth while keeping the raw data on your own server. The self-hosted version is real infrastructure with real maintenance, so be honest about whether you want that.

Umami. Open source and free to self-host, the lightest option on this list, best for developers who would rather run a container than pay a subscription.

Statcounter. Still running after two decades, with a free tier and a live visitor log that some people find more useful than any chart.

Step 4. Decide what you are actually going to look at

Most dashboards are read once and then ignored, because nobody decided what a bad number would mean. Five metrics carry almost all the signal for a normal site.

• Sessions by channel. Organic, direct, referral, social, paid. The trend per channel matters, the total does not.
• Impressions and average position from Search Console. This moves weeks before traffic does, which makes it the early warning.
• Landing pages, not pageviews. Where people entered tells you what is working. Total pageviews tells you almost nothing.
• Engaged sessions or time on page. A crude quality proxy, and enough to spot a page that attracts the wrong visitor.
• Conversions. Even a newsletter signup. Traffic without a defined outcome cannot be judged good or bad.

Set a monthly review, not a daily one. Daily traffic numbers are noise on any site under a few thousand visits a month, and reacting to noise is how people talk themselves into rewriting pages that were fine.

Step 5. Put it in one place

Looker Studio is free, connects natively to both GA4 and Search Console, and turns the monthly review into one page you actually open. Build it once with the five metrics above and stop logging into two interfaces.

For the search side specifically, SEOTesting sits on top of Search Console and answers questions the native interface makes hard, like which pages lost clicks after a specific date. It is paid and it is not necessary, but it saves real time once search is your main channel.

The free stack that covers almost everyone

Four tools, no cost, about an hour of setup, and it answers every question on this page for a normal site. Once you can see the numbers, our guide to how to boost website traffic for free covers what to do about them at the same budget.

• Google Analytics 4 for sessions, channels and conversions.
• Google Search Console for impressions, positions and which queries find you.
• Bing Webmaster Tools for the same on Bing, which matters more than it used to because it feeds several AI assistants.
• Microsoft Clarity for recordings and heatmaps when a number needs explaining.

Add Cloudflare Web Analytics if you are already behind Cloudflare, because it measures server-side and therefore counts the visitors ad blockers hide from GA4. Add Looker Studio when logging into four interfaces becomes the reason you stop looking.

The only good reason to pay is wanting a simpler interface than GA4's, which is what Plausible and Fathom sell, or wanting to own your data, which is what Matomo and Umami sell.

Part two: checking any website's traffic

You cannot install anything on a site you do not own, so every number here comes from a model built on clickstream panels, browser extension data and search volume estimates. Understanding that is the whole skill.

Similarweb is the most complete estimator for total traffic across all channels, and its free tier gives you monthly visits, traffic sources and top countries for any domain. It is strongest on large sites and weakest on small ones.

Semrush and Ahrefs estimate organic search traffic specifically, by matching a site's ranking keywords to search volume. That makes them more accurate than Similarweb for SEO comparisons and useless for anything that arrives from email, social or direct.

SpyFu covers both organic and paid and is the cheapest way to see roughly what a competitor spends on ads. Historical depth is its selling point.

Cloudflare Radar publishes real, aggregate traffic data for domains behind Cloudflare, and it is not an estimate. Coverage is limited to what Cloudflare sees, which is a lot of the web.

How wrong are these numbers?

Nobody selling an estimator wants to answer that, so here is the honest version from checking estimates against sites whose real analytics we can see.

For a large, well-known site, expect the estimate to land within roughly 20 to 30 percent of reality. That is genuinely useful. For a site under about 10,000 monthly visits, the estimate is frequently out by a factor of two or more in either direction, and it is sometimes reported as zero for a site doing real numbers.

Two rules follow from that. Use estimates for comparison rather than for absolute figures, because the model's bias is consistent across sites even when the number is wrong. And never quote a competitor's traffic to a client as fact, because they will check it against a different tool and get a different answer.

Why your own tools disagree with each other

This confuses more people than anything else on this page, so it is worth stating clearly. Search Console and GA4 will never match, and neither of them is broken.

Search Console counts clicks from Google search, deduplicated per query per user. GA4 counts sessions initiated by a page load, including everyone arriving from anywhere else, and missing anyone whose browser blocked the tag. A 20 to 30 percent gap between GA4 organic sessions and Search Console clicks is normal. A gap of 60 percent usually means the tag is missing from a page template.

Tracking traffic from AI assistants

A share of your visitors now arrive after ChatGPT, Perplexity or Gemini named you, and a much larger share read an answer that cited you and never clicked at all. Both matter and they are measured differently.

For the clicks, filter your referrers for the assistant domains, which show up as chatgpt.com, perplexity.ai and similar. Any analytics tool on this page can do this and the number is usually small but growing fast.

For the citations without clicks, referrers cannot help you, because there was no visit. That has become its own measurement category, and the honest position is that if a rising share of your impressions produce no clicks, chasing rankings harder will not fix it. Being the source an assistant quotes is a different objective from being the first blue link.

What to do with the data

Traffic tracking earns nothing on its own. Three moves cover most of the value.

Find the pages that rank but do not get clicked. Sort Search Console by impressions, filter for a position between 5 and 20 and a click-through rate near zero, and rewrite those titles first. It is the cheapest traffic in existence because the ranking already exists. There is more on this in our guide to increasing website traffic organically.

Find the pages that lost traffic. Compare the last 90 days to the previous 90 in Search Console, page by page. Decay is quiet and continuous, and it is easier to recover a page that used to rank than to build a new one.

Find what your traffic is worth. Segment conversions by landing page and channel. Once you know what a visit from each channel is worth, the argument about where to spend the next hour resolves itself. Our breakdown of lead generation cost works through that calculation.

A note on where we sit, since this is our blog. Distribb is not a web analytics tool and will not tell you how many sessions your site had. It reads your Search Console performance to find the pages worth rewriting and then handles the writing, internal linking and publishing. If your problem is measurement, use the tools above. If your problem is that you can see what to fix and nothing gets fixed, that is the part Distribb handles.

Common mistakes

Installing the tag twice. Your traffic did not double.

Reading daily numbers. Below a few thousand visits a month, daily variation is weather.

Treating a third-party estimate as a measurement. It is a model, and on small sites it is a bad one.

Ignoring impressions. They move first and they tell you whether the problem is ranking or clicking.

Never defining a conversion. Without one, every traffic conversation is aesthetic.

Assuming blocked visitors do not exist. Between 10 and 30 percent of a technical audience blocks Google's tag. Your real numbers are higher than GA4 says.

Frequently asked questions

How do I track website traffic for free? Google Analytics 4 and Search Console together cover the whole job at no cost, and Microsoft Clarity adds session recordings for free on top. Most sites never need to pay for measurement at all.

How can I see how much traffic a website gets if I do not own it? Similarweb for total traffic, Semrush or Ahrefs for organic search traffic specifically. All three have free tiers that give you a usable ballpark. Treat the output as a comparison rather than a count.

Are third-party website traffic checkers accurate? Directionally, on large sites. On sites under roughly 10,000 monthly visits they are often out by a factor of two, and sometimes report zero for a site with real traffic. Different tools will give you different numbers for the same domain, and none of them is the truth.

Why do Google Analytics and Search Console show different numbers? They measure different things. Search Console counts clicks from Google search, GA4 counts sessions from every source and misses visitors who block the tag. A gap of 20 to 30 percent is expected.

How often should I check my traffic? Monthly for decisions, weekly at most for a glance. Daily checking on a small site produces reactions to random variation.

Can I track which visitors came from ChatGPT? Yes, by filtering referrers for the assistant domains in any analytics tool. What you cannot see is the much larger group who read an answer citing you and never visited.

Start with the pages that already rank

If you set up nothing else this week, connect Search Console and sort your pages by impressions. The pages sitting at high impressions and near-zero clicks are the fastest traffic available to you, and you can find them in about ten minutes. Our guides to getting blog traffic and increasing website traffic cover what to do once you have that list.