What is website traffic?

Website traffic is the number of people who come to your website in a given period, and what they did while they were there. When someone says "we got 5,000 visits last month", they mean traffic. When they say "most of our traffic comes from Google", they mean where those visits started.

So the word covers a count and a description at the same time. The count tells you how many. The description tells you who, from where, on what device and how often. Both come from the same place: a small piece of code from an analytics tool that runs on every page of your site and reports each visit back.

Traffic is not the same thing as customers. A thousand visits can produce zero sales. But nobody can buy, sign up or read what they never reach, so traffic is the first number most site owners learn to check.

Ask three people how much traffic a site gets and you can get three honest, different answers, because analytics tools report traffic as three separate counts.

A visitor is a person, roughly. The tool gives each browser an anonymous ID and counts the distinct IDs. Google Analytics calls this "users".

A visit, also called a session, is one sitting. It starts when someone lands on the site and ends when they leave or go quiet. Google Analytics closes a session after 30 minutes without activity (Google, 2026). Come back the next morning and that is a second session from the same visitor.

A pageview is one page load. Someone who reads your homepage, then your pricing page, then the homepage again has produced three pageviews in one session. Google Analytics calls these "views" and counts the repeat view of the homepage separately (Google, 2026).

Put together: one person who reads four pages before lunch and two more after dinner counts as 1 visitor, 2 sessions and 6 pageviews. All three numbers are "traffic". Pageviews are always the largest and visitors the smallest, so before you compare your traffic with anyone else's, check that you are comparing the same one. Which number to quote, and why no two tools agree on any of them, is the subject of page views vs. visits vs. unique visitors.

Pageviews are also the number that shows up on invoices, because most analytics tools price by them. Bigdelta's free plan, for example, covers sites under 100,000 pageviews a month, which is more than a small business site usually produces.

The most useful way to split traffic is by where it came from. Every analytics tool sorts each visit into a channel by looking at the page the click came from and any tags on the link. Google's default list runs to about twenty channels (Google, 2026), but six cover nearly everything a small site sees:

There is a newer bucket too. Google's list now includes an "AI Assistant" channel for visits that start in tools like ChatGPT or Gemini (Google, 2026). It is small so far. In Semrush's 2026 study of 50,000 websites, AI assistants sent well under one percent of 2025 visits, though the share grew fast during the year.

Two of these buckets deserve a warning. Organic is the one most people want more of, and what is organic traffic explains how a visit earns that label and why some organic visits get filed elsewhere. Direct is the one most people misread, because it is a catch-all for visits with no clue about their origin. What is direct traffic walks through what really lands there. For the full tour of the buckets, and the rough share each one should hold, read traffic sources explained.

Source is not the only split. Tools also break traffic down by device, by country, by landing page, and by new versus returning visitors. Device matters more than beginners expect. Worldwide, phones and desktops each carried about half of web traffic in August 2026 (StatCounter, 2026), but the split for any one site depends on what kind of site it is. A recipe blog can be nearly all phones. A software company's site can be mostly desktops.

• Organic search: unpaid clicks from Google, Bing or another search engine.
• Paid: clicks on ads you bought, whether on a search engine or a social network.
• Direct: the tool could not tell. In theory a typed address or a bookmark. In practice also every link opened from a messaging app, a PDF or a QR code.
• Referral: a link on another website, such as a blog post or a directory.
• Social: an unpaid link from a social network.
• Email: a link in a newsletter, provided you tagged the link so the tool can recognise it.

The counting works like this. You paste a short snippet of code into your site. Each time a page loads, the snippet sends a note to the analytics tool: which page, at what time, from where, on what device. The dashboard is those notes added up. Setup takes a few minutes, and how to check website traffic walks through it for your own site and for sites you do not control.

The dashboard you get back uses the vocabulary from this post. A tool like Bigdelta, for example, counts unique visitors, sessions and pageviews and shows the visits split by search, social, referral, email and direct.

Two things make the count lower than reality. Ad blockers and some privacy-focused browsers stop the snippet from running, so those visitors are never counted. Links shared in private channels arrive with no trace of where they came from, so they are counted but filed as direct. Dark traffic covers how much goes missing this way and why you cannot fully fix it.

One thing makes the count higher than reality. Bots. Just over half of all web traffic in 2025 came from automated programs rather than people, according to the Thales 2026 Bad Bot Report. Most of that never reaches your dashboard, because search crawlers do not run analytics code and Google Analytics quietly drops traffic from known bots without letting you see how much it dropped (Google, 2026). Some bots do get through, though. When your traffic doubles overnight with no launch, no press and no new ad, a bot is the first suspect, and the giveaway is usually one country, one page and visits that last a second.

So your traffic number is close, not exact. Trust the trend over weeks more than any single day.

For a shop, traffic is people walking past the shelves, and the raw count matters less than which sources send people who buy.

Take a store selling ceramics that gets 20,000 visits in a month. Most of those come from Instagram, a smaller number from Google searches like "handmade dinner plates", and a few thousand are direct. The Instagram visitors mostly look at one product and leave. The Google visitors arrive on a product page already wanting a plate and buy at several times the rate. On a plain traffic chart Instagram looks like the star. Next to revenue, Google is.

Stores also live with seasonality, so compare a month with the same month last year rather than with the month before. And shops lean towards phones more than most sites, so a checkout that is slow on a phone is slow for most of your buyers.

What to look at first: traffic by source with orders or revenue next to it, so you see which sources send buyers rather than which send the most people.

A software company selling to other businesses usually has small traffic, long visits and a desktop-heavy audience. A project-management tool with 3,000 visits a month can be perfectly healthy if 40 of those turn into trials.

The traffic here splits into people who are evaluating and people who found a blog post while looking something up. Evaluators read the pricing page and the docs. Readers land on one article and leave. Mixing them in one total hides whether the pipeline is growing.

Two more things muddy a SaaS traffic count. Existing customers heading to the login page get counted as marketing-site visits unless you separate them. And searches for your own brand name look like organic growth while really being people who already knew you.

What to look at first: visits to the pricing or signup page by source, rather than total visits. That is the traffic that can turn into revenue.

On a content site the traffic is the product. Advertising, affiliate links and sponsorships all pay by the pageview, so here the biggest of the three numbers really is the one to watch.

Content sites usually get most of their traffic from organic search, which means they are exposed to anything that changes how search works. Picture a personal-finance blog with 80,000 pageviews a month where a single post about budgeting apps brings in a third of them. If Google reshuffles its results, or starts answering the question itself at the top of the page, that post loses clicks and the whole site's income moves with it. The stable part of a content site's traffic is the part that does not depend on a search engine: newsletter subscribers and people who come back on their own.

What to look at first: pageviews by landing page, to see how much of the total rides on a handful of posts, and the share of visits from returning visitors, which tells you how much of the audience is really yours.

A plumber, a dentist, a florist or a small agency has the smallest traffic of all and the highest value per visit. A plumber's site with 400 visits a month that produced 20 phone calls did its job. A design agency's site with 600 visits, mostly from LinkedIn and word of mouth, that produced two serious enquiries did its job too.

Much of a local business's traffic starts from the map listing rather than from the regular search results. Clicks on the website link in a Google Business Profile usually land in the organic bucket, so a local site's "organic" number often includes the map. Geography is the other split that matters: visitors from the town you serve are prospects, and visitors from three time zones away mostly are not.

One trap is specific to small traffic. At 400 visits a month, your own visits are a visible share of the total. You, your staff and your web designer checking the site can be a tenth of the traffic. Most analytics tools let you exclude your own team. At this scale, do it on day one.

What to look at first: visits to the contact page, and calls or form submissions if you can track them, split by which town the visitors are in.

There is no universal good number, because the four sites above can all be succeeding at 400, 3,000, 20,000 and 80,000 visits. The benchmark that fits is your own: decide what the site is for, work out how many visits it takes to produce one of those outcomes, and multiply. What's a good amount of traffic for a website walks through that arithmetic and gives the industry medians for the readers who want them anyway.

Once you have the counting straight, the next skill is making the numbers say something: which split to look at, which comparisons are fair, and what to subtract before you trust a total. That method is in how to analyze website traffic.