How can you tell how popular a website is?

Before agreeing a partnership with a publication, paying to be listed in a directory or deciding who your real competitors are, the same question comes up: is this website big or small? Do lots of people visit it, or does it only look that way?

A site’s real traffic is known only to its owner, through their analytics. From outside, all you have are estimates, and one of the most useful and least understood is the domain popularity list: a ranking of the most visited sites in the world, ordered from first to one millionth.

What happened to Alexa

For many years the benchmark was Amazon’s Alexa Rank. Media, advertisers and researchers quoted an Alexa position as if it were an official audience measurement. Amazon shut the service down in 2022.

Even before it closed, it had a fundamental problem: much of its data came from a browser toolbar installed by a very particular group of users, so the ranking reflected those users more than the internet as a whole. And it fluctuated so much from day to day that two studies asking the same question could reach different conclusions purely because of the date they downloaded the list.

What Tranco is and why it exists

Tranco was created in 2019 in academic research, in response to that problem. A group of security researchers from several European universities, led by KU Leuven in Belgium, showed that the popularity lists of the time could be manipulated fairly easily and that their instability undermined the studies relying on them. Their answer was a list that:

• Combines several sources instead of depending on just one. Over time these have included rankings from Cisco Umbrella (based on DNS queries), Majestic (based on links), Cloudflare Radar and Google’s Chrome UX Report.
• Averages the last thirty days, so a one-off spike cannot shoot a site up a hundred thousand places overnight.
• Is open and reproducible: each daily list has an identifier, can be downloaded and cited, and anyone can check how it was calculated.

Today it is the reference list in research and the most widely used alternative to the old Alexa rank.

What it measures, and what it does not

This is the crucial point: a rank is not traffic. It is a position relative to other domains, built from signals that go hand in hand with popularity without actually counting it:

• DNS queries reflect how often a domain is resolved, and that includes machines: API, advertising or software update domains can appear very high up even though no person ever visits them.
• Links measure how many sites point to a domain, which has more to do with authority than with visits.
• Browser data reflects the people who use that browser and share that data.

Three practical consequences follow:

1. The scale is not linear. Between rank 10 and rank 1,000 the differences in traffic are enormous; between 400,000 and 500,000, sites can be almost indistinguishable and move thousands of places without anything real changing.
2. It is global popularity. A site that is very strong in a single country or a professional niche can sit much lower than its weight in its own market would suggest.
3. Being outside the top million is normal. There are hundreds of millions of registered domains and the list covers a tiny fraction. A law firm, a clinic or a specialist shop can do very well without ever appearing.

To estimate a competitor’s visits there are commercial tools that work from user panels, with margins of error worth keeping in mind. For your own site, the reliable source is your analytics or Search Console.

So what is it good for?

Used well, a list such as Tranco answers order-of-magnitude questions, which are often the ones that matter.

Comparing competitors. Not the exact number, but the band and the trend. A competitor at 300,000 and you at 400,000 says little; them climbing for months while you slide does say something, and it is worth looking at what has changed on their site. Finding out what a website is built with helps here, as does reviewing the history of its pages, as explained in how to recover a site’s old URLs.

Judging a link or a partnership. If a publication offers you a sponsored article or a directory charges you to be listed, knowing whether it is in the top 100,000 or outside the top million puts the price in context. It does not replace checking whether the site covers your subject, has real readers and links with judgement, which is what really counts in a link, as covered in backlinks, authority and internal linking. But it quickly rules out inflated offers.

Spotting something odd. A site that presents itself as the leader in its sector but appears on no list, or a domain that suddenly jumps hundreds of thousands of places, are signals worth a closer look.

Tracking your own progress, if you are on the list. Always with a single source over time, not by comparing sources: each list weights things its own way and their numbers do not match.

Common mistakes when reading a ranking

• Treating the rank as audience. “We are at rank 200,000” says nothing about how many customers visit you.
• Mixing different lists. A commercial tool’s rank and Tranco’s are not equivalent.
• Reacting to one day’s movement. In the lower part of the list, fluctuations are mostly noise.
• Turning it into a target. Climbing a global popularity list is not a business goal. Selling more, getting more enquiries or ranking your key pages better is.

Checking any website’s rank

The domain rank checker queries the Tranco list and shows any domain’s current position, its popularity band (top thousand, ten thousand, hundred thousand or million), its change over the last thirty days and a chart of the last ninety. It takes up to three domains at once, which is exactly what you need to put yourself alongside two competitors and compare trends rather than isolated figures.

If yours comes out beyond the top million, that is not bad news: it is what happens to most businesses that are doing perfectly well. In that case, the useful comparison is between your competitors themselves.

---

A popularity list answers “is it big or small?”, not “how many visits does it get?”. Used for the first, it saves time and the occasional invoice; used for the second, it leads to the wrong conclusions.