Alexa site rank: what the number measured, why it ended, and what to use instead

Alexa Site Rank was Amazon's measure of how popular a website was compared with every other website, expressed as a single whole number where 1 was the most visited site on the web. Amazon retired the service on 1 May 2022, so there is no current Alexa rank for any domain and no way to look one up. If you have landed here because a proposal, an old media kit or a directory listing quotes an Alexa rank, treat it as a historical figure from 2022 or earlier. For a current reading of a site's traffic and relative popularity, the practical replacements are Similarweb, Semrush Traffic Analytics, Cloudflare Radar, the Tranco research list and the Chrome User Experience Report top origins list, and each one measures something slightly different.

This page covers what the Alexa rank actually measured, how to read an old one, and which tool now answers each of the questions people used it for. For a side by side comparison of the current ranking tools themselves, see how the web rank tools build their numbers.

What the number actually was

Alexa Internet was founded in 1996 and bought by Amazon in 1999. Its ranking combined two things over a rolling three month window: the estimated number of unique visitors to a site and the estimated number of page views, both drawn from a panel of users who had installed the Alexa toolbar or other extensions that reported browsing data, along with data from sites that had Alexa's certified metrics installed.

Three properties follow from that design, and they explain most of the confusion about the metric:

• Lower was better. A rank of 5,000 meant the site was estimated to be the five thousandth most visited in the world. Companies quoting "an Alexa rank of 1.2 million" were quoting a weak position, not a strong one.
• It was panel based, so it was skewed. People who install browser toolbars are not a representative sample of the internet. Sites with a technical, webmaster or marketing audience ranked better than their real traffic deserved, and the skew was largest for smaller sites.
• It was global by default. A site with heavy traffic in one country could sit at a modest global rank while dominating its own market, which is why the country specific rank mattered more than the headline number.

Why it was retired, and what went with it

Amazon announced the retirement in December 2021 and closed the service on 1 May 2022. The whole product went: the free rank lookup, the site overview pages, the paid subscriptions, the competitive analysis tools, the browser extension and the Alexa Top Sites list that was available through Amazon Web Services. Third party sites that still display "Alexa rank" are either showing a cached value from before the shutdown or computing their own number and borrowing the name. Neither is a current reading.

Note also that Alexa Internet and the Alexa voice assistant were separate products that happened to share a name. The ranking service closing has nothing to do with the smart speaker.

How to read an Alexa rank you find in an old document

• Check the date of the document. Anything after April 2022 is quoting a stale number, and that is worth knowing about whoever wrote it.
• Read the band, not the digits. Inside the top 100,000 meant a genuinely large site. Between 100,000 and one million meant a real but modest audience. Beyond a few million the figure carried almost no information, because the panel saw too little traffic to separate those sites.
• Prefer the country rank if it is quoted. It was the more meaningful of the two for any business serving one market.
• Do not compare it with a current number from another tool. Different panels, different methods, different scales. Rebuild the comparison with one current source applied to every site in the set.

What to use now, by the question you are asking

What you want to know	Use	What it gives you
Roughly how much traffic does this site get?	Similarweb, Semrush Traffic Analytics, Ahrefs	Modelled visits, traffic sources, top pages and countries
How popular is this domain relative to others?	Tranco, Cloudflare Radar domain rankings	A ranked list built for research, with a documented method
Which sites are the most visited in a country or category?	Similarweb category rankings, Cloudflare Radar	Top site lists by market and sector
Is this domain big enough to be worth a link or a partnership?	A backlink tool's authority metric plus a traffic estimate	Link profile strength alongside estimated audience
Which origins do real Chrome users visit most?	The Chrome User Experience Report top origins dataset	A public, documented popularity grouping based on real Chrome usage
How much traffic does my own site get?	Your analytics plus Google Search Console	Measured figures rather than estimates

Tranco deserves a note, because it was built specifically to fix the problems Alexa had. It combines several source lists and averages over a period, which makes it far harder to manipulate and gives a reproducible list that researchers can cite. It is the closest thing to a successor for the "how popular is this domain" question, and it is free.

The habit worth keeping, and the habit worth dropping

The habit worth keeping is comparative measurement. Alexa was useful because it put your site and five competitors on one scale, and that comparison is still the right way to use any traffic estimate. Pick one tool, measure the same set of domains on the same day each month, and read the direction of travel.

The habit worth dropping is quoting a single popularity number as a credential. A modelled estimate of a competitor's traffic is a planning input. Your own measured numbers from analytics and Search Console are the ones to report, and our guide to reading a traffic checker's two numbers covers which of a tool's figures to use for which purpose. If the underlying question is really about site authority rather than traffic, what domain authority is and is not is the better starting point.

Questions people still ask

Can I check my Alexa rank anywhere?

The service closed on 1 May 2022, so there is no live figure to check. Sites that display one are serving a cached pre shutdown value or their own metric under a familiar name. Use Similarweb, Semrush, Ahrefs, Tranco or Cloudflare Radar for a current reading.

Was a lower Alexa rank good or bad?

Lower was better. It was a position in a league table of all websites, so rank 1,000 was a far larger site than rank 1,000,000.

Did Alexa rank affect Google rankings?

It was a third party popularity estimate with no role in how Google ranks pages. Google ranks on its own signals, and a popularity score from an external panel was never one of them.

What replaced the Alexa Top Sites list?

For research use, Tranco is the standard answer, because its method is published and it averages across sources and time. Cloudflare Radar publishes domain rankings from its own resolver data, and Similarweb publishes top site lists by country and category.

My industry still asks for an Alexa rank. What do I send?

Send measured numbers from your own analytics for your site, a current Similarweb or Semrush estimate for the competitive comparison, and one line explaining that Alexa closed in May 2022. That answer is more useful than the number they asked for, and it tends to change the conversation for the better.

Is there a free way to compare my site with competitors?

Yes. Similarweb and Semrush both give limited free lookups, Tranco is fully free, and Cloudflare Radar is free to browse. Combine one traffic estimate with one ranked list and you have a defensible comparison at no cost.

A rank that never existed for assistants

Alexa gave every site one number, and people argued over it for years. Assistants give no number at all. They give a sentence, and a site is either in it or not. The old habit of asking where a site ranks becomes a better question: is it named, for which questions, and from which source? Try it by hand for five minutes. For a longer list, AI Knows Us records the names across ChatGPT, Gemini, Perplexity and Claude.