Accuracy of Similarweb.com (Your Views)

[Malikbhai]

It is sort fairly accurate with a single digit up and down for popular sites, but not at all good for new or websites that have recently lost their google rankings.

Normally, I don't give two hoots about third party metrics, but they seem to be quite accurate on few of our high traffic properties.

That's my take. What do the members here think of their data?

[BBCasinos]

In my opinion they are horrible
their stats in noway match ours

[MichaelCorfman]

Well, Similarweb provides data based on a small sample of overall web traffic (millions of users, but still a very small sample of overall web traffic). I believe it provides information that is very helpful, but especially on sites without a lot of traffic, it is impossible to be statistically significant. Below is some sample data for a site with just north of a thousand visitors a day for the period January to March 2019:

Total Visits: Google Analytics reports 132,076 versus Similarweb's estimate of 235,161

United States: Google Analytics reports 19.86% (position 1) versus Similarweb 10.42% (position 2)
United Kingdom: Google Analytics reports 16.94% (position 2) versus Similarweb 15.38% (position 1)
Canada: Google Analytics reports 3.85% (positiion 3) versus Similarweb 2.96% (position 10)
Indonesia: Google Analytics reports 3.68% (position 4) versus Similarweb 1.45% (position 19)
Germany: Google Analytics reports 3.39% (position 5) versus Similarweb 3.87% (position 5)
India: Google Analytics reports 3.02% (position 6) versus Similarweb 1.01% (position 26)
Malaysia: Google Analytics reports 2.82% (position 7) versus Similarweb 0.61% (position 37)
Australia: Google Analytics reports 2.81% (position 8) versus Similarweb 0.70% (position 33)
Netherlands: Google Analytics reports 1.80% (position 9) versus Similarweb 1.75% (position 14)
Sweden: Google Analytics reports 1.65% (position 10) versus Similarweb 1.90% (position 12)
Malta: Google Analytics reports 1.51% (position 11) versus Similarweb 3.50% (position 7)
Philippines: Google Analytics reports 1.50% (position 12) versus Similarweb 1.09% (position 22)
Thailand: Google Analytics reports 1.43% (position 13) versus Similarweb 0.77% (position 32)
New Zealand: Google Analytics reports 1.43% (position 14) versus Similarweb 0.45% (position 46)
Poland: Google Analytics reports 1.28% (position 15) versus Similarweb 3.45% (position 8)
Ukraine: Google Analytics reports 1.19% (position 16) versus Similarweb 1.74% (position 15)
Romania: Google Analytics reports 1.13% (position 17) versus Similarweb 2.30% (position 11)
Finland: Google Analytics reports 1.07% (position 18) versus Similarweb 1.09% (position 22)
Vietnam: Google Analytics reports 1.03% (position 19) versus Similarweb 0.79% (position 30)
Uganda: Google Analytics reports 1.00% (position 20) versus Similarweb not reported.

It is certainly possible to look at this data and say it is off by a lot, and correct to do that if you are looking for exact figures. However, a traffic estimate accurate within about 40% and identifying the top two countries correctly (but with the order reversed), does provide a decent sense of where a site sits in terms of relative traffic versus other sites as long as you regard it as an estimate with that level of statistical accuracy.

Michael

[euglag]

SW statistics doesn't shows even approximate data. In my case, SW shows x10 more traffic than the website actually have.

[tpivesgarcia]

In my opinion they are horrible
their stats in noway match ours

Yeah, I agree with this, totally not accurate to our data.

[drifter8]

They definitely do not show real data!

[PROFRBcom]

Much like Alexa, SimilarWeb incorrectly estimates our traffic, most popular keywords, percentage of country visitors, time on site, pageviews, and much more.

With that said, I still use these tools as relative measuring sticks that I realize will have accuracy issues that are exacerbated the lower the actual traffic the site I am researching is.

I find them useful, but do not take anything they say as ultimate truth!

[TheGooner]

I just checked out https://www.goonersguide.com on Similar Web - and it was better than random guessing numbers - but not by a whole lot.

For example :
Three of the top five visitor countries were correct, in being US, UK, and Kenya (like a plague of locusts), but two of those are easy to guess given our sports content and language of English ... but Sweden and Slovenia were the other two ? They are countries ranked #25 and #61 according to Google.

The traffic volume and monthly variance is way off (sometimes half, sometimes double), and the visit duration (30% out), pages read (40% out) and bounce rate (20% out) seem sort of related but not very accurate.

It also thinks 99.7% of my social traffic is from youtube - if it is then then I have no idea how it could be I've never linked from youtube - and google analytics doesn't show it at all.

Overall, it sort of gives a feel for the site, I recognise the basic shape of my traffic that it outlines, hundreds of thousands of english speaking sports fans, but ANY STAT that is specific that looks interesting at SimilarWeb just seems plain incorrect or a guess.

I'd give SimilarWeb statistical analysis about 3/10 ... it can tell I get lots of traffic ut the details are all screwed by their user base bias.

[universal4]

Much like Alexa, SimilarWeb incorrectly estimates our traffic, most popular keywords, percentage of country visitors, time on site, pageviews, and much more.

With that said, I still use these tools as relative measuring sticks that I realize will have accuracy issues that are exacerbated the lower the actual traffic the site I am researching is.

I find them useful, but do not take anything they say as ultimate truth!

This is very well said I think, and the disclaimer is spot on. So if you use it regularly, you can study the variances to judge a portion of the results of changes you may perform on your site.

Since so many have wide swings in what they report and what the reality is, I wonder if they use cookie tracking of those that visit and or use their services, and if so a higher concentration of webmasters could be one of the factors in their results.

This may or may not be the case, but many of us know that the toolbar is/was a big factor in Alexa results.

Rick
Universal4

[Rex]

I don’t use many tools like this apart from Ahrefs (which I find good) I took a look at similar web for a couple of my sites and agree with the majority here; it’s not very accurate.. maybe it’s better with a paid version but I wouldn’t bother with that…