Semrush vs Similarweb: which is more accurate? A test on 787 websites
We compared Semrush and Similarweb traffic estimates with Google Analytics data for 787 websites. Both were off by 45-61%. Where each one is closer, and why.
Neither Semrush nor Similarweb measures a competitor’s traffic. Both estimate it, and the estimates are rough: in our test, each tool’s figures deviated from the real Google Analytics numbers by 45% to 61%.
We compared the two against Google Analytics data for 787 websites. The short answer:
| Semrush | Similarweb | |
|---|---|---|
| Sites with 1,000,000+ sessions a month | closer for 57% of sites, 45% deviation | less accurate |
| Sites with 100,000 to 500,000 sessions | closer for 53% of sites, but larger errors when wrong | lower deviation overall |
| Typical direction of the error | no clear direction | underestimates traffic |
So both are good enough to compare websites with each other and to see a market trend. Neither is good enough to quote as a competitor’s actual traffic.
One caveat on age: the measurements are from January 2020. Both tools have changed their data sources since, so read the figures as the size of error to expect from this kind of estimate, not as a current score for either product.
Where Semrush and Similarweb get their data
Where Semrush and Similarweb get their data
Both describe their own sources, and neither has access to the analytics of the sites it reports on, unless the owner shares it. As of October 2026:
- Similarweb lists four sources: first-party analytics shared by website and app owners, anonymous traffic data from Similarweb products installed on devices, public data, and data from partners such as ISPs and measurement companies.
- Semrush says its traffic figures are “based on petabytes of clickstream data”, approximated from the behavior of a panel of internet users. It also explains that Google Analytics collects data directly through a tracking code on the site, which is why the two differ.
The method is the same in both cases: observe a sample of users, then model the whole audience from it. The smaller the site, the smaller its share of the sample, and the rougher the estimate.
How we compared services and calculated accuracy
How we compared services and calculated accuracy
Both Semrush and Similarweb can be used for evaluating competitors, leads, and traffic generated by partners. In the test we used Semrush Traffic Analytics, as the product was called at the time. As we have access to the anonymized Google Analytics data of OWOX users, we assumed we had reasonably precise data on visitors for a number of anonymized websites. We then compared data provided by Semrush and Similarweb for those same anonymized websites against the Google Analytics data. Our task was to describe the inaccuracies in particular business spheres and see how large they are for each service.
The data sample
Our sample consisted of data collected from Google Analytics, Semrush, and Similarweb for all sessions in January 2020 on each of 787 websites. These websites are based in Australia, Canada, the US, the UK, and Germany and were grouped by business sphere:
- Computers, electronics, and technology
- Entertainment
- Finance
- Health and beauty
- Jobs and education
- News and media
- Professional services
- Retail
- Telecom
- Travel
Comparison approach
To find out the accuracy of competitor traffic calculations by Similarweb and Semrush, we merged the following data into one table:
- Anonymized Google Analytics data from 787 websites with more than 100,000 visitors per month according to OWOX
- Data on traffic from those same websites according to Semrush and Similarweb
We excluded websites with low traffic that might be caused by filters for properties.
Then we calculated the absolute value of the deviations between data provided by Semrush and Similarweb compared to Google Analytics.
As deviations might be positive or negative and we were interested only in accuracy, we used the absolute value. This also saved us from a potential mistake with calculations in which positive and negative values could result in zero.
Then we segmented our websites by number of sessions per month:
- 100,000 – 500,000
- 500,001 – 1,000,000
- More than 1,000,000
OWOX Data Marts
See your first report built in real time. 15 minutes.
- Connect your data warehouse
- Pick your metrics
- Get a live Google Sheets report
In the time it takes to write a ticket. Then imagine never writing that ticket again.
Book a DemoWe'll use your actual use case
Research results
The higher the standard deviation, the larger the inaccuracy in data from the given service as measured against the Google Analytics benchmark. The deviation for Similarweb was between 57% and 61% and did not correlate with website traffic volumes. Semrush data showed a clear tendency: websites with huge amounts of traffic (1,000,000 sessions or more) showed greater accuracy and a smaller deviation (45%) from Google Analytics data.
For websites with 500,000 sessions or more, the numbers were 9 to 12 percentage points more precise in Semrush. For projects with little traffic, Similarweb worked slightly better, but both services demonstrated large inaccuracies among this group of websites.
These differences in accuracy are caused by the collection and processing algorithms of Similarweb and Semrush as well as special aspects of clickstream data. Traffic analysis services use artificial intelligence and machine learning algorithms to approximate data on all website visitors based on clickstream data, which is data on samples of website traffic. Thus, the smaller the website, the less accurate the approximations based on clickstream data.
What should you do if your website and your competitors’ sites have low traffic and the accuracy of your data is extremely low? In this case, you should benchmark against bigger competitors in your market. If you compare some big market players, you’ll see not only their performance but also the general market trends. And by comparing performance and trends revealed by Similarweb or Semrush with your own achievements, you’ll be able to see the efficiency of your marketing.
The graph below describes the share of observations: the percentage of websites within each segment for which Similarweb and Semrush respectively were closer to the Google Analytics benchmark. For example, in the 1,000,000+ sessions segment, Semrush gave more accurate data than Similarweb for 57% of analyzed websites:
By comparing the 100,000 to 500,000 sessions segment in the Standard deviation and Share of Observations graphs, we discover an interesting insight: Semrush data has a higher standard deviation, which tells us that the session approximations are generally less accurate. At the same time, Semrush is still more precise in 53% of cases. Long story short, Semrush makes fewer mistakes but the mistakes it does make tend to be massive.
The accuracy of data depends on a variety of factors:
- How Google Analytics is set up, which pages have GA tags, and what exactly those tags measure.
- How active the website is on Google. For a redirect website like from an advertising network or a promo website with massive paid ad traffic, Semrush will show lower numbers.
- How big the share of organic traffic is for the website. Semrush is likely more precise for sites with a larger share of organic traffic.
- The region and business sphere. Let’s suppose Semrush has more projects from the US. In this case, it will have more data from the US and calculations for US sites will be more accurate.
The next two graphs demonstrate the standard deviation and share of more accurate data from both services segmented by business sphere.
As you can see, the deviation of both Semrush and Similarweb depends on the business sphere:
The percentage of websites for which Similarweb and Semrush were closer to the Google Analytics benchmark also depends on the business sphere:
For example, in the Computers sphere, Similarweb was more accurate for 58% of websites and Semrush was more accurate for 42% of websites (the first column in the graph above).
This dot graph shows the positive and negative deviation values for Semrush (blue dots) and Similarweb (green dots):
Even visually, you can conclude that the lower part of the graph has more green dots, which means that Similarweb is more likely to show values lower than the real traffic data from Google Analytics.
To sum up
We found that:
- The accuracy of Similarweb and Semrush is similar, and low: a deviation of 45% to 61% from Google Analytics.
- In the 1,000,000+ sessions segment, Semrush was more accurate than Similarweb.
- For smaller sites Similarweb was slightly closer, and both were far off.
- Similarweb is more likely to underestimate traffic volumes.
Estimates for competitors, measured data for yourself
Estimates for competitors, measured data for yourself
The practical rule from this test: use these tools for comparison, never for absolute numbers. If Semrush shows a competitor growing 20% while your own estimated traffic in the same tool is flat, that difference means something, even though both absolute figures are wrong.
For your own site there is no reason to rely on an estimate. Your analytics, your ad platforms and your store hold measured data, and the question there is a different one: whether everyone on the team reads the same number from it. That is a matter of definitions, covered in marketing metrics and KPIs.
When your own data is spread across several tools, the usual fix is to collect it into your warehouse and define each metric once. With OWOX Data Marts, an analyst publishes that definition as a Data Mart, and marketers read it in Sheets or Looker Studio.
Turn your data into decisions.
Governed data marts give you the clean foundation ML needs to actually work.
- No AI hallucinations
- Analyst-governed definitions
- Every number traces to SQL
Frequently Asked Questions
How accurate is Semrush traffic data?
In our comparison with Google Analytics on 787 websites (January 2020), Semrush traffic estimates deviated by about 45% for sites with more than a million sessions a month and by more for smaller sites. It is reliable for comparing sites and spotting trends, not for absolute traffic numbers.
What is the main difference between Semrush and Similarweb?
Both estimate website traffic from a sample of users and model the rest. In our test on 787 websites, Semrush was closer to Google Analytics for sites with more than a million sessions a month, and Similarweb was slightly closer for smaller sites and tended to underestimate traffic.
How accurate is Similarweb traffic data?
In the same test, Similarweb estimates deviated from Google Analytics by 57% to 61%, regardless of site size, and more often underestimated traffic. As with Semrush, use it to compare websites with each other, not to state a competitor's actual traffic.
