Competitive benchmarking plays a crucial role in marketing analysis for obvious reasons. Indeed, comparing your performance with industry bests can make a considerable difference in analysis completion. Let’s imagine, your marketing team was able to reach 95% of the desired rate during specific period. That doesn’t sound good, right?
On the one hand, if the overall market has shown 15% reduction, then your company was doing great. On the other hand, if you compare numbers and the whole market raised its figures by the same 15%, then your marketing team didn’t get satisfactory results.
There are dozens of tools used for competitive benchmarking. All of them apply different technologies for gathering traffic data and analyzing it, though. Therefore, professionals don’t expect perfect accuracy when using these tools. The only option to make a valid analysis is to identify the most accurate services and fully exploit them.
For these reasons, we decided to conduct research to figure out which one of the two most popular services provides more accurate analysis - SimilarWeb or SEMrush. As a reference point, we compared them to the Google Analytics data from more than 700 anonymized websites.
Before getting straight to the results, let’s examine how exactly these services perform their analysis.
Where SimilarWeb gets data
SimilarWeb mainly gathers data from:
1. Direct reports of website owners.
2. Reports from internet providers, which services are used by millions of users.
3. Public sources. Technologies that have indexing mechanisms for continuous data scanning.
4. Browser extensions, that obtain information about visited websites without user identification (e.g. cookies)
Where SEMrush gets data
SEMrush analyzes traffic by using anonymized data collected from third-party partners as well. Then AI algorithms start processing the data. The algorithm accumulates the data and divides it according to the behavioral analysis of the common patterns of millions of users.
Over the last decade, SEMrush has introduced a few additional tools, that go far beyond competitors’ search positions and website traffic information. These tools also reveal how rivals operate in their PR campaigns, paid search, content marketing what techniques they’re using to enhance their social media presence, and so on. SEMrush tools are useful due to the algorithm’s ability to use the most appropriate data sources for different marketing spheres (e.g. SEO, SMM, PPC).
How the services were compared in terms of accuracy
Both SimilarWeb and SEMrush provide clients with tools for analyzing competitors, the number of leads on their websites and traffic acquired with the help of affiliate partners. Having access to the anonymized data given by Google Analytics allowed us to consider this data fairly precise. All we had to do is to compare reports obtained from SimilarWeb and SEMrush for the same websites to Google Analytics data. The main goal was to identify imprecisions in data depending on the type of business industry and calculate how large were the gaps for both of the services.
What exactly we compared
We collected and analyzed data from SimilarWeb, SEMrush, and Google Analytics for the visits among all 700+ websites in January 2020. These websites are located in the USA, UK, Canada, Germany, and Australia. We divided them into groups according to their niche:
- Technology, electronics and computers,
- Entertainment,
- News and media,
- Finance,
- Jobs & Education,
- Health & Beauty,
- Professional services,
- Travel,
- Telecom,
- Retail.
Comparison method we pursue
To identify the rate of accuracy of rivals’ traffic data by SimilarWeb and SEMrush, we put together the data mentioned below in one table:
- Google Analytics anonymized data from 700+ websites with 100,000 visitors per month.
- Traffic data from the same websites determined by SimilarWeb and SEMrush.
We omitted websites that had low traffic according to Google Analytics. This low rate might be probably caused by filters for properties.
The next step was to calculate the absolute value of the deviations between traffic data given by SimilarWeb and SEMrush in comparison with Google Analytics data.
We had to make calculations using absolute value because deviations might be both positive and negative. Moreover, we were only interested in the rate of accuracy. This decision also eliminated the chances of mistake as positive and negative values could result in zero.
Finally, we subdivided the websites into 3 groups by the number of sessions happened in January:
- 100,000 – 500,000,
- 500,001 – 1,000,000,
- More than 1,000,000.
Statistics showed the following results…
The results are based on the standard deviation (amount of dispersion of all samples compared to the mean). A higher standard deviation implies a higher level of inaccuracy of the data given by SimilarWeb and SEMrush compared to Google Analytics. Here are the results:
- The deviation from the data by SimilarWeb ranged between 57% and 61% and there was no correlation traffic volumes on websites.
- SEMrush data, however, showed more explicit results. Websites from the third group (1,000,000+ visitors) indicated higher accuracy rate and a lower deviation(45%) against Google Analytics data.
Considering websites with 500,000 and more sessions, SEMrush showed 9-12% higher accuracy compared to SimilarWeb. Finally, for websites with the least amount of traffic, both of the services represented a high level of inaccuracy, even though SimilarWeb made a bit better result.
The differences in accuracy rate stemmed from distinction in gathering and processing algorithms as well as some aspects of clickstream data of SimilarWeb and SEMrush. These services operate with the help of AI and machine learning algorithms in order to approximate obtained data on all website sessions on the basis of clickstream data (Clickstream data refers to data on samples of website traffic). Therefore, smaller websites usually mean less accurate approximations based on clickstream data.
Then the question arises: how can you benchmark your website if it along with your competitor’s websites get low traffic, meaning low accuracy as well? In such a case you can compare your website against bigger competitors in your industry. While comparing industry leaders, you’ll notice how they perform and will be able to learn best practices.
Moreover, you’ll get familiar with up-to-date market trends. So the plan is you compare performance and best trends indicated by one of the services to your own current achievements. Thus, you’ll get the idea of how efficiently your marketing works.
The next graph represents the share of more accurate results by SimilarWeb and SEMrush compared to Google Analytics data. For instance, look at 100,000-500,000 visits and you’ll see that data given by SEMrush is more accurate for 53% of websites compared to only 47% of those analyzed by SimilarWeb.
After comparing the 100,000-500,000 sessions segment, using Standard Deviation and this graph we found out a peculiar fact: SEMrush data generally showed higher standard deviation, meaning that session approximations are more inaccurate. Nevertheless, SEMrush results are more precise in 53% of samples. The conclusion is that SEMrush makes fewer mistakes but they are more massive.
There are several factors, that define the accuracy rate of data:
- How Google Analytics system is set up, which pages have Google Analytics tags, and what these tags measure.
- Does Google evaluate the website as active? For example, for redirect website from ad network or promo website mainly used for paid advertising traffic, SEMrush will definitely demonstrate lower numbers.
- What is the share of organic traffic on your website? SEMrush gives more accurate results for websites with a greater share of organic traffic.
- Where it’s located and what niche a business operate in. Let’s assume that SEMrush has more websites and businesses from the USA in the finance industry. Apparently, the service will have more data about the US market in this specific niche. Certainly, calculations will be more accurate for that particular market.
The following two graphs represent the standard deviation and share of more precise data according to both of the services segmented by the industry
It’s clearly evident that deviation of SimilarWeb and SEMrush strongly depends on the industry:
There is one more correlation: the percentage of websites with more accurate results introduced by these services in comparison with Google Analytics also depends on the chosen niche:
Let’s consider the electronics and technology sphere. As you can see SimilarWeb and SEMrush showed 58% and 42% accuracy respectively.
Here is the graph that consists of positive and negative deviation values for both of the services.
It’s evident without expertise that the lower part of the plot (with negative deviation values) depicts more green dots. It means there’s a greater possibility that SimilarWeb will show lower traffic data figures against Google Analytics data.
Conclusion
We figured out that:
- The accuracy of both services is approximately similar.
- SEMrush is the winner in terms of medium-sized websites. There weren’t lots of mismatch in the analysis. But if the mistakes weren’t made, the accuracy would became higher.
- In the group with 1,000,000 sessions, SEMrush was again more accurate.
- SimilarWeb underrated traffic volumes more often.
Please note that there is no 100% guarantee of accuracy from neither SimilarWeb, nor SEMrush. For traffic analysis of your website, you have Google Analytics. However, SimilarWeb and SEMrush are suitable for benchmarking industry leaders and learning marketing trends. While using analytical tools, don’t forget about the origin of collected data and possible measurement deviation.
️Our team is extremely happy to announce that from now on you can use PayPal and Payoneer as payment methods in our network! 
️ We think many of you heard about the problems that Epayments has been facing recently. We hope that it all will be ok but since now we put the payments through this service on pause. Our team strongly recommends you to use other payment methods of our network to get your money:
Payoneer
And do you like The New Facebook?
Feel free to share your lucky source in the comments below 
And don't forget that if you have any questions, you can always contact our manager on Skype - prince94.10 and he will be glad to help you out 
Go to the Affbank Awards site (https://awards.affbank.com);
Find "The best nutra affiliate network" nomination;
Choose 
