Table of Contents for our Analysis of SEMrush’s SEO Ranking Report:
- Part 2: How website security affects SEO.
- Part 3: How backlinking and referring domains influence SEO.
- Part 4: Content length and its ranking factors.
- Part 5: Which parts of on-page SEO are the most important.
- Part 6: How site traffic and behavior cues affect SERP ranking.
As more businesses rely on the Internet to make transactions and grow, the demand for further innovation in web optimization also rises. This means that the ranking battle is only becoming more competitive.
In this six-part series, Edgy Labs evaluates SEO industry leader SEMrush’s latest research and findings.
If you want to win this battle, you need an effective weapon to help you.
SEMrush is a powerful and respected tool when it comes to search engine optimization (SEO). Its credentials are so impressive that it is becoming a desired SEO tool for many businesses.
If you’re not convinced, consider that SEMrush has:
- over 1.5 million users as of writing
- 140 databases all over the globe.
- 30 tools and reports in the SEMrush suite that can help boost online marketing strategy
Add the 2.5 billion keywords and 3 trillion backlinks that SEMrush can track for your business, and the company’s software seems like a “must-have” tool for every modern marketer.
The Ranking Report Research Study and its Purpose
Since the company was founded, SEMrush has never stopped developing tools for better web optimization. Over the years, the company has gathered an enormous amount of data and gained access to tons of valuable online resources in addition to building a solid reputation.
All of this information and experience is what drove the SEMrush team to put together a research study to uncover the variables that affect SERP ranking to date. The results are not only interesting–they’ll make you re-think your SEO strategies.
The team’s main goal? To answer the ultimate SEO question: what makes a website rank higher?The SEMrush SEO ranking report pulls from over 2.5 billion keywords.Click To Tweet
The Research Methodology and its Results
Most studies about ranking factors today are products of correlation analysis.
Correlation analysis is a method aimed at finding the relationship between two or more variables from a subject group. The result is typically presented in a graph as shown in the below:
While correlational studies are helpful in predicting outcomes and, at times, causality, SEMrush did not rely on this method for the present study because of the technique:
- will not work well when a variable is dependent on other variables.
- is sensitive to outliers which affect the correlation coefficients.
Considering these two points: SEMrush chose to deviate from the norm; the team, therefore, used more complex and fine-tuned methods to find the most precise factors behind SERP ranking.
Note: The objective of this study is to hone in on the most current and accurate search ranking factors. Any improvement suggested in this study will not guarantee any increase in ranking position. However, we encourage you to use this as a guide to further expand your SEO strategy.
The SEMrush team kicked off their research with 600,000 keywords derived from the following:
- Worldwide base (US, Spain, France, Italy, Germany, and other leading countries globally)
- Pages’ position in search results
- List of ranking factors
Then, the study used the Random Forest machine learning algorithm to come up with a list of factors.
The Random Forest algorithm works by clustering a data set into different groups called decision trees. In this study, various ranking factors made up the data set. Each decision tree contains random variables taken from the data set.
Repeating the clustering process countless times will produce predictions. Trees with the most predictive powers will be shown as the results. This approach helped to create the list of factors that influence a specific page’s position. The team took all of the pages in the 20th position and verified that the trends for the entire dataset coincided with the results of the 20th position.
Each search result page was analyzed using the following factors:
- On-page factor
Then, the results were classified into the following four keyword volume classes:
- Low- volume
- Mid- volume
- High- volume
- Very high- volume
Following this process, SEMrush arrived with below list of factors. Note that the following factors are represented in order of importance (most influential to the least).
The results of the SEMrush study and approach are surprising, but it will definitely give you a fresh, cutting-edge perspective on the SEO science. The most revealing conclusions are that the study indicates are:
- the number of website visits is the most important page ranking factor. This means that high authority sites that gain the most online traffic have better chances of making it to the top.
- Behavioral signals like time on site, pages per session, and bounce influence the ranking of high authority sites.
- Referring domains, or backlinks, are still important in page’s ranking position.
- The study confirms Website Security as a ranking factor, and proves Google’s statement that “HTTPS is important. However, it has an insignificant effect on a site’s ranking.”
- The remaining on-page SEO factors had very little influence on page ranking. Therefore, they are not considered as major ranking factors.