UX case study: improving conversion for the Semrush free tools
Overview
Semrush’s free tools attract mainly desktop traffic, yet little was known about who uses them, what for, and how. That gap made them hard to improve.
Semrush is one of the largest SEO companies. It serves small brands and companies as well as large corporations. The company wanted to improve its free tool pages, including the SEO site audit, the backlink checker and the keyword rank checker.
I am Nathan Boomsma, a UX designer and researcher, and this UX case study covers my work on conversion rate optimisation for these pages. My role was UX researcher and UX designer. I set out to understand the ideal user and find ways to increase conversion.
My role
UX Design
UX Research
Date
Client
Semrush
The challenge
The challenges
The main challenge was limited influence. The teams at Semrush worked in silos, and the teams I worked with had no influence over the sign-up flow or the pricing of the product. Both affect conversion, so I had to find gains on the free tool pages themselves.
Other constraints made the work harder:
Unclear ideal customer profile (ICP). The ICP was not well defined, so I interviewed a wider range of users.
Lower traffic than expected. Some pages received less traffic than the team expected.
Different time zones. The Semrush team worked across time zones, so async alignment mattered more than live meetings.
Underpowered A/B tests. We ran a series of A/B tests, but many were underpowered. That made it difficult to form the right hypotheses from the qualitative and quantitative data.
These constraints shaped the approach. I needed research that worked without control of the full funnel, and a way to learn from tests that did not give a clear answer.
The result
The result
Sign-ups for a free trial of the paid Semrush account increased by 180%. We ran 4 A/B tests a month during the project.
This work shows how I connect qualitative research, A/B testing and meta-analysis to improve the user experience and conversion, even without influence over the sign-up flow or pricing.
The process
The process
User interviews and website walkthroughs
I started with qualitative research to learn who the users were and what they needed. I interviewed six people:
3 business owners
3 online marketers who work at a larger company
Each session began with discovery. I focused on what participants saw as a problem, what mattered to them, and how they made decisions. Then we did a website walkthrough. Participants explored the pages the way they normally would, and I asked follow-up questions when needed.
Analysis: themes, heatmaps and first impressions
I ran a thematic analysis of the interviews and gave each theme a score from 1 to 5. A 5 marked the worst issue I found. Three themes came out:
Reduce mental effort
Product highlights
Transparency awareness
I also reviewed heatmaps from aggregated sessions and the results of first impression tests. The tests covered what users thought the brand offered and how likely they were to use the website. I shared all findings in a research slide deck.
From findings to A/B tests
Next, we planned a series of tests we could design and run. We gave each idea a short title and placed it on a matrix. We started with the changes that had the most impact for the least effort.
After each test, we analysed the data to look for a better angle. We noted the hypothesis behind each result and built the next test on it.
Meta-analysis of inconclusive tests
Many tests were inconclusive. To get value from them, I ran a meta-analysis of 15 tests. The work showed me why meta-analyses matter, how to run one, and how closely they connect to A/B testing.
UX case study: improving conversion for the Semrush free tools
Overview
Semrush’s free tools attract mainly desktop traffic, yet little was known about who uses them, what for, and how. That gap made them hard to improve.
Semrush is one of the largest SEO companies. It serves small brands and companies as well as large corporations. The company wanted to improve its free tool pages, including the SEO site audit, the backlink checker and the keyword rank checker.
I am Nathan Boomsma, a UX designer and researcher, and this UX case study covers my work on conversion rate optimisation for these pages. My role was UX researcher and UX designer. I set out to understand the ideal user and find ways to increase conversion.
My role
UX Design
UX Research
Date
Client
Semrush
The challenge
The challenges
The main challenge was limited influence. The teams at Semrush worked in silos, and the teams I worked with had no influence over the sign-up flow or the pricing of the product. Both affect conversion, so I had to find gains on the free tool pages themselves.
Other constraints made the work harder:
Unclear ideal customer profile (ICP). The ICP was not well defined, so I interviewed a wider range of users.
Lower traffic than expected. Some pages received less traffic than the team expected.
Different time zones. The Semrush team worked across time zones, so async alignment mattered more than live meetings.
Underpowered A/B tests. We ran a series of A/B tests, but many were underpowered. That made it difficult to form the right hypotheses from the qualitative and quantitative data.
These constraints shaped the approach. I needed research that worked without control of the full funnel, and a way to learn from tests that did not give a clear answer.
The result
The result
Sign-ups for a free trial of the paid Semrush account increased by 180%. We ran 4 A/B tests a month during the project.
This work shows how I connect qualitative research, A/B testing and meta-analysis to improve the user experience and conversion, even without influence over the sign-up flow or pricing.
The process
The process
User interviews and website walkthroughs
I started with qualitative research to learn who the users were and what they needed. I interviewed six people:
3 business owners
3 online marketers who work at a larger company
Each session began with discovery. I focused on what participants saw as a problem, what mattered to them, and how they made decisions. Then we did a website walkthrough. Participants explored the pages the way they normally would, and I asked follow-up questions when needed.
Analysis: themes, heatmaps and first impressions
I ran a thematic analysis of the interviews and gave each theme a score from 1 to 5. A 5 marked the worst issue I found. Three themes came out:
Reduce mental effort
Product highlights
Transparency awareness
I also reviewed heatmaps from aggregated sessions and the results of first impression tests. The tests covered what users thought the brand offered and how likely they were to use the website. I shared all findings in a research slide deck.
From findings to A/B tests
Next, we planned a series of tests we could design and run. We gave each idea a short title and placed it on a matrix. We started with the changes that had the most impact for the least effort.
After each test, we analysed the data to look for a better angle. We noted the hypothesis behind each result and built the next test on it.
Meta-analysis of inconclusive tests
Many tests were inconclusive. To get value from them, I ran a meta-analysis of 15 tests. The work showed me why meta-analyses matter, how to run one, and how closely they connect to A/B testing.
More works
Project
©Nathan Boomsma 2026. All rights reserved.
Project
©Nathan Boomsma 2026. All rights reserved.
Project
©Nathan Boomsma 2026. All rights reserved.