How to get winning testing ideas with conversion research?

How do I optimize my website’s conversion rate?  Should I just try different ideas based on my gut feeling?  Should I optimize the loading speed?  Do I need to refresh the visual identity or rewrite the copy?  What should I A/B test? Without investigating the root causes that keep your customers from buying, registering, or…

How do I optimize my website’s conversion rate? 

Should I just try different ideas based on my gut feeling? 

Should I optimize the loading speed? 

Do I need to refresh the visual identity or rewrite the copy? 

What should I A/B test?

Without investigating the root causes that keep your customers from buying, registering, or reaching whatever goal you’ve set, you’ll be just making uneducated guesses. 

And if science has taught us anything, it’s that uneducated guesses rarely win. 

So, if having a wild guess won’t save your website, what will? 

The answer is Conversion Research.

What is conversion research?

Conversion research is the process of investigating the opportunities on your digital properties (website, mobile app, etc.) and gaining a more profound knowledge of your target customers. It’s the starting point that will fuel your conversion optimization ideas. 

A thorough investigation would provide you with the knowledge you need to achieve the following goals:

  • Write copy that aligns with your customers’ needs, treats their objections, and reduces their fears, uncertainties, and doubts. 
  • Catch the existing technical and UX problems.
  • Understand your users’ behavior (and where the leaks in the funnel are).
  • Create a user-friendly design that will help your users reach their desired goals. 

To reach all these goals, the conversion research process should contain all the following pillars:

  • Heuristic analysis
  • Technical analysis
  • Behavior analysis
  • Qualitative & Voice of Customer (VoC) research

But before diving into these pillars, we need to understand the initial strategy that we’ll be applying at least at the start of the process.

Gathering the low-hanging fruits

While conversion research will enable us to have a large number of insights that we act upon, our first direction should be focused on trying to find any conversion killers that can be implemented right away without having to test them. These could be some bugs in some devices or browsers, serious loading speed issues, mobile responsiveness issues, or design and usability issues. If the issue is not easy to implement, has a low impact on conversions, or has to be tested, then document it and you can get back to it at the end of your research. 

To start the process, we conduct a heuristic analysis. 

Heuristic analysis

A heuristic analysis is the evaluation of your website based on industry standards and best practices. What makes this different from just an opinion is that it’s based on collective knowledge in different industries. So the more you or your expert knows about these standards and best practices, the less opinionated your review becomes. 

Conducting a heuristic analysis means reviewing the website based on:

  1. Usability principles
  2. CRO best practices
  3. General copy principles
  4. Visual design principles


The expert evaluating the website could use Jakob Nielsen’s 10 usability heuristics for example. One of the principles, aesthetic and minimalist design, could be used to judge whether the pages contain much more information than the users need or irrelevant content. 

Let’s take an example:

Although this website is being praised and featured on for being visually pleasing, the aesthetics here are interfering with usability. The priority here should be on displaying the content and the functionalities in a user-friendly way. 

To evaluate the usability of your website yourself, here’s a checklist you can use:

  • Is your website look consistent with the user’s expectation of what a website should look like (the position of the navigation bar, headline or call to action, the shape of a button, etc)
  • Are your website elements (buttons, headlines, links, etc) consistent across different sections and pages?
  • Are you showing relevant, sufficient information with no excess that could cause the users a cognitive overload (too much information)?
  • Are you preventing your users from making any errors such as misunderstanding the function of some elements?
  • Is the text on your website readable (size and contrast)?
  • Do you have any distractions keeping your users from consuming the content (eg: background videos, useless animations, cookie alerts, pop-ups, other promotions, etc)
  • Do your action elements have different states (default, hover, clicked, and active)

CRO best practices

When multiple experts test a certain pattern and find that it’s helpful, a best practice is born. For example, GoodUI has published a pattern called “Gradual reassurance”. The idea is to add relevant and easy questions to your form in the beginning before asking for any personal information. 

This pattern has been tested 7 times by different people/brands and has proven to create an average uplift of +21.8% in sales. This pattern doesn’t mean that you’re 100% sure that there will be an uplift. You should A/B test it with VWO. Take an all-inclusive free trial and see how VWO makes A/B testing easy! 

However, testing an idea like this will have a higher probability of success than a random idea based on someone’s opinion. Many CRO experts are publishing their wins and we should learn from them and even duplicate them. 

The other part of best practices comes from either psychology concepts such as the decoy effect or some lessons we learn from the marketing godfathers like David Ogilvy. 

Copy review

Based on research conducted by Unbounce, based on 40k landing pages, a copy is twice as influential as design in conversions. So reviewing your copy should be an important task you should invest in. 

Here’s a checklist you can go through to analyze your pages. 

As an exercise, choose an important landing page that you want to optimize and go through the checklist below. Rate each question with a number from 0 to 10. 

Note: Reviewing your copy also includes reviewing your offer too. Rewriting your copy is great. Elevating your offer is even greater. That’s why some of the checklist items will be related to reviewing your offer. 

  • Does the headline explain your product/ service value right away?
  • Is your headline language familiar to your target audience? 
  • Does your above-the-fold copy attract the attention of the users to continue reading beyond the fold?
  • Are your claims supported by proof like testimonials, ratings, authority figures’ endorsements, number-based outcomes supporting the transformation after buying your service/product, the total number of customers, etc? 
  • Are you going through the users’ objections and responding to them?
  • Are you offering any incentives (free shipping, money-back guarantee, free trial, etc)
  • Are you using any form of urgency as a motivation? 
  • Do you have only one main call to action?
  • Are you going through all the main product/service benefits? 
  • Can your audience self-identify as the target for your offer?

Visual design principles

To create a visually pleasing user-friendly interface, you need to follow a set of simple but powerful design rules. These will help you:

  • Create a great reading experience
  • Create a visual hierarchy for your elements to communicate what’s most important
  • Get a positive first impression
  • Create a professional and coherent look 
  • Help the user navigate the different elements with ease
  • Help create a story with the design that goes hand in hand with your copy

Qualitative & voice of customer research

It’s business/marketing 101 to understand who we’re serving so we can market better. The goal here is to understand the users’ pains, needs, goals, and objections. Qualitative research will enable us not only to document important information about our audience but will also help us understand how they’re framing their needs, objectives, fears, uncertainties, and doubts using their own words. That will be a gold mine when we start experimenting with different copy ideas. 

To collect this information, we can:

  • Collect your audience quotes from online forums, Reddit, Facebook Groups, and even YouTube comments. This is where users are expressing freely their needs, problems, agitations, goals, or even sharing their wins.
  • Conduct user interviews
  • Analyze emails and chat transcripts
  • Create and analyze surveys

Learn how to capture and action the voice of the customer with a free recording of Ali Good’s webinar.

Technical analysis

The reason your funnels are leaking could be simply technical. It could be bugs, responsiveness problems, loading speed problems, or cross-browser compatibility problems. 

An expert could review all these pillars. But in case you want to do this analysis on your own, here’s how you do it:

Cross-device testing

Go through the user journey from the starting point (generally the landing page) to the ending point (generally the thank you page). Start with your device and see if you have some bugs such as unresponsive buttons or any strange behavior. You can start by manually doing that with real devices like your laptop and phone. But since you don’t have tens of devices, different SaaS apps can help you with that. Apps like Browserstack and LambdaTest can give you the ability to test on different devices and different browsers. And that brings us to cross-browser testing.

Cross-browser testing

Not only devices could affect how your website looks and behaves, but browsers matter too. In 2022, cross-browser compatibility has come a long way and it’s in its best state. That said, there are still issues, especially with old browsers. To investigate the priority for each browser, you can explore the report using Google Analytics by going to Audience > Technology > Browser & OS. 

Not only can you see which browser matters most, but you can also explore the conversions per browser.

If you notice a suspiciously low conversion rate (which can be 0 sometimes) where the sample size is not that low, it’s time for investigation.

Here’s a scenario: 

You notice that IE 10 has a 0.03% conversion rate for a sample size of 3280 sessions. You also notice a 9.34% conversion rate for Opera mini for a sample size of 3040. You check the average conversion rate for all browsers and it’s around 9.5%. With that information, you know that there has to be something broken on IE 10. No big difference for Opera mini though so you can let it off the hook. 

You open the website with BrowserStack and conduct a walkthrough. Aha! The call to action doesn’t work somehow. File this issue to your developer and it’s fixed. No need for testing as we fix bugs, we’ve just gathered a low-hanging fruit.

Page speed optimization

This is indeed a domain of its own, but I believe that most of the problems websites have are easily preventable. So before we tackle optimization, let’s see the most common possible problems:

  • Non-optimized images. With big images come big opportunities. 
  • Non bundled CSS and/or Javascript files
  • Absence of a lazy loading strategy
  • Not using a CDN to serve static assets
  • Overloading your website with multiple CSS and Javascript libraries (way too many Shopify apps or WordPress plugins can cause that)

Other problems may affect your page speed but mainly, the mentioned problems have the biggest impact. 

But before you start investing in page speed optimization, you need to make sure that the improvement will be worth your investment. 

A 10ms improvement could mean a lot for a company like Amazon. Every 100ms in added page load time cost them 1% in sales. So that 10ms improvement in sales means an uplift of $380 million (0.1% of Amazon sales). 

For other companies, that 10ms improvement could mean only an additional $1000 or less per year. So depending on your company’s scale and your page loading speed, you can judge whether the optimization is needed or not. 

If you find that it’s worth it, make sure that you:

  • Eliminate any render-blocking resources. Javascript scripts, CSS files, or images that are not being used in the fold can be deprioritized and lazy-loaded. 
  • Optimize your images and lazy load them.
  • Use a CDN like Amazon S3 to serve your static files.
  • Compress your Javascript and CSS

As a rule of thumb, you’d be targeting a loading speed of 1-2 seconds, especially for your key pages. 

Behavior analysis

Funnel analysis

Conversion optimization is not a random process where we can start anywhere on the funnel. We need to understand where our optimization efforts should be focused first. Funnel analysis is one of the best ways to understand not only the user behavior across the different steps but also where should we start the optimization effort. The closer the users are to the checkout, the more impact our efforts will have on conversion. 
With VWO, you can easily get a funnel visualization to see where the biggest drop-off is. For example, if you see a suspiciously big drop-off on the checkout page, you can prioritize investigating that as this has a big potential for a conversion rate uplift.


Heatmaps are a way to visualize the users’ behavior on a certain page. You can answer questions like:

  • What attracts the users’ attention most?
  • Where do users tend to lose disinterest? 
  • What important information that the users aren’t reaching?

Not only will a good portion of the insights be pulled from heatmaps, but these visualizations are very useful when presenting your insights to a manager, a stakeholder, or a colleague. It’s very hard to argue with heatmaps and that will make your conversion optimization process much smoother. Opt for an all-inclusive free trial with VWO to experience the benefit of heatmaps.

Session recordings

Before the session recordings concept was invented, experts were using controlled testing environments where they bring people into a lab and hide behind an opaque glass to see how users interact with their products. Or they can even be present which, in either case, will affect the behavior of the users. No to mention that this method is so expensive and far from being scalable. The alternative is session recordings

This way you can capture the real unfiltered behavior of hundreds of people.

Form analytics

Forms can be hard to optimize. While applying CRO and usability best practices to your forms is important, it’s even more crucial that you understand how your users are interacting with your form. Session recordings are good for this task but an even more important and complementary tool is VWO’s form analytics. This tool can help you understand which fields are causing more friction, where users are leaving, how many people have reached the form, how many people have interacted with it, how many people submitted it, etc. 

What makes this a better tool than traditional user testing and user recordings (which have their benefits and use cases) is that it quantitatively offers the data. Seeing few users quit after reaching a certain field in user testing is way less accurate than seeing a 65% drop-off rate for a sample size of 9000 people. 

Form Analytics

Take a free trial to try it out for yourself and explore its capabilities.

On-page surveys

Tools like Google Analytics are very important to explain what’s happening. For example, if you have a high exit rate on a landing page and you want to understand the reason behind it, Google Analytics or any similar tool won’t give you the answer. It’s your role to investigate that. On-page surveys can help you understand that by asking your customers directly. You can even target different behavior such as an exit once landed, an exit once reached half the offer section, etc. That way, you can get a piece of much more accurate information based on user behavior. 


Conversion optimization isn’t just about A/B testing. This is one of the most common misconceptions. A/B testing is a very important pillar. But to have experiment-worthy ideas, we need conversion research. This is how we can generate ideas that have an excellent potential to improve the conversion rate. With different tools and approaches, conversion research will give us not only the information needed to run an experimentation program but can also affect other areas such as product/service design. 

Now that you know how to conduct conversion research, it’s time that you understand what comes next: A/B testing.

Why you should test your SEO ideas before you ship them

Your conversion rate optimization (CRO) strategy fundamentally runs on two wheels—Search engine optimization (SEO) and A/B testing. When you optimize your landing pages consistently to improve user experience by engaging visitors on your website, it signals to Google that your website fulfils the user intent. As a result, Google incentivizes you with a high ranking…

Your conversion rate optimization (CRO) strategy fundamentally runs on two wheels—Search engine optimization (SEO) and A/B testing.

When you optimize your landing pages consistently to improve user experience by engaging visitors on your website, it signals to Google that your website fulfils the user intent. As a result, Google incentivizes you with a high ranking in the search.  

SEO has evolved from being a traditional keyword-oriented model to a full-blown strategy as Google keeps changing its algorithm. User intent has become one of the top factors for ranking, but at the same time, it has become challenging for marketers to predict visitor behavior while making any changes in their SEO strategy. 

The logical resolution to the problem of figuring out what Google and visitors want from your website is rigorous testing of your SEO efforts.

This is good news because a) it is possible to run A/B tests without impacting SEO negatively and b) Google encourages A/B testing as long as it is done bearing in mind recommended best practices. 

In this blog post, we have covered the basics of SEO testing and how A/B testing can help you optimize your SEO efforts. 

What is SEO testing?

SEO testing enables businesses to optimize their website performance on the search engine result pages (SERPs). If you have ideas to optimize SEO for your website, you can do so by testing your ideas before you execute them at scale, which otherwise might pose a negative impact on your business if you ship the changes untested. Well-informed decisions also facilitate planning a smooth CRO roadmap in terms of both—strategy and budget. 

The big advantage here is that Google encourages you to A/B test your SEO changes. You may test pages such as product pages, category pages, and elements like meta tags, copy format and relevancy, headline, call-to-action buttons, and other SERP features. These modifications are crawled by Googlebot and help Google learn new information and assess the implications of changes on your website for ranking. 

SEO testing is different from A/B testing, as explained in the following sections.

What is Googlebot?

Google defines Googlebot as the crawler that visits web pages to include them within the Google Search index. In simple words, it is computer software or a bot that keeps learning new information on your web page and indexes the changes in Google’s database. It supports the latest JavaScript features that help them render the web pages to make ranking decisions. 

Google declared the bot as the new evergreen Googlebot. And yes, there is just one Googlebot and it doesn’t like crawling duplicate pages as duplication confuses it. Let’s understand its function in bot-based testing.

Bot-based testing

With SEO bot-based testing, you run the split URL test and assess the control and variation response in a given period. It is done either to accept and deploy the changes to similar pages on your website or reject the changes before moving on with your next experiment. 

For example, if you want to test your call-to-action (CTA) buttons to rope your visitors in organically, you can compare the copy and positioning of your CTA on your website by running a split test. The bot in such a test will be presented with unique web pages with unique CTAs that you want to test. It will crawl down the web pages to index the new information which you have presented through the test, thereby mitigating the chances of confusion for the bot. You will get the winner once the test signals a significant influx of visitors into the sales funnel, which eventually will be reflected in SERPs. 

On the contrary, if you want to run an A/B test for CTA copy and positioning, you present a control version to a set of users and a variation to another set of users. Here, the bot may not be able to distinguish between the two pages (until you use link attributes to your original page, as discussed in best practices below), and it may confuse your testing situation with cloaking—a practice lethal for your SEO.

A/B split testing for SEO
Image source: Semrush

How do A/B testing and split URL testing impact SEO?

Simply put, they don’t. SEO is not negatively impacted by testing activities as long as you’re not deliberately trying to confuse Googlebot. 

Also, A/B testing and split URL testing are not the same. Your visitor will land on a variation or control purely by chance in an A/B test, whereas in split URL testing, they will land on statistically similar groups of pages as tests are hosted on different URLs. So, in A/B testing, you split users to test experiences, and in the latter, you split pages with distinct experiences. 

For example, if you want to test your title tags which are key to SERPs, you can compare your title tags with SEO split testing, which will allow you to create groups of pages, say product pages, where you want to ship your changes to. These pages are categorized into unique testing groups and hence appear alike to visitors and Googlebot. This type of testing is straightforward and effective for SEO. Whereas, it gets a bit tricky in A/B testing if you don’t follow the best practices. 

Split URL testing is preferred for SEO when you need to test major changes like content format, navigation, design, etc. However, you can A/B test your SEO efforts as well keeping in mind a few factors that we will discuss in the following sections.

Automation Page Graphics V4 0 Implent Seo Testing Final
Image source: SEO clarity

Best practices for A/B testing SEO activities

Google advises stringent guidelines on SEO A/B testing if you don’t want to hurt your SEO:

  • Avoid cloaking

According to Google Webmaster guidelines, cloaking is showing one version to humans and others to bots. It’s deceptive to the search engines, bringing more harm to your website. Try always to serve the original content, otherwise, Google can demote your website.

  • Use 302 redirects

Always use a temporary 302 URL rather than 301. You may need to redirect some of your URLs to consolidate similar content pages into one or move your site to a new address.

  • Use rel=“canonical” to avoid duplication 

While running an A/B test for minor changes on your web page, always add a link attribute to your alternative URLs for your testing pages. This linking enables Google to distinguish between the original and the variation. For example, Google can index your homepage as a test variation if you want them to understand that you are running a test on your homepage. You can do so by identifying the page by using rel=”canonical”. It also signals to Google that other variations are close copies and should not be confused with the original URL.

  • Experiment runtime

Figure out what you want from the changes on the website. Since Google changes its algorithm all the time, be mindful of your test’s duration as your SEO might get affected if Google updates its algorithm while your test is running. This may, in turn,  impact the metric you are testing. Such situations certainly get you skewed test results. It is advisable not to run SEO tests for long periods, unmonitored. If you want to see the positive impact of the variation, then you may want to conclude the test and make the changes live. However, if you want to gauge the scale of impact, then wait until your test reaches statistical significance. 

Blog Banner Impact Of Ab Testing On Seo

How does VWO help in running A/B tests and split URL tests without impacting SEO?

According to Google webmaster, website speed is one of the key factors impacting website ranking. The VWO Asynchronous SmartCode loads in parallel with your website, thereby having zero impact on your website load time. You can customize time-out parameters for the VWO SmartCode to execute and make sure that your visitors do not experience any inconsistent behavior while using the website. If the VWO SmartCode fails to execute within a specified time, it stops and then loads the original content of the page.

VWO Testing ensures that there is no negative impact on your website ranking while A/B testing your SEO elements. For example, if you want to A/B test some minor changes in the website, such as the copy, image, or color of a button, it usually has no impact on the page’s search result snippet or ranking. However, if you want to make changes like copy length and redesign your website, it is recommended that you follow best practices such as using query parameter rel= “canonical”.

For instance, if you have two different URLs ending with A.php and B.php, you can use the below code on B.php to avoid content duplication: 


    <link rel=”canonical” href=”” />


Although Google will index both the pages, you are enhancing Google’s knowledge to consider the original A.php for SEO by adding the above code. 


SEO is a black box. Nobody knows the future of Google’s algorithms and how the search engine perceives your unique web pages. 

The closest you could get to understanding it is by acknowledging that Google is getting smarter every minute as you read this blog piece. The only way to keep your SEO game up and running in this competitive age is through consistent experimentation. 

Just like the COVID vaccines. There are so many other factors at play you can only say with some reasonable amount of confidence that you can expect the results to come out a certain way.

Kevin Indig, Head of SEO, Shopify

The only solution to crack your SEO efforts for revenue generation is to keep testing, keep optimizing!

Blog Banner Impact Of Ab Testing On Seo Bottom