The Ultimate Guide to CRO
“There are two ways to generate more revenue online through Internet marketing. The first is driving more traffic to your website. The second is converting more or your existing traffic. Conversion Rate Optimization focuses on the latter.”
Nick Slettengren, Co-Founder Power Digital
In this article you will get:
- A clear definition why CRO matters
- Well defined and proven process for conducting your CRO
- A step-by-step CRO guide you can follow along with
…Let’s start here.
Conversion Rate Optimization (CRO), Why? Why does CRO matter, that is?
1) Make more money
CRO directly correlates to your bottom line, that EQUALS money. Then when you increase your conversion rate you can also…
2) Increase advertising spend
As you convert more of your existing traffic you can afford to reinvest in bringing in more traffic e.g. PPC, SEO, Social Ads, etc.
3) Increase market share
Now let’s tie it all together. The more your website visitors convert into leads or sales, through increasing your conversion rate, the more traffic you can now buy more of, the more NEW customers you can acquire, the more business you get. Simply put, market share increases through CRO and it all starts with conversion rate at the core.
What is CRO?
Let’s begin by defining what CRO is and what it is not.
- Finding the truth in why visitors are NOT converting
- Fixing the problem
- Increasing the conversion rate
- That’s pretty much it at a very high level
CRO is not:
- Changing your buttons from red to green
- Guessing at what you should A/B or split test
- A one and done fix
To my first bullet point on what CRO is not, like most digital marketing channels, there is always some “guru” swearing they have a magical CTA, button color, or messaging sequence that wins all the time with conversion.
That cookie cutter approach and those kinds of assumptions we want to stay away from in CRO. Your industry and website is unique and cannot be blanketed in a ‘one size fits all’ solution for increased conversion.
CRO starts with analysis and never assumption. With the tools available now to analyze user behavior on your website, why would you ever guess what you should test based on a hunch?
Last, CRO is not a one and done fix. Testing for conversion is ongoing because unless your conversion is 100% it can always be better!
Conversion Rate Optimization Defined
Conversion Rate Optimization (CRO) is an Internet marketing channel that creates a system for increasing the percentage of visitors to a website that converts into customers, or more generally, take a desired action on the webpage.
More specifically, conversion rate optimization evaluates your website’s sales funnel. Here we look at ways to improve a website in order to attract customers to your service of product offering.
CRO starts with analytics and user feedback. From looking at user data you can determine your hypothesis to test. From there you create a new variation of your webpage that supports your hypothesis and you test against your current version. The winning page is determined by the increased percentage of people converting e.g. signing up for your service, buying a product, opting into your newsletter, etc.
CRO Keywords to Familiarize Yourself With
Here are a few concepts that will come up time-and-time again.CRO Analysis
A conversion rate optimization analysis looks specifically at user behavior on your website. This process uses analytics data gathered from various analytics platforms that measure user behavior. Examples of these platforms are Google Analytics, Crazy Egg, Hotjar, and several others we will get to later.
User testing is getting real feedback from real people on the user experience of your website. Whether you hold a focus group, or use a virtual testing platform like UserTesting.com you can give direction and get critical feedback for your CRO hypothesis.
Calls to Action (CTA)
Calls to action are buttons, links, web forms, etc. that people can take action on your website that take them to a next step. The CTAs move things forward in a website experience and always point towards conversion. Examples include a “Buy Now” button you see on Amazon.com or a “Get in Touch” link you would find on Powerdigitalmarketing.com.
This is the path the user takes, sometimes referred to as the ‘user journey’ on a website. This flow can be different for each user but an example for an eCommerce website could be: Home Page => Category Page => Product Page => Checkout => Thank you page
A/B of Split Testing
A/B testing and split testing are one and the same. When we are ready to test our CRO hypothesis on a specific user metric we will develop another webpage that supports our hypothesis and test it against the current webpage. The winner is determined by the highest increase in conversion.
Multivariate Testing (MVT)
Instead of testing just one element of a webpage against another webpage like A/B testing, multivariate testing allows as many page elements tested as you want. In a multivariate test you may test, copy, CTAs, and photography all in one test against another webpage to determine the winner with increased conversion.
CRO Statistics We Found Interesting
According to Google Searches and interest from Google trends ‘conversion rate optimization’ is on the rise! – Google (link to Google Trends Report)
The average ROI on CRO is 223% -Venture Beat
7 to 10 marketers practicing CRO use their findings to inform other marketing channels. –MarketingSherpa
On average companies engaging in CRO spend $2,000 a month on CRO tools. -Forbes
62% of online marketers use CRO to increase lead generation on their webpage forms –Ascend2
How to Calculate Your Conversion Rate
- Total conversions is defined by the number of people who took the action to convert for a specific task.
- Examples – email newsletter, purchase, filling our request form, etc.
- To calculate your conversion rate, divide the number of conversions by the total number of visitors to your website.
- Example – 10 conversions divided by 1000 website visitors is a conversion rate of 1%
What website metrics should you be looking at to identify for Conversion Rate Optimization?
Your website’s bounce rate is a percentage of visitors to a particular website who navigate away from the site after viewing only one page. If your bounce rate percentage is high for a series of webpages that is not good. Finding the outliers of webpages that have high bounce rate above your website average is a good place to start to find webpages fitting for CRO.
Each webpage also has an Exit Rate that is defined as the percentage of visitors to a site who actively click away to a different site from a specific page, even after visiting other webpages. Not to be confused with bounce rate, exit rate can have a series of webpage views before an exit. High exit rates are a great indicator of a hole in your user journey that can we optimized with CRO.
Average Time on Site
An Interest / Engagement Metric, Average Time on Site tells you how long people stick around on a specific webpage. These metrics show interest and lead to engagement on the webpage. If they are low and even combined with say a high bounce rate, that can point to poor performance and an opportunity for CRO.
Average Page Views
Average Page Views also serves as an Interest / Engagement Metric. It tells you how many pages visitors visit on average for your website before leaving. If people are visiting more multiple pages on average, it tells the search engine people are enjoying your website by digesting more of the content on numerous pages. On the other hand, it can also point out a lack of clarity from the user journey and may need to be evaluated with CRO.
Starting Your CRO Journey
When starting your CRO roadmap it is best practice to start with your goal conversion in mind. This could be email sign ups, product checkouts, lead form submissions, etc. After all, your end goal conversion will have the biggest impact to what you are trying to achieve.
With the end goal conversion in mind, you are going to want to reverse engineer your funnel. By starting at the bottom of your funnel, you will avoid measuring micro-conversions that may not lead to your final conversion goal that are more typical towards the top of the funnel.
A Quick Lesson…
A good example of this is focusing first on clicks from the home page to a service level sign-up page. After running a test, you find 20% more people move to the sign-up page with your new webpage variation. That’s a big win, right?
What you are not seeing is that even though there was increased click-through-rate (CTR) from the home page to the sign-up page, it decreased page sign-ups by 5%. By not starting with the conversion goal in mind, you sent people to a weak sign-up page that could not convert the newfound traffic. As a result, the home page test was putting the cart before the horse and negatively impacted the end goal conversion, which was page sign-ups.
This is not an uncommon scenario and must be considered when architecting where to start your conversion rate optimizations. Always start with your end goal in mind.
There are two data types you will be looking at in CRO quantitative and qualitative. Quantitate referring to measuring by the quantity of something numerically. Where qualitative refers to the quality, size, appearance, value, etc.
When collecting data for CRO it is good to address both.
One of the best ways to start gathering data is through Google Analytics. Here you will be able to see numeric data on how people are using your website, which steps in the funnel they are taking on their user journey, and where they are dropping off from, to name a few.
Funnel Visualization for CRO Analysis
Funnel visualization is a great way to start with gathering data for CRO within Google Analytics.
Getting started click on “Conversions” then “Funnel Visualization” to view your funnel conversion rates on your website.
If you do not have funnel visualizations set up yet, click here for our funnel visualization set up guide.
The funnel visualization will show you what percentages of people are moving through the steps in your funnel. For example, if you are running a website selling shoes, your funnel may track people in the shopping cart moving from the shipping options => payment details page => complete purchase page.
Above is an easy three-step funnel, but funnels come in all shapes and sizes depending on your conversion goal and can become more complex with multiple user journeys.
If you have a large website with many user paths, consider using Users Flow within Google Analytics to define the most popular user flows in your website and start to build funnel visualizations from those.
When it comes to evaluating your funnel and looking for CRO opportunities, begin looking for drop offs in your funnel. These are gaps in the percentage of people going from one stage in the funnel to the next. Here is an example from a Bike Tour company.
Now that you have identified the drop offs with the largest percentage of visitors not going to the next step in your funnel you can begin to add those pages to the list to evaluate with CRO. With these webpages in mind, begin to consider some of the on-page elements that might be holding the page back from higher conversion.
Load Time Analysis
Do you hate waiting for websites to load? You are not alone most people do and according to Kiss Metrics, a 1 second delay in load time can result in a 7% reduction in conversion. Additionally, 40% of visitors abandon a website if it takes more than three seconds to load.
With conversion rate optimization the goal is to increase conversion. Site speed is a low hanging fruit opportunity to juicy to ignore.
To run a site speed analysis in Google Analytics, click on “Behavior” => “Site Speed” => “Overview.” Here you will get the average site speed of your website and can drill down to individual URLs that are part of your conversion funnel you have identified.
Under a two second load time is excellent, a three second website load time is good, four to five seconds is average and should be improved. Above a five second website load speed is poor and should be addressed sooner rather than later.
Here is a link to our site speed optimization guide.
Conducting Qualitative Research for CRO
When it comes to conducting your qualitative research there are really three proven ways to go about it:
- User Behavior
- Customer Surveys
- On-Site Surveys
- Usability Testing
Testing for current website visitors for user behavior is one of the best ways to understand and collect valuable qualitative data. The best way to test for website user behavior is through a software. Two of the most popular website user behavior software are Crazy Egg and Hotjar.
Crazy Egg is one of the first, if not the first, user behavior platforms for developers and marketers. It has three main testing components (1) heat mapping, (2) click mapping, and (3) scroll mapping. Each test allows you to measure interest in different parts of your webpage.
Heat Mapping - allows you to see where web visitors hang out on your webpage the most and what they are looking at based on where they keep the pointer on the screen they control with their mouse. The heat map will light up I the areas people spend the most time and marketers and developers and enhance those areas and improve “cold” areas on the webpage as well.
Click Mapping - counts where people are clicking on the webpage. If people are avoiding clicking on buttons that are tied to your conversion, the click map will confirm this action and show where they are clicking instead.
Scroll Mapping - measures interest in a webpage by seeing how far down visitors will actually scroll down the page. If people do not care to scroll down and read all your information, perhaps you need to reconsider the order and value of the way you are presenting your information on the webpage.
All three of these tests can come in handy for CRO and give you insight as what elements you should test first.
Hotjar is the web 2.0 version or next generation of Crazy Egg. It has a few more features that help out with identifying user behavior on your website. Like Crazy Egg, Hotjar has heat mapping, scroll mapping, and click mapping.
However, it also has a few more premium features:
- Funnel Tracking
- Feedback Polls
- And more…
*I should also note that many of the CRO testing platforms I will mention later have website user behavior testing built into their software as well.
Customer surveys are a great way to collect information from active users on your website that have purchased from you and experienced your website sales funnel.
Through a survey, you can begin to understand what motivated them to make a purchasing decision and what they liked and did not like along that journey.
Here are three proven techniques for a successful survey:
Streamline and Ask Only the Necessary Amount of Questions
Less can be more with surveys. Of course as a marketer, you are going to want as much data as possible. But the truth is, most people do not enjoy surveys. Therefore asking fewer questions that are precise and driving towards gathering information around your end conversion goal typically yields accurate information and more people participating.
To my first point, people generally do not like to fill out surveys on their own accord. Offering an incentive is a surefire way to increase the conversion of people taking your survey. Keep your audience in mind, when incentivizing. Amazon gift cards, discounts off your product or service offering, are all good places to start.
Questions Should Be Open Ended
Yes and no questions do not deliver the true user experience. That’s why subjective responses are necessary for CRO. The details are what matter and create the insightful user data that can be interpreted for hypothesis and CRO testing. Multiple choice will not cut it either, that assumes you know the multiple choice answers are exactly how the user feels, versus giving them the freedom to answer freely through their experience.
Below is a list on the top survey platforms that will allow you to quickly and easily deploy your very own CRO survey:
Here are some questions you may want to add to your survey made up by CRO experts.
- How would you describe this product or service to a co-worker of good friend?
- How many other options did you consider before purchasing this product or service?
- What were the names of those similar products or services?
- Why did you decide to go with our product or service?
- What almost prevented you from signing up / or purchasing?
- What were the questions you had about the product or service?
- How could we have made the answers to those questions more clear or obvious?
- How could we do a better hob persuading your friends and colleges to choose our product or service?
- How do you persuade others when you speak positively about our product of service?
- How does our product and service help you in your everyday life?
- When did you first realize you needed a product or service like ours?
- What is the number one biggest problem our product or service has solved for you?
- What is the biggest flaw in our current product or service?
- What are two adjectives that describe our product or service?
*Note some of the questions may need to be refined and tweaked to fit your offering.
Not all will make perfect sense and therefore you may not want to include them. Also, there are many other questions you can ask that were not covered here.
Another kind of survey is an on-site survey. On-site surveys allow you to ask questions to visitors on your website in real time. This survey allows you to gather information on (1) people who will eventually make a purchase and (2) those who will not.
Focusing in on those who will not purchase from you is a HUGE opportunity.
Understanding what is holding your website customers back from a transaction is critical. Was the price to high, did you not clearly understand our offering, could you not find your way around the website, etc.
These are the questions you can get answered from on-site surveying. The top tools for on-site survey deployment are Qualroo and Google Feedback Surveys.
Here are some expert questions you can ask your visitors about your website:
- What are you finding most confusing on this webpage?
- Is there an element you are looking for and can’t find?
- Do you have any questions at this point?
- What is your biggest concern up until now about purchasing our product or service?
- What is the biggest reason stopping you from making a purchasing decision today?
- What is missing from the current shopping experience?
- Why didn’t you complete your purchase today?
- What was the missing element that would have allowed you to complete your purchase today?
- What is the biggest problem our product or service can help you solve?
- What else can we place on this webpage to convince you to buy?
The purpose of an on-site survey is always to drive towards answers around why people are not making the purchase online. That information will prove to be invaluable when formulating a hypothesis and testing.
In a usability test you provide steps for a user to complete. The user takes your direction and offers feedback on where they got hung up on their user journey.
Now you can begin to decipher where the speed bumps are in your user journey as potential areas for conversion rate optimization.
User testing can be the catalyst to some obvious and quick wins that will help you button up your user journey on the path to your conversion goal.
Here are some of the best tools for user testing you can try:
Who makes a good candidate for a user test?
Your Own Team
Getting critical feedback from your team should always be step one. Not to mention, they are your team driving towards the same goal and an easy resource that is “close to home” to rely on.
Customers are another great place to start for finding user testers. After all they know your brand and product / service offering. Getting customers to take the user test is commonly easier said than done. Consider an incentive to get more participants like a free giveaway of your offering, amazon gift card, etc.
Even users that are not part of your core audience group can be helpful and should not be overlooked. These users can point our glitches, point to messaging issues in communication and understanding and much more.
When conducting a user test you want to come up with a series of tasks.
Here is a quick example of tasking for user testing:
Task 1: Go to Google.com. In the search field search for this “phrase” in relation to one of your product or service offerings. Locate “example website” and click on it. How easy was the website to find in Google search?
Task 2: Go to the home page and look at it for 5 seconds. Then turn away. Write down what you remember about the offering?
Task 3: Find an item you would consider buying. Proceed to checkout with that item (without purchasing) and record your experience in the checkout process.
When evaluating your survey responses, you are going to want to look for three elements:
When you pull all your surveys together, what patterns do you see? Are there multiple people giving critical feedback about your lead form or shopping cart for example. If so, make sure to document those as areas for CRO.
Look carefully at the descriptive words the survey taker used to describe your business. Keep your eyes peeled for the who, what, why, when, where, and how surrounding your business. This will help you (1) understand how the user sees your business and (2) you can repurpose popular user phrases into your CTAs and other important copy, so you are speaking their language.
What are the specific objections people are having? Here you want to focus on the negative and critical feedback. What is the thorn in the user’s side? Identify these specific hurdles the user is having and document them for further evaluation.
Stack rank your survey findings into pillars.
Pillars represent important areas of opportunity people have identified through your surveys to improve. You can represent your pillars in columns on an excel spreadsheet to better visualize your feedback.
Based on the patterns you have identified, you can start to see strategic pillars of your business people care about.
If you are an eCommerce company these pillars may look like:
- Free Shipping:
- Free Returns:
After identifying your pillars. Start to calculate out how many different surveys identified one of those pillars as being most important to them. You can imagine the pillar mentioned the most is a great place to focus your CRO testing.
Selecting a CRO Software
After you formulate your CRO hypothesis, you are going to want to test them. There are many CRO software to choose from, here are three that fit different service levels.
Google Experiments is free and offered as feature within Google analytics. Here you will have to design your variation page and add the Google experiments code to the variation page to be served against the current version of the page or your constant. This is more work, but the program is free. Google Experiments will also (1) measure your results of your tests, (2) reveal the conversion rates between the two pages you are testing and (3) select the statistically significant winning variation.
Virtual Web Optimizer, also known as VWO, is an all-in-one CRO platform. You can start by running visitor user testing with heat maps, click maps, scroll maps and more in order to collect data on how users are currently using your website. Further integrated testing includes on-site surveys to gather even more insights. Then you can record your findings and push right into a test.
The VWO platform is very streamlined and walks you through a CRO process A-Z. It is great for beginner to advanced CRO users.
Optimizely is another enterprise level CRO service. They are the leaders in the space and even have a few more bells and whistles, although it is very comparable to VWO. The service level is a bit higher, however the usability is a bit more advanced and not as intuitive.
Whatever platform you choose, the main objective of a CRO software is to help you arrive at a statistically significant result where one webpage out performs the other webpage in its conversion goal.
When running your CRO tests, there are two main rules you want to try and follow.
- Run your tests till there is a chance of 95% of beating the original. This will insure your conversion is statistically accurate.
- Run your test for at least 7 days. This will make sure even if you have enough conversion data to prove the test is 95% winning, the longer period will allow to iron out any anomalies in the data over time.
Setting Up the Test
The next step in your CRO journey is setting up the test. We will use VWO as our platform because of its ease of use and fast set up options.
First we are going to go into the VWO platform and “create a test.” This is the first thing VWO will prompt you to do.
Next you will want to select the type of CRO test you will be performing. In this example we will start with an “A/B test.”
Enter the URL of the webpage you wish to conduct your A/B test.
Now we can use VWO’s powerful WYSIWYG editor to create a new headline to test, for instance. Here you will not need any code to create the variable page you wish to test against your current webpage. However, if you are a developer you can use HTML and CSS to further develop the test page.
When our variable landing page is designed and ready to deploy for the test we are going to want to set goals. An example goal would be going from the home page to the sign up page. Then you could set up another goal that measures the sign up page to the confirmation page of people actually signing up.
Goal #1: Home Page => Sign Up Page
Goal #2: Sign Up Page => Sign Up Confirmation Page
Note that tracking down to your final conversion goal is really what matters here. In this example we are after increasing the “sign up confirmation page.” However, we want to test people from the home page to the sign up page to make sure that step is not broken in the funnel.
For the last step of the CRO test set up we are going to want to define: (1) the name of our test, (2) notes on what we are testing for, (3) define the percentage of traffic being sent to the webpage, and finally (4) target test to a segment. Below we will further define each one of these inputs.
Name – What you are naming your test to keep track of your tests you are running.
Notes – Here you will keep track of your hypothesis you are measuring in your test.
Percentage of Traffic – This will define the amount of traffic you are sending to the test. You do not need to send all your traffic to the test if you do not want to. However, the faster the test will go the more traffic that is sent to the test. Usually you will want to split test your traffic 50/50 or in the VWO platform run 100% of the traffic to the test.
Target test to a Segment – Here you will be able to set parameters on filtering different types of traffic to your test e.g. mobile traffic, desktop traffic, organic traffic, etc.
After step 5, the VWO platform will prompt you to start the test. Once you start the test half your visitors will see your new variation of your webpage the other half will see the constant or webpage you started with.
Interpreting You CRO Test Results
When examining your CRO test results in VWO you are going to want to click on the “Results” tab and then “Running Tests.” Here Visual Web Optimizer will show you a summary of your existing tests.
The dashboard will show you:
- Total Visitors
- Total Conversions
- Conversion Rate
- Active Test Variations
- Goals Tracking
The “Summary Table” will outline with a red or green bar if the variation is improving conversion (green) or decreasing conversion (red).
To get even more granular you can click on “Detailed Reports.” This section allows you to toggle between goals and see your conversion rate between your two pages you are testing and the timeline in which the test has been running
Below the chart you will find the results for the goal you have selected. It compares the goal vs. the variation alongside the conversion range, conversion rate, percentage improvement, and the percentage chance to beat the control and the conversion and visitors for each test page.
When interpreting the data above we can see the test has been running from January 16 to February 19 (and beyond) for approximately 34 days. The new variation increased Conversion 100%. The variation has a 100% of beating the control page after testing for eleven days. The control is converting at 2.41% and the Variation is converting at 5.86%.
This test was a clear runaway winner. I thought it was fun to show but certainly not always the case. Here we ran the test past seven days and have over a 95% statistical significance and we can declare the variation page the winner!
Now we can hand off the changes to the web development team to make the changes to the webpage permanently.
Final thoughts on CRO
Conversion Rate Optimization has benefits for every website trying to convert more leads or sell more product. With a low risk and high return, CRO is a digital channel I recommend to anyone marketing their website. It does take a little time and brain power, but the cost of the tools are relatively low, and testing your conversion hypotheses can be fun and helps out your bottom line.
If you have questions or thoughts on this article, please participate below.