Chapter 2: Your First Steps in CRO
In the first chapter, we covered the basics of CRO, how it works, and how businesses can benefit from it.
By now, you are probably ready to dive-right-in and run your first conversion rate optimization campaign. Before you can do that, however, there are a few things you need to know about what CRO is not, what to expect from it, and the process for running a CRO campaign.
Section I: What CRO is NOT
If you’ve gone through the first chapter, you already have a good idea of what CRO is all about.
However, there are still a number of misconceptions about what CRO is and what it can do for you.
In this section, we’ll help you understand what CRO is not.
1. CRO is NOT a List of “Best Practices”
If you’ve spent time on any marketing blog, you’ve likely come across articles on “CRO Best Practices” backed by isolated case-studies. These case studies reflect how making certain standard changes, such as changing the CTA button color from red to green or increasing the font-size of headlines leads to dramatic improvements in conversion rates.
This can give rise to the impression that CRO is essentially a list of “best practices” and if you implement them, your conversion rates will be better than your peers.
The truth is that while there are certain guidelines all websites must follow for better conversions (making the CTA more prominent and maintaining visual hierarchy, for instance), there are no fixed best practices for getting the results you want.
For example, as per color psychology, a color like green is easier to process visually and helps the viewer relax. It is also associated with “go”, thanks to its use in traffic lights.
Red, on the other hand, is frequently associated with stopping because of its widespread use in warnings, road signs and of course, traffic lights.
In fact, some case studies even show that green buttons convert better than others. In one test, switching from a blue to a green button bumped up conversions by 35.81%.
Based on this and the negative associations we have with the color red might make you believe that using a green button is a “best practice”.
Yet, in another test, a red button outperformed a green button by 21%.
This disproves any best practice you might have heard about button color.
However, if you look at both the tests shown above, you’ll notice that what matters more than button color is whether the button actually stands out on the page.
Test #1: The original blue button is similar to the blue font. In contrast, the green button stands out.
Test #2: The original green button is similar to the green in the Performable logo. The red stands out in comparison.
The general principle you can draw from this is: a CTA that stands out on a page draws more clicks, regardless of exact color.
This is essentially how you must view all “best practices” - as indications of a general principle, not an actual result you need to follow through.
Don’t follow best practices. Instead, focus on diving deeper to understand the results from different tests and draw out general principles from it.
2. CRO is NOT a Magical Bullet
Looking at the stupendous results some businesses manage to achieve - 60% more app downloads, 25% more leads - you can be misled into believing that CRO is a one-size fits all solution to all your marketing problems.
In truth, while CRO is incredibly useful, it is only one part of your marketing, sales, product customer success strategy. Without the others, no amount of conversion rate optimization will yield results.
Consider an example. Suppose you sell a $10 product and get 10,000 visitors to your site every month. Your conversion rate for this traffic is 2%. Thus, your monthly revenue is:
2% * 10,000 * $10 = $2,000.
Now if you wanted to increase this revenue to $10,000, here’s all that you could do:
Traffic-focused solution: Increase traffic to 50,000 visitors/month while maintaining the same conversion rate (2%).
Product-focused solution: Increase product price to $50 with the same conversion rate and traffic.
Conversion-focused solution: Improve conversion rate to 10% while keeping price and traffic the same.
Depending on your core competence, budget, and existing marketing strengths, you might opt for any one of these solutions. For instance, a business that already excels at generating traffic might focus on getting more qualified visitors to its site. A business that can use upsells, on the other hand, might use a product-focused approach to get 4x more revenues.
For most businesses, the actual approach will likely involve all three solutions above:
Increasing traffic to 20,000 visitors/month.
Increasing product price to $20.
Improving conversion rate to 2.5%.
Understand this before you start any CRO campaign: your results will often depend on other parts of your business. Without traffic or a product worth buying, no amount of CRO wizardry will help.
CRO is only one part of the marketing puzzle. You still need quality traffic, a product people want to buy, and customer service to back it up.
3. CRO is NOT Quick
Often, businesses start a CRO campaign with a lot of enthusiasm only to end it after a couple of weeks without results.
A well-executed CRO campaign will require two things:
Traffic to test out different variations.
Time to divert enough traffic to each variation.
It is only once you have both these elements can you declare a particular variant as “better” with statistical confidence.
Also understand that the more traffic you have going to a test, the faster you will get results. A page with just 10 visitors/day will require at least 20+ days to divert enough traffic (100 visitors per page variation) to each variant.
However, if you have 1,000 visitors/day, you can rapidly test out different page variants and get results almost immediately.
Here’s a graph illustrating this time-traffic relationship:
Also understand that the more complex your website is and the more elements you want to test, the longer it will take to get results.
A blogger who wants to test the headlines on his blog posts can easily see results within days. On the other hand, a large eCommerce store with hundreds of elements to test - product descriptions, CTA buttons, navigation style, etc. - will likely take weeks, if not months to get results.
Ensure that you have enough time to devote to a CRO campaign. Do not expect immediate results. At the same time, prioritize your testing to get big wins first (more on this below).
4. CRO is NOT Consistent Across Websites and Verticals
This is one of the fundamentals of CRO: every niche, every website is different. CRO tactics that work on one site will not necessarily work for another.
For example, when retailer StacksAndStacks.com added videos to its product pages, it discovered that customers who viewed a video were 144% more likely to buy than those who didn’t.
From a marketer’s perspective, this makes clear sense: videos show a product as it actually is, as opposed to a static image.
However, another test showed otherwise. Brooksdale Living, a community home for seniors, tested two variations of a landing page. One version used a static image:
The other version used a video above the fold instead:
Even though you’d expect the video version to win out, test results showed that it was the opposite. The image version converted better, adding $106,000 to the company’s bottom line.
While the reason for these results could be many (for example, Brooksdale Living’s target audience of seniors might not have fast connections for video), the point is that you can’t assume anything with CRO. Proven practices that work on one website might completely falter on another.
The only solution is to create variants based on sound principles and to test them out without assumptions.
CRO results are seldom uniform across websites and verticals. Results from one site don’t always apply to another. Instead of making assumptions, test out different page variations and rely on data to make decisions.
With this wrapped up, let us now look at what a conversion rate optimization process actually looks like.
Section II: The Conversion Rate Optimization Process
As stated previously, CRO is an ongoing process.
It is not something you conduct a few times and then forget about when your goal is achieved. Your prospects’ desires are constantly shifting and new roadblocks can arise suddenly. You need to be able to respond to these changes quickly to ensure your conversions match your expectations.
As long as your website is live, you should be constantly testing and finding ways to refine and optimize for best user experience.
You can broadly divide the CRO process into three distinct stages:
Discovery: This is when you understand your audience through qualitative and quantitative data and formulate an optimization plan.
Testing: The actual experimentation phase when you test out different hypotheses you brainstormed in the above step.
Review & Analysis: The post-experimentation phase where you analyze results from different tests and make changes accordingly.
We will cover each of these stages in more detail in later chapters. For now, let’s look at the four things you will need to implement a CRO campaign.
5 Things You Need for a CRO Campaign
To achieve your goal of maximizing conversions, you need to be able to:
Come up with solid hypotheses based on qualitative and quantitative data that will move your business goals.
Create page variations based on the above hypotheses.
Run a test long enough to gather actionable insights.
Analyze test results and implement changes accordingly.
You can’t simply blindly optimize pages without understanding what your audience actually cares about. You have to find your “must-have experience” which turns passive visitors into valuable conversions.
Once identified, you can then embark on designing the shortest path towards that experience which reveals the “aha moment” for your visitors - the CRO process in a nutshell.
For an effective CRO campaign, you’ll need these four things to be successful:
1. Customer Insights
What your audience wants, what pages it lands on, what hurdles it faces in completing a purchase - these are critical questions to answer in the CRO process.
This is also the first thing you need to run a competent CRO campaign. To jump blindly into a campaign without thoroughly understanding your audience would be a waste of time and resources.
There are two types of insights that you will need to gather and analyze to be able to determine what changes need to be made.
Quantitative insights: Quantitative insights help you answer questions like about the what, where, and how much.
Qualitative insights: Qualitative insights answer the why behind the above.
We’ll explore both these two data types in-depth in Chapter #3. For now, let’s go over them briefly so you better understand what they involve.
The goal of analyzing quantitative insights is to figure where losses are occurring in your conversion funnel. By using analytics data, heatmaps, and session playbacks you can monitor what exactly visitors do on your site.
Quantitative data can help you discover insights such as:
How long do visitors spend on a page, where they come from, and where do they leave from.
What pages are most popular with your visitors and what do they do there.
While quantitative data can reveal where problems are occurring on your pages, they do not reveal why they are occurring. At best, you can hypothesize solutions with insights from quantitative data.
Think of this as a doctor checking your pulse and taking your temperature. By itself, these measurements don’t tell the doctor what you might be suffering from, but they do help him in figuring out whether something is wrong, and what might be the cause behind it.
For more insights, the doctor has to ask you pointed questions, look at your past medical history and draw out hypotheses from it.
In CRO, this would be the equivalent of qualitative data.
The goal of qualitative insights is to gather an in-depth understanding of user behavior and the reason behind that behavior.
Qualitative data allows you to peek inside the mind of your prospects via a direct communication channel with your visitors. Essentially, qualitative insights attempt to find out the why behind a what or a how.
With qualitative insights, you’ll answer questions like:
Why are people clicking away from a sign-up form?
Why are 90% of my buyers abandoning their carts?
What do customers actually want when they land on a page?
Here are two of the most popular methods for gathering qualitative data:
Interviews: Talking to your customers directly is the best way to discover the gap between your marketing message and what your customers would prefer to hear. Interviews can reveal fundamental flaws in your value proposition.
Customer surveys: Gathering feedback on a large scale from your visitors reveals the kind of traffic you attract, their desires, the type of language they use, and the doubts they face. This insights can reduce points of friction in your conversion process.
Not every idea you test is going to be successful.
While insights are crucial, you can’t really run a test without traffic.
2. Quality Traffic
If you are testing elements on your site, you naturally want to be able to quickly identify which variation leads to the metric you want.
To be able to analyze statistically significant results in a timely manner, your site needs to have at least a few thousand unique visitors per week.
Keep in mind that if you begin testing several variations of your page at once, the amount of traffic required will rise significantly to be able to achieve in the desired time.
What if you don’t have this kind of traffic?
Not to worry, you can still carry out A/B testing by buying traffic. Factor the traffic acquisition cost into your CRO campaign if you’ll be taking this path.
A Note About Traffic Quality and the Sustainability of Results
The quality of your traffic is a massive factor in the success of your CRO campaigns. After all, traffic is the starting point in the entire conversion funnel. If the traffic quality is low, no amount of optimization will help you grow conversions.
Many businesses fail to take this into account before starting a CRO campaign. A business that gets high-quality organic traffic through a target long-tail keyword (such as “buy auto insurance in new york”) will naturally have higher conversion rates than one that gets generic traffic through social media, even with a poorer design.
This is why you must always tailor your conversion goals in accordance with your traffic source. If you get a lot of “window shoppers” from social media, expecting them to buy something at the first go would be foolhardy. Similarly, if your traffic comes from targeted organic keywords with high commercial intent, hiding away your money-making offers won’t make much sense.
Broadly speaking, your conversion goals should be as follows in relation to your traffic:
Untargeted Traffic (social media, random referrals, and type-ins) -> “Soft” selling by capturing emails/phone numbers or engagement.
Targeted traffic (PPC, targeted organic keywords) -> “Hard” selling through discounts, offers, etc.
If you are buying traffic to run to your tests, also keep in mind whether you can sustain a similar traffic quality once the test is over. Else your results might not be the same once the traffic tap runs dry.
3. Testing Tools
There has been a revolution of sorts in the conversion optimization industry in the last few years, led mostly by the commodification of testing tools. Everything from gathering customer insight to running split-tests is now within the reach of average businesses.
Testing tools generally fall into two broad categories:
All-in-one tools: These are dedicated tools for running CRO campaigns and combine several tools in one.
Standalone tools: These are usually single-function tools that help you accomplish a specific task in the CRO process (such as creating surveys or running a split test).
Let’s look at each of these in more detail.
These are specialized conversion rate optimization tools designed specifically for running CRO tests. While the actual features might vary, these tools usually have at least:
A way to gather qualitative insight using surveys.
A way to run A/B and/or multivariate tests.
A way to analyze results from tests and segment audiences.
A great example of this would be Omniconvert, which has built-in tools for conducting surveys, running tests, segmenting audiences and showing them personalized pages.
Of course, you might still need additional tools to complement this. For example, if you don’t have in-house design talent, you might want to purchase a landing page creation tool (though Omniconvert also has a visual editor to help you make changes yourself). You will also need a way to capture quantitative data such as Google Analytics.
All-in-one tools have several advantages and disadvantages:
Easier to use: A single software means that all features integrate nicely with each other. You don’t have to open multiple tools to gather data.
Cheaper: Instead of paying for dozens of tools, you just have to pay for one, which can result in substantial cost savings.
Faster: You don’t have to worry about importing/exporting data to and from multiple tools, which can result in a faster and more streamlined testing process.
Fewer features: Since an all-in-one tool has several functions rolled into one, each function (such as conducting surveys) might not be as feature-rich as a standalone tool.
Missing functions: Because of the sheer breadth of the conversion optimization process, any all-in-one tool will invariably have a few missing functions.
These are usually single-function tools that help you accomplish a specific task in the CRO process. Often, you can integrate these tools with each other (such as a landing page tool that integrates with an email marketing tool) to run a CRO campaign.
For example, you can use a tool like Typeform to conduct user surveys:
You can complement this subjective insight with data gathered from Google Analytics.
With this data, you can come up with a hypothesis. You can then use LeadPages to create landing page variants.
You can then split traffic to these pages using Omniconvert, and so on!
Standalone tools vary greatly in complexity, features and cost. Some of these might be useful outside a CRO context as well (such as using Typeform to gather customer insight or using Google Analytics for general analytics).
Some of the advantages/disadvantages of standalone CRO tools are:
More choice: Instead of relying on an all-in-one tool’s built-in features, you have the option of buying tools based on your own requirements.
Better features: Since most standalone tools do only one thing (such as creating landing pages), they often have better features.
Useful outside CRO: Many standalone tools will be useful outside of CRO as well.
Higher costs: It’s not unusual for standalone tools to run into hundreds of dollars - substantially more expensive than a single all-in-one tool.
Less streamlined: You don’t have a single dashboard to gather data and run tests, which can make it difficult to run a streamlined campaign.
Integration issues: With an all-in-one tool, you can be sure that every feature will integrate with another. With standalone tools, this might not always be the case. There might be no way to integrate your landing page tool with your split-testing tool, for instance.
Whether you choose a standalone or an all-in-one CRO tool will depend largely on your preferences, budget, and marketing needs. There is no one-size-fits all; it’s entirely a matter of individual requirements.
4. A Testing Framework
Most successful CRO campaigns follow a testing framework. This testing framework is essentially a distilled, actionable version of the CRO process we covered above. The exact framework you use will depend on your goals, your business and your existing expertise, but more often than not, it follows this structure:
Step #1: Identifying your business, marketing and conversion goals.
Step #2: Identifying the reasons why you are unable to meet these goals.
Step #3: Coming up with a hypothesis to solve the problem(s) identified in step #2 above.
Step #4: Creating page variants and running split tests.
Let’s walk you through an example to help you understand this framework better.
A Hypothetical Testing Framework
Suppose you are running an online store that sells shoes. You define your business goal as follows:
Business goal: Increase sales of shoes on site by 100%
You now need to define how marketing will help you reach this goal. Thus you say:
Marketing goal: Increase conversion rate from 1% to 2%.
After defining the marketing, you need to define how exactly you will meet this marketing goal. You might have something like this:
Conversion goal: Add larger images, improve checkout process and reduce cart abandonment to increase conversion rate.
Keep in mind that at this stage, your conversion goal is based largely on rudimentary data and conjecture. The actual conversion goal might change depending on the existing and fresh data you capture.
After defining these three goals, you can focus on identifying the reasons why you can’t meet your conversion goals. Once you’ve gathered and analyzed your data (qualitative and quantitative), you might recognize your conversion impediments as:
Customers can’t figure out what the product actually looks like.
High shopping cart abandonment.
Stray clicks during checkout process impede conversions.
With this, you can come up with different hypotheses to solve the above problems, such as:
Use large images to show off the product and its craftsmanship.
Offer free shipping to cut down on shopping cart abandonment.
Remove navigation and all other external clicks from the checkout process to prevent stray clicks.
After doing this, you can create your own page variants and start testing these hypotheses.
As mentioned before, CRO is a long-term process. Only 14% of A/B tests deliver a winning result. For sites with low traffic, this often means that you might spend weeks testing different variants only to get negative results.
A successful CRO campaign, therefore, requires a lot of patience. You should be willing to devote anywhere from 3-6 months before you see a marked improvement in results.
However, just because a particular test failed doesn’t mean that it wasn’t useful. There is real value in every result, good or bad as you can learn insights that may have been previously unknown to you.
For example, if you’re testing out different CTA button colors, you might realize that your existing color converts better than any tested variants. This doesn’t mean that the time spent testing was a failure; it means that either:
Your existing color stood out on the page and didn’t need to be changed.
Your tested variants were not remarkable enough to improve upon existing colors.
Here are some potential things to ask yourself after failed results:
Did you test based on intuition or on actionable insights from research?
Are you blindly following best practices advised on the internet?
Do the metrics you are testing match your business goals?
Was the variation tested bold enough to be noticed by visitors?
Did you end the tests too soon?
Review your tests and ask yourself these questions and redesign your campaign.
This wraps up chapter 2. In the next chapter, we’ll look at how to gather data for your tests.
CRO does not yield quick results. It is also not a magical bullet to all your marketing woes.
CRO is NOT a list of best practices, nor is it consistent across websites and industries.
Every successful CRO campaign follows a process.
If you want to run a successful CRO campaign, you have to start by gathering insights into your audience.
You also need traffic, high-quality tools and a proven testing framework.
Lastly, you need patience; a CRO campaign might not yield results for weeks, if not months.
Go back to the Table of Contents.