What is Conversion Rate Optimization?

Chapter 8

How to Collect Data for CRO

We’ve talked at length about optimizing the entire marketing funnel. Discussed the techniques to use with your ads, website, post-click landing pages, thank you pages, and remarketing campaigns to make it easy for visitors to take the desired action at each step.

However, conversion optimization isn’t only about a list of practices; not every website or post-click landing page simply needs a bigger, brightly colored CTA button to get more conversions. Knowing the best optimization techniques to use in your funnel helps in creating a more action-oriented funnel. But, finding out and solving your particular funnel’s conversion problems is what makes your own funnel optimized.

In this chapter we’re going to focus on the second part of conversion rate optimization — that is collecting and testing data to improve your marketing funnel conversions.

Let’s look at the techniques you can use to gather data from your marketing funnel:

Heuristic Analysis

Heuristic analysis is an experience-based assessment of a web page, it can be done relatively quickly and involves reviewing the page to identify problems with usability and the user interface (UI) design.

There are three main frameworks most widely used in heuristic analysis: The Neilsen Norman Group’s Heuristic evaluation, Marketing Experiments’ Model, and the LIFT model.

Nielsen Norman Heuristic Analysis

The NN/g has highlighted ten steps that need to be considered for heuristic analysis:

1. Visibility of System Status:

The visitor shouldn’t feel lost on a web page; they should be informed about the next step they should take with the help of navigation and user onboarding messages etc. Call-to-action button copy that explains to the visitor what they’ll get once they click the button is an example of this step.

2. The match between system and real world:

The copy on the web page should reflect the language that visitors use and are familiar with. The information on the web page should also follow a logical order.

3. User control and freedom:

The visitor should always be able to easily exit the web page if they feel the need to. Popup screens that hijack the entire screen and don’t let you exit is a perfect example of what not to do on your website.

4. Consistency and standards:

A web page should always maintain consistency and standards, i.e. not use same words for different situations, because this could confuse visitors.

5. Error Prevention:

Prevent errors on your web page. However, you should also prepare to notify visitors accordingly when they do occur. For example, if your page server is down, you should notify visitors via a message on the web page.

6. Recognition rather than recall:

All instructions of your web page should always be visible to the visitor. They shouldn’t have to remember information they want to access. A sticky navigation bar that scrolls with the visitor promotes the principle of recognition rather than recall.

7. Flexibility and efficiency of use:

Make it easy for visitors to perform frequent actions, with the help of personalized CTAs and tailoring the experience for them. An example could be showing customers customized in-app messaging, such as using their first names, etc.

8. Aesthetic and minimalistic design:

Keep things simple when it comes to web page design, if an element doesn’t add to user experience there’s no benefit in using it. Don’t clutter web pages with unnecessary elements just because they may look good.

9. Help users recognize, diagnose, and recover errors:

In case a page error occurs, it’s best to help users identify the error by telling them in plain language as opposed to code what has happened and what they need to do to fix it. Designing a 404 page is an example of this step.

10. Help and documentation:

It is necessary to provide visitors with help and documentation to improve their web page experience. A knowledge base and an FAQ section are examples of adding help and documentation on web pages.

Marketing Experiments Model

The marketing experiments heuristic model is represented by the following equation:

C = 4m + 3v + 2(i-f) – 2a


  • C: Probability of conversion
  • m: Motivation of user (when)
  • v: Clarity of the value proposition (why)
  • i: Incentive to take action
  • f: Friction elements of process
  • a: Anxiety about entering information

Essentially, the formula dictates that the probability of a conversion depends on the match between the offer and visitor motivation plus the clarity of value proposition plus the incentives to take the action, subtracting anxiety from the equation.

LIFT Model

The LIFT Model, introduced by WiderFunnel is nearly similar to the marketing Experiments model, but contrary to an equation, the model is explained by a graphic:

The LIFT model focuses on six conversion factors to identify issues and opportunities for testing web page hypotheses and design.

The main thing to remember about heuristic analysis is that it doesn’t tell you the solution to a problem. It doesn’t even tell you how important the problem is. What heuristic analysis does is help you understand the overall layout of your webpage.

After performing heuristic analysis, ideally you’ll want to conduct other tests to improve the landscape of your web page.

Web Analytics Analysis

Web analytics gives you answers to a whole array of conversion questions, including:

  1. What visitors are doing once they arrive at your website and post-click landing page
  2. The impact and performance of every page element, feature, and widget
  3. Where your website and post-click landing pages are losing money

Some of the most popular web analytics tools on the market are Google Analytics, Adobe Analytics and Quantcast.

With web analytics you gain an insight into funnel visualization, learning how much traffic is coming at each stage of the funnel. You can also find out where you’re failing to convert visitors and understanding which parts of your website or post-click landing pages need fixing.

Analytics is also a good way of checking if your website is browser-compatible. Check if you’re losing out on conversions on a specific browser and if you are, it’s time to analyze your site’s UX from that browser.

With analytics you can collect and calculate data for a wide range of metrics, such as:

  • Demographics
  • Interests
  • Geo
  • Behavior
  • Bounce rates
  • Page load speeds
  • Conversion value
  • User flows
  • Traffic sources

Before you begin analyzing data it’s important that you set up proper key performance indicators (KPI’s) so that you can measure how successful you were at achieving your marketing goals.

Where Heuristic analysis gives the answer to the question “why,” web analytics analysis tells you about the “what,” “where,” and “how.”

Visitor Behavior Analysis

With visitor behavior analysis you can collect data on what visitors do with their mouse when they’re on your web pages. This allows you to understand if they are having any problems with any specific page elements — such as clicking something that’s not clickable and changing those elements accordingly.

To track visitor behavior, heat maps and user session replays can help you.

Heat Maps

Heat maps are visual representation of data and these help you record what visitors are doing with their mouse or trackpad when they’re on your web pages.

The heat map can be in monochrome, with the values ranging in black and white. However, heat maps are mostly designed in the following five color gradients from lowest/coldest to highest/warmest:

  • Blue
  • Cyan
  • Green
  • Yellow
  • Red

To ensure that the data you collect from heat maps is accurate, it’s important that you have an ample sample size before you start generalizing results and making changes to your web pages.

It’s recommended you have at least have 2,000-3,000 page views per screen, and per device (desktop and mobile) before you start generating heat maps. That’s because heat maps won’t be as beneficial and uncover as much for you if your page has low traffic.

Changing website elements based on data collected from a heat map with very little traffic and data isn’t going to help you make any optimization decisions.

There are different types of heat maps that you can use.

Hover map

Hover maps record visitors’ mouse movement and tell you where visitors have hovered when interacting with your website and post-click landing pages.

This is what a hover map looks like. In this example, the user mostly hovered over the text fields with percentages:

Although hover maps provide additional insight into your visitors’ cursor movements, they are considered a poor version of eye tracking, because the accuracy of mouse tracking studies is questionable. presentation by Google’s Senior User Experience Researcher, only 6% of people showed some positive correlation between mouse movement and eye tracking. Furthermore, 10% of visitors hovered over a link and then continued to read different things on the page.

Just because a user’s cursor is in one spot, that does not mean that’s where they’re looking. For example, maybe they’re having trouble locating the CTA button while their cursor sits on the headline. And research supports this, too.

Although hover maps can be used to record user behavior on your website and post-click landing pages, there are some concerns as to the accuracy of their data. That’s where other heat maps can pick up the slack, such as click maps.

Click Maps

Just like their name suggests, click maps show you a visual representation of aggregated click data on any given web page. The brighter the color, the more “clicks” that particular area received:

Click maps explain which page element is attracting the most attention, and which element is a cause for concern for visitors. For example, if visitors are clicking a certain element on the page that isn’t even clickable, a click map would show you that data. Then, you can make that particular element clickable while optimizing the user’s experience.

Attention Maps

Attention maps help you determine which areas of web pages are viewed the most by the user taking into consideration the horizontal and vertical scrolling activity.

Attention maps are best summarized by Peep Laja at ConversionXL:

Attention maps are useful because they take into account different screen sizes and resolutions, and shows which part of the page has been viewed the most within the user’s browser. Understanding attention can help you assess the effectiveness of the page design, especially above the fold area.

Scroll Maps

Scroll maps show you how far visitors scroll on your web pages. And since not everybody likes to scroll, a scroll map will generally look darker the further down the page you go:

Scroll maps are beneficial for collecting user behavior data on long-form sales pages as you can find out at what point visitors usually stop scrolling the page. Once you determine that particular point, you can make that section of the page more engaging (like inserting a CTA button, for example).

You can generate a variety of heat maps for your website and post-click landing pages using tools such as Crazy Egg, Inspectlet, VWO and Hotjar.

User Session Replays

In contrast to heat maps, user session replays don’t require any color interpretation. Instead, user session replays (or user recordings) give you a video account of what visitors did on your website and post-click landing pages. Simply watch a video of how visitors are interacting with your web pages. All the data you collect with user videos is qualitative data.

User session replays help you identify any usability and optimization issues that may be affecting your websites and post-click landing pages. You can find out which elements your visitors are responding well to, and which elements that are still not being interacted with. For example, if user recordings show visitors exiting the web page after they attempt to complete the form, it’s probably best you look at optimizing your form.

Some tools you can use to generate user session replays are Inspectlet, Hotjar and Mouseflow.

A/B Testing

A/B testing allows you to create, measure, and then test variations of your web pages to determine which combination of elements on a page are more successful at converting visitors.

You can use A/B testing to improve conversions on your website, post-click landing pages, and marketing emails.

Before you begin A/B testing, though, it’s important that you have enough traffic on your web pages. Starting out, for best results your pages should at least have 350-400 conversions per variation. That way, your results can be generalized and you can optimize the elements that need to be changed.

All A/B tests start with the original version of your post-click landing page, this is also called the “control” variation. Once you have a control page, there are four steps to follow when conducting A/B tests:

1. Set Conversion Goals: The conversion goal you set depends on the campaign you’re running at the moment. For example, if you’re setting goals for your ebook post-click landing pages the conversion goal would be ebook downloads.

2. Create Variations: You can create page variations simply by changing one element in the original variation. For example, if your control page has an on-page form, your variation could have a two-step opt-in form.

3. Start Testing: When you begin testing, the testing platform you use will randomly send equal amounts of traffic to both variations so that the results generated are more accurate.

4. Analyze Results: After your test reaches 95% statistical significance, an average A/B test takes approximately 4 weeks to run. Only then can you begin to analyze the results. Keep in mind, the results you get also depend on the A/B testing platform you’re using because different tools test for different metrics. Most tools calculate metrics such as unique visitors and conversion rates.

Some popular A/B testing platforms are Optimizely and Visual Website Optimizer. However, we recommend Instapage’s advanced A/B testing analytics because it simplifies the testing process. It allows you to quickly and easily create variations with our fully customizable builder. You also have the option to duplicate, transfer, and delete any page variation you want at any time.

Not only that, but all your A/B testing results are displayed in an easy-to-understand analytics dashboard, which shows the following metrics:

1. Unique Visitors: the number of unique visitors that have viewed a particular post-click landing page variation.
2. Number of Conversions: the number of visitors who have completed the post-click landing page goal.
3. Conversion Rate: the percentage of visitors that completed the post-click landing page goal on a particular page variation.
4. Improvement: the difference between the conversion rate tested against the control version and variation B.

To see how easy it is to create and A/B test post-click landing pages in Instapage, follow the next steps.

Click “Create an A/B Test” and then click “Variation A” and “Duplicate.” Once that’s done, you’re all set to make changes to the original page:

You have the option to duplicate, transfer and delete your variation:

Duplicate: copy a specific variation when creating a new test
Rename: write in a new name for a specific variation
Transfer: transfer the variation to another post-click landing page
Delete: remove a variation completely (all stats accumulated will also be deleted)

You can always monitor the results of your tests in the analytics dashboard. Remember though, it’s best to wait and make conclusions from your results after your A/B test has reached 95% statistical significance.

Collecting and analyzing data is an important part of conversion rate optimization because with data you’re able to see exactly which element of your website and post-click landing page is underperforming. Only then can you understand what needs to be fixed to create a more optimized marketing funnel.

Choose a Chapter

Get Started Now

Try the most powerful all-in-one landing page solution today. Start a 14-day free trial, no credit card required, or schedule a demo to learn more about the Enterprise plan.