What is Conversion Rate Optimisation (CRO)?

An Introduction to CRO

Conversion Rate Optimisation (CRO) is a bit like User Experience. Suddenly every marketer is talking about it.

But what is CRO and what will it do for you?

CRO is a collection of different approaches to getting more value from the users of your website, app or platform. Let’s break down its three parts: conversion / rate / optimisation.

Conversion

In the weird world of digital marketing, Conversion is the language we use to describe the actions you want your users to take.

It has nothing to do with St Paul or the Road to Damascus.

Examples of conversions might include:

• clicking a ‘find out more’ button
• entering information on a contact form
• adding items to a basket in an e-commerce store
• actually making a purchase
• upgrading from a free plan to paid plan.

Rate

The rate of conversion is the percentage of a group of users who do the thing you want them to do.

Your conversion rate could be:

• The percentage of visitors to a ‘contact us’ page who actually make an enquiry
• The percentage of people viewing a product who then add it to their basket
• The percentage of total app users who make a purchase

Conversion rates vary massively. Be careful when looking for benchmarks that you always compare apples with apples. 3% of users making a purchase on a fashion website could be a fantastic result; 20% of visitors on a landing page for a B2B white paper could be a bad result. Rates are always relative to the profile of the audience group and the action you want them to take.

Optimisation

Unsurprisingly, optimising the conversion rate is all about increasing the share of users who do the thing you want them to do. Often, this involves making small changes that ‘nudge’ people in the right direction: moving the position of a button, time-limiting a special offer, making your language clearer.

Bringing it back together

So CRO is the art and science of making improvements that optimise conversion rates. It involves constant iterative experiments – testing and learning – with changes to design elements – copy, layout, colour palette, font etc – so that more users convert and your marketing spend gets more efficient.

Often, these experimental changes can seem incredibly small, and their impact can be tiny in percentage terms – but dramatic in revenue terms when a user base is very large.

Google’s $200m question

Google – unsurprisingly – was one of the early pioneers in this field of data-driven experimentation.

In 2009, there was lively debate between Google’s graphic designers and product engineers (led by Marissa Mayer, who subsequently left to be CEO of Yahoo!). The debate was about the specific shade of blue to use for hyperlinks in adverts placed inside Gmail and alongside Search results.

The engineers’ method won they day. They picked 41 different shades of blue, and then ran a series of ‘1% tests’: experiments that showed a different variant of blue to 41 groups of users, each representing 1% of the base.

The results showed a small uplift in clickthrough rates for the most preferred shade.

But that small increase equated to $200m additional revenue for Google.

We can’t guarantee that CRO can do this for your business – but in the rest of this article we’ll give you a jargon-free guide to the basics of CRO and show you some tools you can use if you want to try it for yourself.

---

How to do CRO

Smart Insights provides an useful diagram explaining the CRO process. 

Let’s look at each part of the process in detail. 

Analyse 

Evaluate and audit your data e.g. Google Analytics data first. Ask what you’re web visitors are doing, where are they dropping off. What’s the typical journey for them? 

Research 

Qualitative and quantitative research as this stage including user testing would be ideal. 

Our user testing guide provides insight on what it is and how to do it. It also details how to run website surveys, website visitor recordings and heatmaps. 

This is the stage to truly understand what your visitors are doing. 

Generate insights 

Produce insights from the data. Be clear on where the barriers are on your site. 

Hypothesis development 

Create hypotheses for example visitors are doing X, if we do Y they will do Z. 

Experiment 

Run your AB tests until they are statistically significant i.e. you have enough traffic which has experienced the test pages to know if it’s a beneficial change or not. 

Repeat 

CRO is an ongoing process. Onto the next test or further research! 

---

A Practical Example

The first step in any CRO effort is a data audit to understand what your web visitors or app users are doing. Armed with data about your current situation, you can then create a series of hypotheses about what improve your conversion rates – and then test those hypotheses to see if they make a positive difference. 

Let’s imagine you have a website whose main call-to-action (CTA) is to get users to make an online enquiry.

But you see that many visitors to your ‘contact us’ page don’t actually fill out the enquiry form. You assume they want to contact you because they navigate to that page – but only 1% of visitors are actually entering the details you ask for in your contact form. 

From this, you could produce a series of hypotheses and tests for example: 

Situation: the form has 12 fields, all of which are mandatory. 

Traffic: 20% of all visitors to the site go to the ‘contact us’ page. Only 1% of that group are converting.  

Hypothesis: the form is too long. 

Now you decide what sort of test to run: A/B Test or Multivariate test (MVT).

A/B Testing

A/B Testing is a binary comparison: it compares one variant against your original page design.

In this example, your test might involve reducing the total number of fields from 12 to 5.

Visitors to your page would be randomly assigned to one of two groups. One group would see the original 12-field design; the other would see the variant 5-field design.

Once you have a statistically large enough sample to identify any differences between the two versions, you either change to the variant (if it performs better) or stick with the original.

Multivariate testing (MVT)

Multivariate tests compare several variants simultaneously alongside the original. This works well for high traffic websites or apps with a large user base.

In this example, you might design three different tests for your hypotheses:

Test 1: Reduce the number of required fields 

Test 2: Reduce the number of total fields 

Test 3: Justify why you’re asking for certain pieces of information e.g. why you need their postcode.

Visitors to the ‘contact us’ page would then be randomly assigned to one of four groups – one for each variant plus the original (control) group. The best performing variant – once you have enough data to be confident – wins.

This all sounds complex to design and manage, but there are many tools that take care of splitting the samples, displaying the variants and even switching to the winning design once the right thresholds are met. Google Optimize, ABTasty, Optimizely, Instapage and other CRO tools can help with this.  

You can read about examples and case studies of A/B and Multivariate tests on the Smashing Magazine website, including this one about simplifying the multi-step checkout in the Just Eat app:

Uses and Benefits of CRO

What can you test with CRO?

CRO using A/B and Multivariate tests can be applied to almost any website element: 

• Landing pages including landing page copy, headline, 
• Website copy across the site 
• Checkout buttons, basket and checkout process. 
• Call-to-action (CTA) buttons across the site 
• Forms 
• Website navigation and structure 
• Upsells e.g. ‘you may also be interested in XYZ’ on appropriate pages. 

What is the impact from CRO?

There are many benefits of CRO including:

• Increased revenue
• Increased average order value 
• Increased revenue per person 
• Efficient marketing spend – make the most of your marketing spend 
• Website changes rather than marketing changes are faster to implement.

Remember, CRO is an ongoing process. Revenue increases are never guaranteed, but CRO can help you make the most of your website and increase revenue from your existing visitors.

A/B and Multivariate testing also only works if you have enough traffic to get a statistically significant result. If you don’t have large traffic volumes, other approaches to CRO can include UX reviews and user testing. 

UX (User Experience) reviews of your site by a conversion expert will help to pinpoint barriers and identify changes which can be made in line with best practice. 

User testing involves getting feedback from users on your website. The insights will inform website changes which can lead to improved conversions. We have a whole guide for you on user testing here. 

CRO Platforms for A/B and Multivariate Testing

Explore the A/B and Multivariate testing tools in the Insight Platforms directory: 

CRO Platforms User Testing

Explore the User Testing tools in the Insight Platforms directory: 

Final Word 

CRO works really well with a UX focus, it helps you to optimise a site to user needs.

By understanding your users and by tailoring your site to them as much as possible you can make it easier for users to achieve their goals on your website.