IES606

MN-389-E
March 2016

Conversion Rate Optimization

Introduction
Since the advent of e-commerce, many tools have been developed to drive this new sales
process. The revenue of a company competing in digital business still depends on traditional
marketing variables, such as the quality and pricing of the products it sells, but in the digital
world new techniques associated with recently created tools have made it possible to influence
variables such as the conversion rate in real time.

For example, when a company’s management committee wants to boost sales in its e-
commerce channel, it can choose between two alternative approaches. The first is to increase
visits while maintaining the same conversion rate. The second is to increase the conversion
rate, taking it as given that the number of visits isn’t going to change. A combination of both
approaches would of course be ideal; however, for illustrative purposes we will focus on these
two alternatives.
Table 1

2015
Visits Conversion Sales Average order value Revenue
1,000,000 1% 10,000 €120 €1,200,000
2016
Visits Conversion Sales Average order value Revenue
1,500,000  1% 15,000 €120 €1,800,000

This technical note was prepared by Luis Ferrándiz, partner at KPMG and CEO of ADN, Pilar Soldado, research assistant,
and Professor Julián Villanueva. March 2016.
All of the material contained in this document has been developed by the author unless otherwise stated.

Copyright © 2016 IESE. This translation copyright © 2016 IESE. To order copies contact IESE Publishing via www.iesep.com.
Alternatively, write to [email protected] or call +34 932 536 558.
No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form
or by any means – electronic, mechanical, photocopying, recording, or otherwise – without the permission of IESE.

Last edited: 5/5/17

MN-389-E Conversion
C Rate Optimization

Table 2

2016

Visits Conversion Sales Average order value Revenue

1,000,000 1.5%  15,000 €120 €1,800,000

The figures in the tables above illustrate the two alternative approaches. In the first case, the
company gets 1 million online visits a month and converts 1% of these into sales. With an
average order value of ê1 , this generates a turnover of ê1. million a year. If the firm
managed to increase the number of annual visits by %, with the same conversion rate, sales
would increase to ê1. million a year. The second table shows that the same company, with
the same number of visits, could achieve exactly the same turnover by increasing the
conversion rate to 1.5%.

In the drive to improve the profitability of digital business, conversion rate optimization (CRO)
is generally considered the most suitable approach and tends to be the goal companies focus
on in their efforts to boost sales. This is because increasing the number of visits—especially for
a mature e-commerce company that has already optimized traffic generation and has a
sophisticated media plan in place—will almost always require more investment and yield
diminishing returns; that is, the average cost per acquisition (CPA) will rise as investment
increases, and the average quality of the customers acquired will probably decline (resulting
in a lower average conversion rate).

Using CRO techniques will help build a healthier business, strengthening a firm’s ability to
improve the conversion rate and, therefore, the cost per acquisition. These positive outcomes
are possible because CRO leads to a better understanding of the online platform and enables
companies to optimize budget distribution among the most profitable traffic sources.

1. What Is CRO and How Does It Work?

When people talk about CRO, what they usually mean is A/B testing and processes aimed at
improving a website, but there are various actions (which we will discuss at a later point) that
can be carried out beyond the platform and that also facilitate the optimization process. These
actions have an influence on the entire conversion funnel, that is to say, on the user’s journey
toward the goal determined by the company.

2 IESE Business School–University of Navarra

 Conversion Rate Optimization MN-389-E

Figure 1

Reputation

Awareness

Upper
Consideration

Preference

Visit
Middle
Conversion

Experience
Lower
Loyalty

Referral

The fulfillment of this goal, known as “conversion,” isn’t necessarily a monetary transaction.
It could be to get the visitor to complete a registration process, share some news, invite a friend,
become a fan of a sports team, download something, request a vehicle test drive, and so on.

Conversion is expressed as a rate or ratio—the number of visitors who take the desired action
divided by the number of unique visits to a website in a given period—and the goal is to
increase this rate.

Successful CRO depends on a series of disciplines, including design, programming, traffic

generation, e-CRM, usability analysis, and web analytics, among others.

2. The CRO Process

Conversion rate optimization should follow a series of prescribed steps to maximize its
effectiveness. The five stages of the process—all linked to an initial analysis to identify the
company’s business objectives—are as follows: preliminary analysis, formulation of
hypotheses, performance of specific tests, examination of results, and implementation of
continuous improvement based on what has been learned.

IESE Business School–University of Navarra 3

MN-389-E Conversion
C Rate Optimization

Figure 2

BUSINESS OBJECTIVES

04.
01. 02. 03.
Study of
Analysis Hypothesis Testing
results

05. Continuous improvement

2.1 Identification of Business Objectives

To understand how and where to apply CRO, it’s essential to understand the business and
identify its short-, medium- and long-term goals. This means first establishing whether what
the company wants to find out is how to earn more or how to sell more. Identifying specific
goals will make it possible to define the best actions for an optimal CRO process, such as
acquiring new customers, increasing the average order value, or increasing shopping cart
entries.

2.2. Investigation and Analysis

The analysis performed should be based on data, and the data used must be reliable. At this
stage, the analyst must become a detective who’s able to assess a website without any
preconceived ideas in order to identify all the factors that affect its performance, beyond a
positive or negative feeling about the site.

The first step is to analyze traffic to the site, which can be done using tools such as Google
Analytics, Omniture and Webtrends. Tools of this kind can be used to understand what traffic
sources feed the site (e.g. organic traffic, display ads, SEM, social media, etc.). This analysis
should determine the conversion rate for each source. Online tools usually provide average
rates for all traffic, but it’s important to dig deeper into the data. Investigating which Google
AdWords campaigns are most effective would be an example of this more in-depth approach.
While in principle conversion can be measured based on the last click (i.e. attributed to the last
traffic source), whenever possible it’s preferable to use attribution models that assign credit
fairly among the various traffic sources that led to conversion.

It’s very important at this stage to study behavior in relation to different devices (smartphones,
tablets and desktops). In general, the conversion rate for mobile traffic is lower than for desktop
and tablet traffic. Bear in mind that the rate can also vary significantly between different
operating systems (Android, iOS, Windows, etc.).

4 IESE Business School–University of Navarra

 Conversion Rate Optimization MN-389-E

It’s also important to focus on landing pages, which are the points of entry to the website. The
conversion rate for each landing page is a good indicator of how effective it is. Identifying
inactive pages that continue to appear in search results without producing any conversions
(and which divert traffic from the desired goal) is without doubt a good practice that should
be kept up.

The third step is to identify the conversion funnel and the stage at which the process is
interrupted. Most of the traffic generated by a website doesn’t lead to a purchase, and working
out the reasons why visitors abandon their shopping cart, a form, etc. is crucial to determine
where improvements need to be made.

An improperly configured web analytics tool will hinder efforts to optimize the conversion rate
and result in a CRO process that’s very costly and far from optimal. It’s therefore vital to ensure
that web analytics provide the information needed to identify exit points and solve this
problem.

The fourth and final step in this first stage of analysis is to study the behavior of both current
and potential customers when faced with the action that marks achievement of the objectives
set. This means looking at the way visitors navigate the website’s most significant pages. At
this point it may be useful to conduct eye tracking studies that identify the areas of a page
users pay most attention to. Also useful are programs such as Lucky Orange, Hotjar, Crazy Egg
and Clicktale, among others, which can provide a range of information, including click maps
and details of scrolling behavior.

2.3. Hypotheses
Having analyzed and understood what factors may be affecting the behavior of website users,
the next step is to formulate hypotheses concerning actions that may improve results.
Hypotheses can be generated based on what’s been learned, by drawing on the creativity of
those involved, and by using brainstorming techniques. For every problem identified, potential
solutions need to be proposed. Techniques for generating hypotheses play a particularly
important role in advanced CRO processes (when the most obvious improvements have already
been made).

After developing all relevant hypotheses and identifying every opportunity that exists, the next
step is to organize and classify them. The idea is to prioritize initiatives according to two
variables: the effort it will take to test each one, and its expected impact in terms of conversion.
This can be done using the hypothesis prioritization matrix shown in Figure 3, which is based
on the Eisenhower Decision Principle.

IESE Business School–University of Navarra 5

MN-389-E Conversion
C Rate Optimization

Figure 3

High

Rapid benefits Major projects

Impact

Secondary tasks Tedious tasks

Low

Low Effort High

Source: https://en.wikipedia.org/wiki/Time_management#The_Eisenhower_Method, last accessed December 2015.

This matrix helps identify activities that have a high impact and require less effort, so they can
then be prioritized accordingly.

2.4. Testing
There are two ways of approaching this stage. One is to perform A/B testing, which involves
analyzing two different options using two samples of statistically similar customers. For
example, this approach could be used to test two different landing pages: A and B. The results
will indicate which one works best to achieve the proposed objectives.

The other option is to do multivariate testing, which involves applying various changes to
elements on the same web page (buttons, text, colors, etc.) and testing the various options with
different groups of customers. After obtaining the results, a regression analysis is performed
in which the dependent variable is usually the conversion rate and the independent variables
are the elements that were changed on each page (usually represented as zeros and ones,
according to whether or not each change was present or not).

It’s very important to be patient when performing these tests. Obtaining a representative sample
can be a long process, especially for sites that get few visits. The aim is to pick out from among
all the factors analyzed the ones that lead to an incremental change in the conversion rate.

2.5. Reviewing Results and Drawing Conclusions

After testing has been performed, results should be compiled to determine whether the business
has “gained” or “lost.” When success is achieved, the test will indicate a higher conversion
rate; when this is not the case, there will be a negative outcome that will hurt daily sales.

6 IESE Business School–University of Navarra

 Conversion Rate Optimization MN-389-E

Failures in A/B or multivariate testing lead to lost sales that cannot be recovered. Therefore,
the performance and analysis of tests should not be left in the hands of non-experts or people
with a limited knowledge of statistical analysis. The conclusions drawn from the testing
performed are a key part of the process, and it’s important to be sure that they are statistically
significant. Otherwise, there’s a risk that actions taken on this basis will hurt the business—the
opposite of what’s intended.

2.6. Repetition and Continuous Improvement

The CRO process should be iterative so it contributes to continuous business improvement. The
learning curve will be very steep at first but will level off over time. At some point, CRO will
involve complex processes in which it will be increasingly difficult to make changes that lead
to positive outcomes. Any gains at this stage will be barely discernible.

When pursuing continuous improvement in this way, one must constantly ask to what extent
processes can be improved and whether benefits achieved in some areas of the site can be
replicated in others. It’s also important to consider the cumulative impact of changes
implemented over the medium and long term. The combined effect of various changes may
result in a user experience that’s quite different from what was initially intended, and this is
easy to overlook.

3. Useful Applications of CRO

Implementation of CRO methods often falters because those involved lack technical knowledge
of key tools and how to use them, and as a result feel they don’t adequately understand the
process. However, given that CRO can be applied to the entire conversion funnel, in many
cases it’s not necessary to be an expert on A/B testing to have a hunch about what stage to
focus on first. Common sense is often sufficient to identify various areas where improvements
can be made. These relatively straightforward areas are discussed in the following sections.

3.1. Reviews
Brand and/or product reviews are essential to build a good reputation and attract high-quality
traffic to a site. They’re also the first step on the path to conversion. People are increasingly
influenced by reviews left by users who they don’t know, but whose comments they see as
having high credibility and authority. Paying attention to online feedback and reacting
appropriately to both positive and negative comments is an effective way to influence the
quality of traffic and thus the conversion rate.

3.2. Persuasion
Making the investments required to persuade users to move forward through the conversion
funnel is a top priority, particularly given that 97% of users don’t convert on their first visit
to an e-commerce site.1 Retargeting techniques and creative ideas to call users to action, invite
visits, and drive conversion can be useful approaches.

1 See http://www.convinceandconvert.com/customer-experience/cart-abandonment, last accessed December 15.

IESE Business School–University of Navarra 7

MN-389-E Conversion
C Rate Optimization

3.3. Products
Twenty percent of the products on an e-commerce site will account for % of sales, so it’s
well worth identifying those items (especially the ones that work best) and making them very
accessible to the consumer. This simple measure will automatically deliver a quick win.

3.4. Forms
Twenty-four percent of users abandon forms they find too complicated.3 A relatively fast
improvement can be achieved by shortening or simplifying this process.

Eliminating nonessential steps can be a good starting point. If it’s not absolutely necessary to
ask for an ID number, for example, it’s better not to. It’s okay to request information, but this
should be done in a gradual way. Customers can also be offered a number to call, or given the
option of having the company call them.

3.5. User Registration

Visitors to an e-commerce site are often there for the sole purpose of making a purchase, so
forcing them to register to complete a transaction isn’t usually a good practice. To make users
want to register, the site must offer added value they can perceive immediately. The registration
process can come at checkout, or the customer can simply be allowed to make a purchase
without registering. The key is to make the transaction as easy as possible.

3.6. Shipping Costs

Fort-four percent of users abandon their order because of added shipping costs, 4 so it’s
important to keep these charges as low as possible and ensure that they’re not used to offset
other expenses.

3.7. Calls to Action

Helping users find their way around the site also boosts conversion: 53% of people take more
than three seconds to find what they’re looking for.5 The idea is to reproduce in the virtual
realm the same kind of signposting we see in the physical world. For example, in the El Corte
Inglés supermarket there are signs to help customers quickly find what they’re looking for;
replicating such elements in the digital environment can significantly increase the conversion
rate.

3.8. Security
Given that many users are wary of making online purchases, it’s vital to display trust seals.
The less the user knows about the brand of the products or services being sold, the more

2 See http://backlinko.com/conversion-rate-optimization, last accessed December 15.

3 See https://econsultancy.com/blog/ 71 -1 -ways-to-reduce-basket-abandonment-on-your-ecommerce-site, last accessed
December 15.
4 Ibid.
5 See http://www.convinceandconvert.com/customer-experience/cart-abandonment, last accessed December 15.

8 IESE Business School–University of Navarra

 Conversion Rate Optimization MN-389-E

important this is. The reverse is also true: the more the user knows about the brand, the fewer
trust seals will be needed. Nevertheless, it is strongly recommended to include them in all cases.

4. Benefits of CRO
Applying CRO techniques yields both immediate, direct benefits and indirect gains. The former
will be apparent in the bottom line and are associated with higher and better conversion rates;
the latter will enrich the site thanks to improvements in SEO and SEM. CRO makes a site more
usable and reduces the bounce rate, which helps improve SEO and, as a result, search engine
positioning. The impact in terms of SEM is also clear: good usability and a high conversion
rate, combined with more effective landing pages, will result in a better AdWords Quality
Score, resulting in a lower cost per click (CPC).

IESE Business School–University of Navarra 9