Finished thesis

I finally finished my thesis. Not quite on time, but with good results! I defended my thesis today, 30 January 2013, and received the mark 10, which is second-best on the new Danish scale.

For anyone who might be interested, the thesis is available for download below:

Conversion Rate Optimisation: Barriers to Adoption by Sisse Marianne Bertelsen

Are technical obstacles a poor excuse?

In my interviews for this thesis, some companies expressed concerns regarding technical obstacles. For instance, a Java-based website needed to be replaced because it could not be crawled by search engines and the code would basically screw up data collected by the web analytics tools. The code would reload the page several times, making it look like a page had been visited multiple times rather than just a few.

The authors of Actionable Web Analytics are very insistent that technical issues are poor excuses for not collecting good quality data. They write:

“All leading web-analytics tools can work with almost every possible development scenario. They can track many sites out of the box, and for the rest, they have plug-ins or other applications to help retrieve data. It doesn’t matter if your site is dynamic, is Flash-heavy, relies on a content-management system that scrambles URLs like eggs, or even uses a cutting-edge technology like Ajax. It can be tracked.”

So I’m wondering: Are technical barriers always just a poor excuse? I mean, is it really always possible to correct tracking issues one way or another? I suppose that even if it is possible, we will still meet the issue of companies lacking the knowledge to implement their web analytics tool correctly.

Definition: Conversion

When I try to explain the topic of my thesis and mention conversions, people sometimes ask “conversion to what?”. The word conversion can be many things, and so I have tried to devise a definition for the purpose of my thesis.

There are many definitions of a conversion in the online marketing context – what most of these definitions share is that a conversion is an “action” which is tied to “business goals” or “business value”. This underlines the importance of tying the website to the overall goals of the business. Some examples of specific actions could be: buy, download, opt-in, register, refer a friend, make a phone call and click to chat.

Brian Clifton chose to define “goals” and “KPIs” separately:

“Goal conversions, also referred to as simply goals or conversions, are any actions or engagements that build a relationship with your visitors.”

“… a key performance indicator is a web metric that is essential for your organization’s online success.”

With goals, Clifton is talking about micro conversions. However, for the purpose of this thesis, I would like to lean on the definitions where conversions are tied directly to the company’s business goals. This is because we here focus on conversion rate optimisation in a way that helps companies improve their performance online – a focus on a greater level than micro conversions, though these are steps on the way. So Clifton’s definition of KPIs, which would be closer to macro conversions, becomes more relevant. What is interesting about this definition is that instead of “action”, Clifton uses “web metrics”. This adds a dimension to the definition: actions must be measurable. Tim Ash confirms this: “A conversion happens when a visitor to your landing page takes a desired conversion action that has a measurable value to you business.”

So, for the purpose of this thesis, a conversion is defined as follows:

“A conversion is a measurable action, taken by the visitor and defined by the company, which is crucial to the company’s business goals.”

Any constructive comments are welcome.

Web Analytics Maturity Model

I am currently having a look at Stéphane Hamel’s Web Analytics Maturity Model. What strikes me is that the model goes way beyond what I used to think of as “web analytics”. Hamel’s definition of web analytics in itself says a lot:

“The extensive use of quantitative and qualitative data (primarily, but not limited to online data), statistical analysis, explanatory (e.g. multivariate testing) and predictive models (e.g. behavioral targeting), business process analysis and fact-based management to drive a continuous improvement of online activities; resulting in higher ROI.”

This definition also clearly encompasses something which I see as belonging in the world of CRO: continuous improvement of online activities; resulting in higher ROI. Does that mean that we cannot distinguish between web analytics and CRO? I mean, CRO clearly is very dependent on data and thus web analytics. But are the optimisation activities not something that goes beyond web analytics? I confess to being slightly confused by this overlap.

Although Jim Sterne is quoted for saying “it was a mistake to call it web analytics, it is business analytics” is analysis about taking action on one’s findings?

Definition of Conversion Rate Optimisation (CRO)

Conversion Rate Optimisation (CRO) is the practice of reconstructing a website, or parts of a website, with the purpose of getting visitors to perform certain actions. These actions can be a purchase of a product (for e-commerce sites), signup for a newsletter, clicking a certain link and whatever else that you think of which is relevant to your website.


The changes can consist of revisions of colours, text, shapes and images. Furthermore, an important part of CRO is to test what the effect is of these changes. It is important to test one change at a time, in order to get valid statistics on which change actually made a difference – whether positive or negative. If the change is negative, you simply roll back. But in order to do that, you need to know exactly which detail gave you those results.

Most tests are carried out with the help of tools for A/B or multivariate testing. These tools are useful, because they do all the statistical calculations for you. With such software tools, it is possible for anyone to carry out testing on their own website. If such tools did not exist, you would need a skilled statistician to make sure you got valid test results.

The importance of CRO

Now, the importance of CRO is crucial. There are many tools for getting visitors to a site, such as various SEO methods and many kinds of advertising. However, once a website gets a certain amount of visitors, the question comes up: What to do with them? Why do you have visitors? What do you want them to do? If you don’t really want them to do anything, why have a website? Surely, you must have a wish for them to do something specific. This can be one action or many actions – but you need to get clear on what they are. Once you have these actions or goals in mind, you can start working with CRO.

CRO consists of many different activities. For instance, working with colour theory, copywriting, call-to-actions and images.

So where should you start?? Well, that is the real problem… (see also A Beginner’s Guide To CRO?)