Wednesday, 18 March 2015

Debunking the 5 Myths of Conversion Rate Optimization

Some conversion optimization myths are not easy to spot. You hear about them in conversion conferences, you read about them in books but years of practice reveal that you have to be more careful before a myth can cost you money and valuable time.
Let’s take a look at the most common myths of conversion rate optimization (CRO), keeping in mind that the same thing won’t work for everyone. You need to keep an open mind and remember to take a step back to make sure you’re not doing something just for the sake of it.

Myth 1: Always Be Testing

When I started in the field of conversion optimization about 10 years ago, it was difficult to convince anyone of the importance of testing. It was typical for us to hear a customer say, “you are the experts, just tell me what I need to change on my website and I’ll do it.”
We trust you so we do not see a need to test.
Most clients wanted to take optimization recommendations and not spend money on implementing different designs and running AB or multivariate tests to figure out which design will increase conversion rates. Many thought that these tests were a waste of time and money. In order for those of us working in the field to convince people of the importance of testing, we had to sell the idea of A/B and multivariate testing in conferences, articles, and seminars. Nowadays, I think it is fair to say that the culture of testing is taking its roots in the online community. However, the early push to convince people about the importance of testing might have oversimplified the process of testing. We did not focus on how to select the right pages to test or what elements should be tested on these pages. Another side affect was that people confused the process of conversion optimization with process of testing.
Conversion optimization includes several steps:
  1. Analyze analytics data to determine which page should be tested
  2. Analyze selected page to determine problems
  3. Prioritize problems
  4. Create a hypothesis on how to fix the top problems
  5. Create mockups for the new designs
  6. Implement the new designs on website
  7. Test (A/B or multivariate) the different designs to determine the winning design (hypothesis testing)
  8. Analyze the results of the test and determine marketing insights
As you can see, testing is only one step within the process of conversion optimization. As you attend marketing conferences or read articles about conversion optimization, you will start noticing the new push to persuade the online community about the importance of following a conversion optimization methodology.

Myth 2: Always Apply Best Practices

What elements should be tested on a particular page? When evaluating a particular page, people tend to fall back on best practices to determine which changes should be made to the page. There has been a lot written about usability best practices, however that field has a large mix of good, bad and ugly advice. I always give the example of an online retailer that I worked for back in the 1990s. We had one simple best practice that we followed: Copy That best practice did not prove very effective because no matter how much copying we did, our conversion rates remained around 1%.
Best practices are good in theory but difficult in practice
What works for one website, may or may not work for your website.

Myth 3: You Need to Invest in Expensive Testing Software

As more online companies focus on testing, there are many options you can select from to deploy an A/B or multivariate test. The cost of the different software packages can vary a lot. You can conduct the tests for as little as for free or as much as many thousands of dollars. While the features of the different packages vary a lot from the simple point and click testing packages to the more complicated testing software. There are also differences between these packages on how they determine a winning design and which calculation methods they use. For most of our customers (and these include some top online retailers), we have been successful in conducting tests in one of five different packages:
  • Google Analytics content experiments
  • Pii Testing Engine (our own)
  • Visual website optimizer
  • Optimizely
  • Omniture
Google Analytics content experiments is free but has a limited set of features. It also does not support multivariate tests. Omniture on the other hand is the most expensive of the five options I mentioned above. Pii, VWO, and Optimizely are close in price and features to one another however one of them might work for a particular test or a customer better than another.

Myth 4: The More Tests You Run, the Higher the Conversion Rate You Will Get

Customers want higher conversion rates and they want them now. I would say that half of our customers are always asking us to conduct two or three experiments in a given month. We usually push back. First off, we do not recommend conducting concurrent experiments.
Let’s say that you are testing category and product pages concurrently. A customer views one design in category pages and then he views another design in the product pages. Neither of these designs is the original. That customer might not convert as a result of that particular combination, or even if he converts, there is not way to link the two designs that caused the conversion. Things can get messy.
In addition, the most powerful of conversion optimization is the marketing insights which you should be learning from conducting a particular test. Pushing for multiple experiments means that portions of the conversion optimization process will be ignored and its full impact will be minimized.
There are of course few exceptions to this rule and that will depend on the size of the dedicated team that handles CRO for your company. For 99% of cases, conducting multiple tests within a month backfires.

Myth 5: Your Design Team Can Handle CRO Implementation

One of the biggest roadblocks to successful conversion optimization is not understating the amount of investment it requires. Producing mockups for new designs suggested for a particular page or set of pages on your website is just half of the technical problem. You must have planned for the technical resources who will be able to convert these mockups into actual code that can be deployed on your website.
The team must also be familiar with the particulars of the testing engine that you will use to deploy the tests. The cost of implementing a test suggested by a conversion optimization team can add anywhere from 20% to 40% to the overall cost of the project. In addition to that, relying on non-dedicated internal resources ignores the fact these resources will have other responsibilities.
In business, priorities are constantly changing. And a resource that is not focused on conversion optimization might push implementing a test down the line. At the end, you are not able to keep up with testing and in many cases might give up on it.
You have to understand the full commitment that is required to conduct a successful conversion optimization project and assign all the required resources to it.


We’ve taken a look at 5 myths that you’ll want to keep an eye out for in CRO, but we’ve also learnt that the same thing won’t work for everyone and in some cases one of these myths might be the right way to go. The important thing to remember is that you shouldn’t blindly follow what others have done or what has worked for a different project in the past.
You need to make sure you’re paying attention to what your data is telling (or not telling) you and make appropriate adjustments.

To view the original article
Click Here

No comments:

Post a Comment