Increasing Website ROI

The last couple of posts have been regarding ROI and how to measure it. Once you know how to measure ROI how do you increase it? The simple answering is testing. Testing means different things to different people. Ask a programmer and their definition of testing involves clicking on links and testing the functional aspects of a site. Others, mostly non-marketing professionals, will assume testing is having the client view the Website or page and approve it. However, testing to a marketing professional involves much more than the functional aspect of the site.

Testing involves usability test, A/B split and multivariate testing. For best results you should be using a combination of all three. One of the most common misconceptions about testing is that is expensive. Our answer, is how expensive is it to NOT to test? If you want to put some numbers with this answer find out what your conversation ratio is and determine how much profits will increase by increasing your conversion rate.

We’re going to look at usability testing. First, let’s explain what it is. Usability testing means that you observer how a test subject uses your site. This can be as scientific and high tech as some of the labs we’ve used with two-way mirrors, several observers, multiple cameras and software recording every mouse move and keystroke. Or the test can be extremely low tech by simply observing the testing.

If you’ve been in marketing for sometime, I’m sure you’ve been involved with focus groups. This is somewhat similar. In focus groups you gather in a sample of your target audience and they use and discuss the product. In usability testing you gather a sample of your target audience and ask them to complete several timed tasks on your Website.

Each task it timed and recorded. Since the site owner should not be present at the test it is best to record the session. Designers and programmers also benefit greatly from viewing the testing as well. Many times sites are developed and designed based on our own preferences. You many notice that designers get stuck in “template” mode. The designs are appealing but they all begin to all look very similar. Participating in usability test is one way to stretch the designer’s creativity and sharpen their skills.

Each usability test we’ve done have always uncovered at least one “gold nugget” that made the testing worth doing. Usually there are several areas on a Web project that can be improved as a result of usability testing. The reason is simply. If your target audience can easily use your site the likely hood of them taking action increases.

One of the misconceptions about usability testing is that it uncovers problems with navigation or ease of site use. However, much more than those areas should be tested. One of the first questions we ask the subject is “what is this site about” or “what service/product does this company offer”. Then note how long it takes the subject to answer the question. These questions have nothing to do with design or navigation but rather content. If the subject is unable to quickly say what the site is about how can the site be successful?

A common mistake in usability testing is that the tester fails to test the test. Never conduct a test without first testing the test yourself. The test developer and tester should BOTH take the test.

I’ve often been asked from new clients how do they know if their current site has been given a usability test. Since I know this important phase of development is skipped by most developers (many are totally unaware of the process) I know can safely say their site was not tested. My response is that if the developer never mentioned it then it safe to say that it was not tested. There is cost involved in testing and no one is going to do extra work and not get paid for it.

How important is usability testing? Well the answer goes back to the earlier topic about how expensive is testing. If you know that a 1% increase in your conversion ratio will yield 300% in gross profits – then you can easily how important testing can be