Conversion Optimisation - Are You Closing the Deal?

Step 1: Analyse

There are a number of ways in which one can analyse how users are interacting with a website, and thus determine the effectiveness of your web pages.

Listening labs
An evolution of usability testing whereby a user, supervised by a moderator, sits at a computer in a lab and interacts with the target website. The user is normally observed through a one-way mirror, and his actions and reactions to different elements on the site are investigated.

In a typical usability test the user would be given a defined objective (e.g. complete the purchase of a certain stock-item from Amazon). Listening labs differ in that the task is not defined beforehand. Rather, users are interviewed at the beginning of the test and are asked how they currently relate to the site - what services/products are they interested in, what competitors would they consider using, where they normally access the site from, how computer literate they are, etc.

Then, a specific task is set according to the user's context. While the user performs this task the moderator asks him questions to better ascertain how he is interacting with the site.

Website stats
There are currently a number of software tools that can aid in gaining an understanding of user behaviour on the site. ClickTracks and Google Analytics are two of the best known tools. One of the most important contributions that these tools can make is that they provide a means for tracking the most common click paths on a site.

The analysis is not only focused on the actions on the site itself but should extend to the collaborative efforts of all your eMarketing efforts up to this point. So instead of just investigating what your visitors are doing on your site, you should be checking that your Email Marketing, SEM and Online Advertising campaigns are driving the correct visitors to it. This can be done by seeing what links are driving traffic to your site or which search phrases are leading the most users to you.

Split testing
Split testing involves coming up with a hypothesis, and then testing it. This involves two sites with a single difference - for example, Option A might contain the words - "buy now" in the copy while Option B has a much more casual tone, or else the one option might ask for a consumer's phone number before completing the purchase while the alternative option only asks for it upon completion of the purchase.

Visitors are then randomly sent to either option in a statistically equal manner and the conversion rates for both sites are compared.

Multivariate testing
A more advanced form of split-testing, multivariate testing takes multiple options of the page elements and then creates a separate page for each unique combination.

The effectiveness of multivariate testing can often be limited by the amount of time that it will take to get a statistically valid sample of visitors to each of the options. Additionally it can also require a lot of time to successfully analyse a large number of alternatives.

Marketers are often very wary of undertaking split or multivariate testing, but one should bear in mind that direct response advertisers have used split testing for a long time. Testing of this sort can cause your site to operate at less than ideal for a short period of time due to some bad options being tested, but one bad day can very easily translate into a future lifetime of rewards.

Other measurables
eMarketing is in itself an extremely measurable discipline, and many of the component services that make up an eMarketing strategy can provide a wealth of valuable data. PPC data, banner advertising statistics and affiliate tracking can all be used to better understand your users.

Tracking the return on investment across these services can determine the effectiveness of each and highlight which should be optimised in order to get visitors to their desired locations.

download Pdf

Prev | 1 | 2 | 3 | Next