Sep 13th 10 Posted by CJ Cunniff in Digital, Featured Articles

Optimizing the User Experience with Web Analytics

In the marketing world, we are constantly faced with conflicting priorities. In higher education, those conflicts might look something like this:

• President: Our institution should grow in number, quality, and ethnic mix of students.
• User: I’m expecting to see information about student life, academic quality, financial aid, etc., and I’m expecting to see it in a certain way.
• Marketing Department: Our website needs to attract more qualified students, with a limited cost, and try to meet all of their needs.
• Dean of _________  : We can’t title that page “Welding Engineering Information”; it should be “Materials Joining Information.”

These are some dilemmas we will definitely be discussing during our SIM Tech® Seminar in Las Vegas this year, but I thought I would go ahead and get the discussion started now.

Quantitative information is one tool we can use to make decisions on how to prioritize each of these needs. We can get a lot of what Avinash Kaushik refers to as “What” data with analytics data. For example, I could tell you that from the school’s home page x% of prospective student go straight to financial aid, x% go to student life, and x% view academics information. I can create unlimited ways of segmenting this data and eventually give you a report on the section of our home page that seems to drive the most traffic and therefore might be seen as the most important area on the site. What I’m not telling you at this point, is why they are doing that.

This is where the art of user experience design and the science of analytics based site optimization meet. Understanding the user experience requires a lot of qualitative data in most cases. While the question of “Why” is more difficult to answer, it can provide tremendous value to you as well. There are several attributes that define a user experience.

http://semanticstudios.com/publications/semantics/images/honeycomb.jpg

In each of these attributes the user is asking himself questions like:

• Can I find what I’m looking for?
• Is the information I found valuable and/or useful?
• Is the website I’m on usable (in other words can I do what I need to do)?
• Is the information presented to me in a way that is desirable (in other words, will I do what I need to do on this site)?

The key goal behind user experience based design is to present an experience to our users that will incentivize them to take a desired action. You can gather this information in several different ways:

• Do focus group studies
• Run usability tests
• Use your clickstream data (generally your Google Analytics data) to create conjectures as to what is happening on your website for a particular segment of your audience

This is really very key information. With our analytics data we might determine that the academics area is the most important area on our home page.  However, when we begin to analyze why that is, we might find that our users are actually looking for admissions information but because of the graphics and confusion with the terminology they are getting confused and going to the wrong area from our home page.

So the question is really, how can we combine the “Why” information with the “What” information and use that to optimize my site for my institution and the user at the same time?

I would propose two ways to analyze the same basic information which can give us at least some insight into the “institutional need verses user need” dilemma.

The Top-Down Approach

Take a Top-Down approach where you define your Key Performance Indicators (KPIs) for your site from the institutional perspective. Then utilize your reporting and analyzation capabilities to determine how well you are meeting your goals, and which areas would provide you with the highest possible return on investment if optimized further. What you might find is that people are getting to the correct degree program but not from the correct geographic area. By using geographic segmentation you might determine that a program that is only available on campus is being included in your Pay Per Click campaign which is running nationally. This is naturally going to both cost you a considerable amount of money and a considerable amount of frustration for your users.

The Bottom-Up Approach

In this approach (as described by:  LOU ROSENFELD on A List Apart) you are not pre-defining what you are looking for. Instead you are looking for trends in what your audience is looking for and trying to make determinations about why they might be doing that. For example, you might notice that a particular degree program is always at the top of the bounce rate (bounce rate is defined as  what occurs when a user visits a single page on your site and then exits) report. A natural question might arise as to what is causing that higher than expected bounce rate.  Next you might look at a site overlay, and even ask people what their experience has been on these pages. By the end, you might find out that a very prominent link is taking people to another site due to an error, or that the link is hidden and is what people are looking for which is causing frustration.

Choose Wisely!

So which approach should you be using? That’s simple… use them both. By doing your research with both methods you will force yourself to look at the data through the lenses of both your institutional goals and your user’s goals. When you find areas of overlap where your return on investment would be high for both types of approaches, you have a very strong sign that this is an area where you should be focusing some time on your site.

So, have you had a personal experience with combining your institutional goals with your user’s goals?

What did you find and do you have any tips on how to deal with similar situations?

And, if you’re in the giving mood, could you please share your institution’s home page click data?

Shameless Plug…

If you want to continue this discussion further, you can always tweet me (@cjcunniff) or better yet use our Sim Tech 2010 hashtag (#simtech10) on twitter.  Of course I would also love to discuss this further with you in Las Vegas at this year’s Sim Tech as well.