What do website visitors really want?

Implementing most of the best practices highlighted in the Gartner report requires a major assumption – a tacit understanding of what customers who visit your website really want. But how in fact do you determine that? There are really only three options available for delivering an optimal user experience to your visitors.

One method is to guess what users want. The majority of websites today seem built on the guess method. Content authors post content they think will interest users. But how do they really know? For the most part, they make an intelligent guess. Even Mr. Maoz advises, "Use a cross-departmental team to discuss what end customers most want to hear from you."

The problem with the guess method is that website authors and designers more often than not guess wrong. The proof of this is all around us: users who leave websites after spending only seconds on a page; poor conversion rates on landing pages; high abandonment rates of shopping carts. When a website truly delivers an optimal user experience, visitors will stay longer, access more pages and content, take more actions, and convert at a higher rate.

A second method of finding out what users want is to survey them. Says Gartner Analyst Maoz: "Correlate [your discussion of what customers want to hear from you] with actually asking the end customer."

The problem with surveys is twofold. First, the answers you get are based on what people say, rather than what they do. The other problem with surveys is that a visitor who completes a survey about his user experience has already had the experience. You may be able to implement changes to improve future visits for him and others. But the experience he already had was less than optimal, and you were unable to do anything about it.

In addition, surveys are not continuous. They have a start and end date. To keep them going is time-consuming and potentially expensive. Also, unless you hire professional surveyors, you are mostly limited to surveys of your opt-in list, which can generate skewed data.

A third method of discerning what users want is through analytics: tracking a variety of metrics that show us things like what pages and content visitors looked at, how they arrived at those pages, and how long they spent looking at each.

Now we are on to something. Analytics tell us what visitors like and don’t like based not on what they say, but on what they do while on a website in real-time aggregated over thousands of user sessions. So the data is more accurate.

Analytics deliver continuous data; unlike a survey, they do not end. In addition, the cost of running analytics software is not prohibitive. And, analytics reflect the behavior of literally every visitor to the site, not a cross-section.

One potential problem with analytics-based thinking is the "action chasm" – defined as the lag between when metrics are measured and when analytics reports are read and acted upon.

When typical "stand-alone" analytics packages are used, there can be a long lag time between analysis of web metrics and the actual taking of corrective actions that will improve the website and create a more optimal user experience.

A new breed of web metric analysis, "integrated analytics," may be the key to shrinking that lag time from days or weeks down to minutes or even seconds (real-time), enabling the website to serve users more relevant content based on the actions they are taking during their visit.

It is integrated analytics that can enable us to understand the preferences of our users more intimately and thus achieve the optimal user experience conforming to the best practices from the Gartner report.