At SAMHSA, our Web improvement efforts center around a “customer first” approach in everything we do. Much of our recent work focused on understanding how information on the agency’s website should be organized and prioritized.
How are we doing this? A variety of tools and techniques are helping us meet this goal, such as directly soliciting feedback from customers via comments and card sorting exercises, using trending social media stories to identify what issues may be of greatest interest to SAMHSA customers, and examining website metrics to find out which website features are our customers’ favorites.
Understanding Web metrics using measurement tools like Google Analytics or Webtrends helps uncover what visitors find most and least useful on our website. When paired with usability testing, surveys, or direct feedback, Web metrics guide how we deliver information to our customers effectively.
But Web metrics can’t tell us everything, so we don’t stop there. To ensure we’re delivering the most relevant information for our visitors, we must keep in mind that site placement often drives clicks—even if that information is not what visitors may need the most. Pages linked from the homepage often have the highest click-through rates, giving the impression that this is the most popular content. However, it may be that the prominent placement of this content allows visitors to more easily locate and visit them.
Search engines add another level of complexity to interpreting Web metrics. Google, Bing, and others drive more visitors to webpages linked from a website’s homepage than to pages not connected to the homepage directly. In SAMHSA’s case, this may mean that greater numbers of people are finding pages on SAMHSA.gov because those pages are linked to from the website homepage, not necessarily because visitors are more interested in those particular pages. Taking a deeper dive into our Web metrics allows us to properly analyze visitor trends and page popularity in more depth, providing us with the whole picture of visitors’ interests.
We know our information is only as good as the ease with which it’s accessed. To that end, we’re always thinking about sorting out these potential data biases and understanding how the data is being provided. In our case, we’re leveraging many different datasets simultaneously to ensure we have a full picture of what our customers like and what they need.
Be sure to look for more about our use of these tools and methods in upcoming posts.