The Golden Metric

Jun 19, 2014

Photo of an iceberg
Are you looking for the “golden metric” that is the best measure of your agency’s website performance and cross-comparable across .gov websites? If so, stop looking. The concept of the golden metric is a dangerous one because it oversimplifies performance analysis of your website and overlooks the truth hidden behind other, more relevant metrics.

Don’t get me wrong—it is easy to fall for the concept of the golden metric. Everyone wants to see more visits and pageviews on their agency websites. It’s quite common to hear management ask for a chart with “Total Visits” or “Total Pageviews” as if it is the key indicator of the website’s success. Naturally, a growing, positive trend of total visits and pageviews tells a great story and makes everyone happy. On the flip side, when total visits and pageviews numbers are declining, the perception is that of a “bad website” leading to potentially premature conclusions.

In reality, it’s not that simple. Let’s say that your website focuses on providing information to help people apply for a government grant. And, your total visits are declining, which seems like bad news. However, by segmenting your visitors by grant eligibility criteria, the data reveals that the number of eligible visitors for your program is rising, and their rate of website visitation and grant application is growing.

Or let’s say your website’s bounce rate (% of visits that resulted in only one pageview) is changing significantly (growing or declining). That could mean completely different things depending on your website goals. For example, for a navigational website, such as USA.gov, a high bounce rate is an indication of success because the site goal is to help users navigate to another agency’s website as efficiently as possible (i.e., with fewer pages per visit). On the other hand, if you are a heavy-content website, such as Data.gov, with a goal to provide users with lots of content and relevant data files, a low bounce rate is better.

What Really Matters

At a recent Google Analytics Summit, there was a quote mentioned that resonated with me,

“Big Data at scale requires clarity of scope, strategy, context, governance, metrics and data standards.”

Before gravitating to total visits and pageviews trends, website analysts should have a clear idea of what their website is set to accomplish in the first place and then interpret performance results within the context of the goals.

Think about these questions before starting your website performance analysis:

  • What are the objectives of your website? Do you want visitors to stay on the website and read pages and pages of content, or, do you expect them to move to another website after the entry page or the search results page?
  • What are the top success scenarios for your website with respect to visitor behavior?
    • For instance, if you are a website such, as GSA Advantage.gov (GSA’s transactional e-commerce government website offering products and services for purchase), you may want visitors to land on the website, perform an on-site search for a product, browse the product pages, add a particular product to the shopping cart and complete a purchase.
    • If you’re NASA.gov, one of success scenarios may be a visitor landing on the site, searching for a particular topic or reading an article, and watching a video.
    • If you are Data.gov, a success scenario could be a visitor coming to the site, performing an on-site search for a particular topic, and downloading a data file on that topic.
    • Finally, if you’re DigitalGov.gov, your top scenario may be a visitor coming to the website, reading a blog post, joining a community, signing-up for a service, and subscribing to the DigitalGov Email Updates.
  • What is the website’s outreach goals? What efforts are being done to reach the website audience (e.g. email campaigns/subscriptions, social media, government referral links, SEO etc.)?
  • Who is the website’s key audience—general public? College students? Disaster survivors? Veterans? Are you targeting U.S. visitors, international, both? Are you targeting desktop or mobile visitors? What are the specific goals for each audience segment reach?

So, for a website like DigitalGov.gov, looking only at total visits and pageviews would tell you little about the website’s success. To measure its effectiveness, you’d want to look at visitor to subscriber conversion rate (visitors who visited the site and subscribed for the DigitalGov Email Updates), growth in returning users, level of engagement (e.g. % of visits with comments on posts, pages per visit), bounce rate etc. As part of this analysis, you also may find that you have a segment of visitors who visited the website only once and did not return. In that case, you can perform user segmentation analysis to compare visitor acquisition methods, demographics and other relevant information of the “one-time website visitors” to that of the repeat visitors/subscribers to identify possible reasons for the difference in behavior.

Metrics With Meaning

In today’s world, there is data everywhere, and everyone is wearing an analyst’s hat. Data in the hands of folks doing Web performance analysis without clear website goals at hand, without a clear analysis scope and without an understanding of each metric can be disastrous. Looking for that golden metric, or simply comparing total visits, pageviews, and other common metrics across .gov websites makes no sense without a strategic context.

The DAP Web analytics platform allows government agencies to go from high-level (executive dashboards and overall trends) to in-depth, multi-dimensional analysis of agency website performance through segmentation, channel and attribution analysis, popular content analysis, event analysis (outbound links and downloads), and many others. So, bring together your .gov website goals and the robust DAP analytics data to make your performance analysis truly meaningful, insightful, and actionable.