Web Metrics 101
Understanding the numbers leads to better content, stronger marketing and more efficient use of resources.5 Tips for Beginners
If you're just starting out, the terminology and massive amounts of available data can be daunting. Here's some free advice based on my experience:- "Hits" is a four-letter word. It's not meaningful for audience development; don't use it. Instead, use the appropriate term -- Page Views, Visits, etc. -- and encourage others to do the same.
- Know the difference between a Visit and a Unique Visitor (and be able to explain it to others in plain English).
- Unfortunately, every metric or methodology has some imperfections. Just be aware of what they are, so you can make educated decisions.
- Avoid making assumptions based on any one metric. It's rare that a single measurement is useful by itself, without the context provided by others.
- Help your team focus on improving a few key performance indicators (KPIs) that are most important to your strategy, rather than trying to improve on everything.
Definitions of Basic Web Metrics
This is hardly a comprehensive list of metrics, and you may find that not all these are worth focusing on for your business. But being able to use and speak about these will get you off to a great start.
These terms are defined in my own words; for additional insight, try one of the resources listed here.
Hits: Simply put, requests made to the server to load something onto a web page, like a photo. So, imagine your site has multiple pages with several photos each. You could have dozens and dozens of "hits" but few actual page views. Further, all that activity could be generated by just one really click-happy visitor. Hence, hits are not a useful indicator of audience growth.
Page: An analyst definable unit of content. How I describe it: A group of content elements (pictures, links, etc.) that you can view without clicking someplace else.
Page Views: The number of times any page is viewed. Page Views are additive: Add them together over any time period without worry of duplication or overstating the figure.
Page Views per Visit: Total Page Views divided by Total Visits gives you an average number of site pages viewed during a typical user session.
Page Views per Unique Visitor: Total Page Views divided by Unique Visitors for a given time period yields the average number of pages viewed by a typical visitor (person) the reporting period.
Panel-Based Measurements: Rather than measuring everything that happens on your site as Omniture or Google Analytics do, a panel-based methodology gauges the activity of a sampling of internet users -- similar to a phone survey in which random households are selected, but results are reported as applying to the general population. A web panel may be comprised of volunteers who allow tracking of the sites they visit. In other cases, providers form partnerships with ISPs to collect and report aggregate data. Whatever the panel does is then deemed indicative of all internet users in a particular geographic area or audience segment.
Some well known panel-based metrics providers are Hitwise, comScore, Nielsen and Quantcast. Useful questions to ask about panel-based data:
- How big is the panel? Too small a panel probably means there's too big a margin of error for the measurements to be useful. (If the provider is reluctant to provide panel size, that's a potential red flag.)
- Is it truly representative of the total audience?
- Does it capture at-home internet use only, or does it also capture at-work, school or public access (like libraries)?
Referrers/Referring Domains: A list of websites that have sent you traffic via a link to a specific page on your site. For example, at PalmBeachPost.com, we often see another site, like ESPN.com or Phins.com linking to one of our sports stories.
Time Spent: The number of minutes, on average, spent on a page of your site, or the site as a whole. Omniture provides us with time spent per visit as well as per unique visitor. Variations of Time Spent include Average Visit Duration and Total Minutes.
Unique Visitors: The number of inferred individual people, counted only once, within a designated reporting timeframe (usually a day, week or month, but may also be available by hour, week or quarter). This is a non-additive metric: Unique visitors cannot be added together over time, because one visitor can make multiple visits in the time frame studied. For example, a person who visits your site twice in the same day is one daily unique visitor who generated two total visits.
- Common error: Adding the month's Daily Unique Visitors does NOT equal Monthly Unique Visitors. I find analogies are helpful in explaining this, so let's imagine people who visit Grandma's house: If last month you visited her daily, during the month Grandma would have had 30 total visits, but just one unique visitor per day. (Poor Grandma. Round up some siblings or cousins, will ya?)
- Quick caution on the word "visitor": This is an attempt to count individual people, but is actually individual web browsers. In theory, if I visit your site once from home and once from work in the same reporting period -- or even twice from home, once using Internet Explorer and once using Firefox -- your analytics will likely count me as two unique visitors. Because of this, some organizations have opted to use the term "unique browsers" in their reporting.
Visit: An interaction with a Web site consisting of one or more page views. If an individual does not take any action on the site within a certain time frame (typically 30 minutes), the visit or session will "time out" (terminate). Remember, a Visit (activity) is not a Visitor (person): In our Grandma's house example from above, Grandma received 30 visits.
Visits per Unique Visitor:
Visits divided by Unique Visitors = the average number of times each person
came to your site during the reporting period.