Key Metrics
Google Analytics measures and reports on a lot of information.  Much of it is useful, but there are key elements that will be more important than others.  Since you truly can make statistics mean whatever you want them to mean, it is important to use the right metrics in order to draw meaningful conclusions.

The following list of metrics describes those which we have found most important and how they can be used.

Pageviews
Pageviews refer to the absolute total of pages downloaded.  This information is very commonly asked for, but in truth is important only in setting a baseline of site traffic and in making relative comparisons in-site. For example, it can allow you to track improvements based on design tweaks, identifying underperforming pages, etc. Otherwise it is considered a “vanity metric” that really doesn’t tell us much in and of itself.  The number of hits doesn’t matter, for example, if nobody is reading the page once they get to it.  Pageviews may also represent someone clicking around your site looking for one thing but not being able to find it.  Always look at pageviews in conjunction with other metrics such as bounce rate, time on site, and unique visitors.

Sessions
Sessions are generally perceived to be more useful than pageviews.  Sessions can be defined loosely as “user visits.”  A session may include one page view or several, depending on the user’s actions.  It is basically a measure of how many times someone comes to your site looking for information.

Unique Visitors
The unique visitors metric gives you the reach for your site – how many people have visited one or more of your pages. This is another borderline vanity metric, but can give you a realistic idea of where you should be setting your expectations and a way of measuring improvement. Note that the number will not give you a completely accurate reading.  For example, one person who visits the site on two browsers on their desktop and then on their phone would be seen as three unique visitors.  Removing cookies from your browser and then visiting the site would also cause the viewer to be seen as a different unique visitor.

New/Returning Visitor Ratio
Your ratio of new to returning visitors can be important, but requires a bit of business-related analysis when analyzing the numbers.  The ideal ratio will differ depending on the nature of your site.  A public informational website, for example, would expect to have a much higher percentage of new users than an internal site with a limited audience.   In general you would use this metric to see the balance of who is coming to your site, and thereby allows you to know what audiences you might need to target in order to achieve specified goals. A healthy amount of return traffic is good, but (at least for public facing sites) new visits are going to be what extends your reach.  A good comparison is running a restaurant – you need your regulars for a base revenue, but you will only grow by attracting new visitors.

Another key way you can look at new vs returning visitors is to set Segments for both of them and look at how behavior differs.  You can often get insight into actions or behaviors that your returning visitors take that new visitors don’t.  Knowing these differences, you can streamline your process to make it easier for new users to carry out those behaviors.

Frequency & Recurrence
This metric is a bit hidden in the interface, but can provide some insightful information.  It indicates the number of sessions and page views for users who have visited the site a given number of times.  That is to say, the number of sessions/pageviews that have come from users who have visited the site once, who have visited the site twice, etc.  This is a good indication of whether you have a lot of one-off traffic that never returns or whether you have dedicated users who visit the site repeatedly.

Days since last session
Also rather hidden, this is an extension of Frequency & Recurrence.  It lists the number of sessions and pageviews by users who have visited the site a given number of days ago.  Hence a high number on the “0” row indicates high traffic today; on the “1” row indicates traffic yesterday, and so forth.  This report can show both how regularly people visit your site and if there is any pattern for its periodicity.

Bounce Rate
Bounce rate is one of the most important metrics collected, and you absolutely need to be looking at it regularly.  It indicates the number of visitors who visited only one page on your site and never clicked through to a second page.  This gives a view user’s levels of engagement with the site.  A high bounce rate should be a red flag that your site is failing in some way to engage with the user.  The exception to this would be tiny, and especially one-page sites.  If there is not a second page for users to visit, bounce rates would be expected to be high.

Pages per Session
This is another way of measuring bounce rate and user engagement.  You will have to interpret the numbers based on your individual site goals and expectations.  A small site without  many pages would expect to see a value close to “1,” while a large site with lots of informational pages should be seeing a significantly larger value.

Time on Site/Time on Page
Ideally this metric is another way of measuring engagement.  This tells you how long users took to read your content. For sites with high bounce rates, though, this becomes problematic.  Time on Page is measured by the difference from when a person loads Page 1 and when a person loads Page 2.  If there is no second page load, though, there no measurement for that visit.

Referring Source
This is a vital piece of information, particularly when you have crossed the line from “looking at the numbers” to using analytics to improve your site traffic.  It is important that you know where your traffic is coming from.  Looking at the referrer – both in terms of channel (search, direct, referral, social, etc) and in terms of specific referring page – to see who is sending you traffic.  Analyze that page’s traffic as well if you control that site.  Look at all of your data broken down by channel to see how one medium’s patterns differ from others.  This will give additional information on areas to target to improve performance.  Your pages’ largest referrers are the ones most important to your site, so it is often helpful to form a relationship with the owner of those referring sites.

Queries
For most sites, search is the primary channel which visitors use to get to the site.  It is important, then, to know what terms they uses in their search.  Google Analytics provides some of these, and Webmaster’s Tools can provide more.  We are limited, though, by the growing “not set” value that is quickly becoming the largest part of the collectable data.

Conversions
The setting of goals should be a part of any site’s design process.  Analytics makes these goals measurable.  The conversion rate for each goal gives you a measurable indication of how well your site is meeting the specific needs which you have identified as being most important. A conversion may be as simple as tracking the number of people clicking from Page-X to Page-Y, or it can involve a longer chain of links creating a traffic funnel.

Campaigns
If you have any sort of email or online advertising campaigns you should be linking them to Google Analytics so that you can track how successful they are.  This is done through the use of campaign tags that are added as query variables in the link URL.  Google even offers a URL builder to help you create the links.  Proper tagging will allow the incoming traffic to be accessible both through the Campaigns menu within Google Analytics and also in the Channels and Source/Medium areas of the Acquisition report section.

Mobile
Connectivity over mobile devices has made it imperative for us to have mobile friendly web pages.  The mobile metric will show you the breakdown of your site’s traffic in terms of desktop, mobile, and tablet traffic.  You can then use this to show how many people are visiting your site on these devices and so to justify spending more time on mobile displays.  If you want to dive deeper into the report, you can also see traffic breakdown by device type (Android vs. iOS) and even specific device make and model.

Demographics
Depending on the nature of the site, you may want to analyze demographic information that gets included within Analytics.  Because Google Analytics can pull in information from Double Click and other marketing cookies, it can give you a general idea of the age of the person viewing your site, their gender, and their geographic location.  Keep in mind that all of these numbers will be approximations, but they can at least give you an idea of whether you are hitting your demographic targets.

Segment
Once you have mastered the basic reports, you are then ready to use segments to supercharge your ability to get an even finer level of detail.  In essence, a segment is a filter that you can apply on top of all of the standard reports.  For example, you may want to apply a “new users” segment.  This would then allow you to look at all of the analytics reports – sessions, demographics, bounce rate, etc. – but only include traffic coming from new users.  Segments are incredibly powerful at pulling out specific information that you are looking for, but do generally require that you know ahead of time the types of data that you want to see.

Scroll
Moving outside of Google Analytics itself, many heatmap programs now measure not only where users click but also how far down the page they scroll.  This gives important information on how long your content should be, page elements that affect abandonment, and placement of content. It is important to make sure content that you want the majority of people to see is placed far enough toward the top of the page that people are actually seeing it.


Annotations
Annotations offer a simple way to track notes in the Google Analytics reporting interface by date, so you can mark important events based on that may have impacted your data in otherwise seemingly inexplicable ways. In this way, it can explain reasons for jumps in the data to your entire team (if shared with everyone) that otherwise may be unclear.