Google Analytics metrics for dummies (and clever people): Part 1
Hi, I’m Emma and you may remember me from previous blog posts such as How to study for the GAIQ exam and How to calculate your Click-Through Rate using Google Webmaster Tools. Today’s blog post is not a ‘how to’ but a ‘simple guide to.’
Part of my job is to offer training on how to use Google Analytics. I accept questions about Google Analytics from team members, clients and whoever wants to ask questions. There are no stupid questions. I was hesitant to put “for dummies” in the title of this post because a) it’s cliché and b) I don’t actually think you’re a dummy for not knowing.
I always give people an example of the most basic question I’ve ever been asked:
What is a ‘visit’?
Good question. It sounds pretty obvious. “That’s when someone visits your site, right?” Well, yes, but what makes something into a visit? It also might be worth noting that visits are now known as sessions in Google Analytics. I apologise in advance for using them interchangeably or both at once.
When does a visit/session begin and when does a visit/session end?
A session begins when somebody comes to a page on your website that has your Google Analytics tracking code on it, from a page that did not have your tracking code on it. The session ends when they then go from a page with your tracking code to a page without your tracking code (closing the browser counts).
- Someone is on your website at midnight. At midnight, magic happens and they begin a brand new session. Otherwise they would have the same session twice on two different days and data would be strange.
- Someone is inactive (not clicking anywhere) on your website for a period of time (usually 30 minutes, though this can be changed). When someone picks up after a period of inactivity it becomes a new session.
Now we have a good idea of what a visit/session actually is, we can move on to more metrics. Something that commonly puzzles people is the difference between three metrics that are related and bandied about in tandem.
What’s the difference between a visit, a pageview and a visitor?
This is easiest to explain with an example.
Edward and Gordon both visit your website. Edward looks at five pages and is able to find what he wants. Gordon looks at two pages, gets frustrated because he can’t find what he wants and leaves the site. He calls Edward who explains how to find what he wants. Gordon goes back to the site, visits five pages and finds what he wants.
Gordon (first visit): P1->P2
Gordon (second visit): P1->P2->P3->P4->P5
We can see here that there were
Two users (or visitors)
Three sessions/visits (because Gordon visited twice)
Five unique pageviews (because only P1-P5 were visited between the two of them)
12 pageviews (because in total, between the two of them, they loaded pages 12 times)
Users/Visitors/Unique Visits = The number of people who visit your website
Sessions/Visits = The number of times people have visited your website
Unique Pageviews = The number of different pages viewed on your website.
Pageviews = The number of times pages were loaded on your website
Once you understand the difference between these, you can understand this one:
How does Google Analytics distinguish between new and returning visitors?
The answer to this one is “easy if you know the answer”, which doesn’t help if you’ve never thought about it before.
Google Analytics recognises users by the cookies they have on their computer. Think of cookies like a passport:
If someone has visited your website before they get a stamp. If Google Analytics has never seen their cookie before they have no stamps and they are a “new user”. If Google Analytics has seen their cookie before they’ll have stamps from your website and are a “returning user”.
Similarly, cookies can expire if they aren’t used often enough (i.e. by revisiting a website) and when a visitor gets a new cookie, it won’t have any stamps on it.
Speaking of stamps…
How does Google Analytics know how long someone is on a website or a page for?
The simple answer is “through using timestamps”.
…when does Google Analytics use a time stamp?
Any time someone loads a page with a Google Analytics tracking code on it, their time is ‘stamped’. The first page that is loaded will have the timestamp 00:00:00.
Let’s go back to Edward and Gordon.
Edward was able to find the answer to his question because he spent time looking at the website and reading. Gordon was much more impatient and left the website before he could find an answer.
So this would be how Google Analytics would stamp their visits.
Edward: P1 (00:00:00) -> P2 (00:01:30) -> P3 (00:04:05) -> P4 (00:05:00) -> P5 (00:05:30)
Gordon: P1 (00:00:00) -> P2 (00:00:30)
Each time shows what time they landed on that page at. Google Analytics can only calculate how long someone was on a page if they then went to another page because it requires the timestamp of the next page.
So the actual times they spent on each page is as follows:
Edward: P1 (1min 30s) -> P2 (2min 35s) -> P3 (55s) -> P4 (30s) -> P5 (unknown) (Total time on site: 5min 30s)
Gordon: P1 (30s) -> P2 (unknown) (total time on site: 30s)
Because Google Analytics does not know at what time Edward and Gordon left the website, it cannot understand how long they spent on the final page. This means that when Google Analytics shows you ‘time on site’ it is only showing you the time it knows was spent on the website – it is always missing the time people spend on the final page.
This is particularly troublesome if people only view one page. Google Analytics has no frame of reference and records both Time on Page and Time on Site as 00:00:00.
This leads us neatly to the next topic.
What is a bounce rate, exactly?
Bounce Rate is something that many people who report on websites use to report how well a page is performing. They know that a low bounce rate is good and have an idea that it has something to do with the amount of time or number of pages spent on a website.
What is a Bounce? On your website, when someone gets there (lands on a page) and they don’t like it, they leave without seeing any more pages because “this sucks” (i.e. – bad user experience). A bounce is when someone only views one page.
Bounce Rate is the percentage of sessions in which only one page was viewed.
I think that’s enough for now. Class dismissed.
Stay tuned for Part 2 in which I talk about goals, conversions, ecommerce and the other ‘value’ metrics.