If you’re planning to analyze websites, it’s only a matter of time before you’re thrown into Google Analytics and asked to look at the health of your site. UX designers do this differently than marketers, so today we will learn about the basics:
Designing with Data — Summary Statistics
(If you’re just starting the UX Crash Course: Start Here.)
****
There are 7 basic stats that you will need to understand if you want to do anything with a website. Not just what they measure, but what they mean.
Numbers are not good or bad: 1 million users is only good if it’s higher than last month. If you’re Facebook and you have 1 million users this month, you’re in trouble.
Before you analyze anything, take aminute to think about what your site is supposed to do, and realize what numbers would indicate that type of behaviour.
****
1 & 2) Sessions (Total Visits) & Users (Unique Visits)
This may surprise you, but Total Visits measures the number of times your site has been visited, total.
Unique Visits is a little different though. If I visit your site 50 times, I am only 1 unique visitor. And technically, it counts unique devices not “people”.
What they mean: Comparing these numbers will tell you whether your traffic is:
High quality — Total is much higher than Uniques.
High quantity — Total is about the same as Uniques, and Uniques are higher than last month.
Both — Uniques are higher than last month and Total is much higher than Uniques.
Neither — Uniques are lower than last month and Total is about the same as Uniques.
“Uniques” are a more authentic representation of your traffic, but I would rather have 1000 people who visit every day than 10,000 people who visit once a month. Then again, if 1 guy visits my site a million times, that’s not particularly useful either, and he probably has issues.
****
3) Pageviews
What it measures: One pageview is counted every time any visitor loads any page.
What it means: You can think of pageviews as a “general indicator” of traffic, because it describes the total amount of content being seen, overall, and it ignores most other factors.
If your site runs on banner ads this is an important number to watch.
If your site is content-driven, like news, then having a lot of pageviews overall might be more important.
****
4) Pages-per-Visit
What it measures: the average number of pages each visitor will see, each time they visit. You can also think of it as the “clicks” per visit, if that helps (although technically not accurate).
What it means: If your site is more focused on tasks or social interactions, then the Pages-per-Visit might be a better metric to watch than pageviews.
On the other hand, if you are Google Search, you want this number to be as low as possible, because good search results give you what you want on the first page.
****
5) Time-per-Visit
What it measures: how long each visit lasts (on average). It can be very important to compare Time-per-Visit and Pages-per-Visit.
What it means: In a perfect world, content-based sites want visitors to read the whole article, and read a lot of articles. That would be a long Time-per-Visit, and many Pages-per-Visit.
If people see a lot of Pages-per-Visit, but don’t spend much Time-per-Visit, they might be looking for something and not finding it (bad), or they might be completing tasks really quickly (good). It depends!
If the Time-per-Visit is high, but Pages-per-Visit is low, your navigation might suck (bad), or it might mean the articles are long and the users are really engaged (good)! It depends!
If both Time- and Pages-per-Visit are low, that might be a bad sign, unless the purpose of your site is to get in and get out quickly, like Google.
****
6) Bounce Rate
What it measures: the people that only see one page and then leave without clicking anything.
What it means: In general, we usually think of this like people saying “no” to your site. BUT… there are some exceptions.
Blogs tend to have a higher bounce rate, because they are designed for single-page viewing: you either scan the list of recent posts, or you came for a specific post.
The structure of the site and the source of your traffic can have a huge effect on your bounce rate, so even though it looks simple, bounce rate can be a complex indicator.
****
7) New vs. Return Visitors
What it measures: If a person (actually a device) has been to your site before, they are “returning”. If not, they are “new”.
What it means: Returning visitors know more about your site, so they often “bounce” less and see more pages. If they are returning they must like what you do, so they often spend more time on the site as well.
However, “new” visitors are good, because it means our site getting exposure to more people.
The main thing is the ratio of new to returning. If you have no returning visitors then your site is either new, or terrible, because nobody is coming back. If you have only returning visitors then your users might be loyal, but the site is dying, because nobody new is finding it.
And, in general, the more mature a site is, the more return visitors you will have — as a percentage — because return a lot of visitors who come back often are hard to beat with normal search traffic or ad campaigns.
****
Tomorrow we will learn how behavioural data looks when you make it visual: Designing with Data — Graph Shapes.