This is the third post in our “Google Analytics Fundamentals” series at analytics-for-humans. In this series, we will (with a focus on SMBs) teach you how to distill actionable business insights from your Google Analytics data.

In our last post, we showed you how to identify the right dimensions, metrics, and units of analysis in your Google Analytics data to answer your business questions.

In this post, you will learn:

1. Basic strategies to convert Google Analytics data into actionable business insights

2. How to identify your website’s “core audience” with just four groups of metrics

3. How to create good hypotheses if you don’t yet have an identifiable core audience

4. Understand how to step beyond Google Analytics to validate your hypotheses

If you’ve ever used Google Analytics, you’re probably familiar with the dashboard shown at the beginning of the article.

At first glimpse, this dashboard seems pretty straightforward in showing you who your website visitors are, their behavior, and how valuable they are for your business.

It is often difficult, however, without analytics expertise, to step beyond the basic information provided by this dashboard and uncover insights that can help you not only understand how your users behave on your site, but also (1) why your users behave this way, and (2) what action to take based on these insights.

That’s where this article comes in.

In this post, we will provide you with a simple framework for analyzing Google Analytics data and generating testable hypotheses, so you can not only know something about your visitors but also immediately do something about it — i.e. improve their level of engagement and conversion on your website.

Let’s begin.

How to strategically approach basic analysis in Google Analytics

Before diving into how to read these dashboards, let’s first briefly discuss some guiding principles for analyzing Google Analytics data.

Essentially, the entire data analytics process in Google Analytics can be boiled down to comparing certain metrics (e.g. time on page, number of sessions) across different dimensions (e.g. age, gender, geography) to identify different user segments, or groups.

For example, if you want to understand how user engagement on your website differs by age (i.e. by a certain dimension), you need to first divide your data into different age groups.

You can then compare the bounce rate and average session time (i.e. metrics) of these age groups.

Think of your analysis as putting together a jigsaw puzzle that, when assembled, shows you a picture of the demographic profile and behavior of your site visitors.

For example, maybe first you discover that the gender of your ideal customer profile (ICP) is female.

Then you learn that she is a millennial between the ages of 25 and 35, who lives in cities like New York with an average income of $65,000.

The more slicing and dicing you do on your metrics and dimensions, the better you will be able to understand your customers.

If you are wondering what metrics and dimensions Google Analytics offers, you can find a complete list of them (with explanations) in our previous “Google Analytics Fundamentals” article.

How to read the Google Analytics dashboard to identify your core audiences

Now let’s look at some ways to identify your core audiences.

The process is in fact pretty simple. The goal here, when analyzing site visitors, is to identify high impact user groups, who have one or more of the following attributes:

• Most Frequent Visitors (High total sessions): This group visits your page most frequently

• New Visitors (High percentage of new users): This group is your emerging user group, where many of your new visitors are coming from

• Most Engaged Visitors (Low bounce rates, high pages/session, high avg. sessions duration): This group is most engaged with your website

• Highest Converting Visitors (High conversion rate): This is the group that most frequently accomplished your business objectives

If you find a user group that meets all 4 of the above criteria, congratulations! You’ve identified a “core audience” — the users who contribute the most visits, engagement, and conversions on your website.

All of your marketing activities (including your website design) should be targeted, tailored, and optimized for this core audience (you can have different core audiences for different product lines).

However, if you only find user groups that satisfy one, two, or three of the criteria, you’ve identified a potential core audience (especially if a conversion is one of them).

To see if they are truly the core audience you should be targeting, you should create a hypothesis for why they don’t meet all four criteria.

How to create a great hypothesis if you cannot find your core audiences

In order to nurture your potential core audiences, you need to create some hypotheses for what is holding these user groups back from having the optimal experience on your website (i.e. having maximum sessions, engagement, and conversions).

Here are some sample hypotheses that you might come up with based on your data:

• If user groups are both the “most engaged visitors” and “highest converting” visitor groups but does not belong to the “new visitors” group, a valid hypothesis might be that you are focusing too much effort promoting outside of your core audience group (i.e. users who meet the four criteria). You should, therefore, refocus your marketing efforts on the group that is actually bringing in revenue for you.

• If your specific group is one of the “highest converting” groups, but not the “most engaged visitors” nor “most frequent visitors”, a valid hypothesis would be that your highest converting visitors are buying your product or service even though their website experience is not optimal. If this is the case, you should redesign your website to drive even higher conversions from this highest converting group.

You get the point. Based on the interaction between these different variables, you can generate multiple hypotheses for why your customers are not engaging or converting on your website as much as they could be.

You can only create your hypotheses on a case-by-case basis, but you should follow these guiding principles:

• Pay attention to interactions between multiple variables: you may be unable to come up with a hypothesis if you only observed that the total sessions are high and the conversion rate is low for a specific user group. However, if you add the fact that the same group also has a high bounce rate and low session duration, it becomes much clearer that bad on-site experience may be driving the low conversion rate.

• Use common sense and qualitative knowledge: it may surprise you, but Google Analytics experts are humans too. You know your past and present business activities better than anyone, so draw connections between what you see in your data and your previous business decisions. Think about what you and your sales team know qualitatively about your website visitors and customers, and consider how these factors might influence a specific metric. For example, perhaps a spike in the number of sessions for a specific user group can be explained by a recent Facebook ad campaign that targeted that demographic. Of course, correlation is not causation, but don’t ignore what you know about your customers from experience.

• Compare across time: maybe the phenomenon you have identified is just a short-term spike caused by some marketing campaigns. Therefore, it is important to look at metrics with a long-term time horizon (a minimum of 1 month). If possible, also look at the differences across different periods.

• Be careful with small data points: If there is a user group that only has 1 total session but a 100% conversion rate, any data on this group is meaningless since you don’t have a large enough sample size. Set a threshold for the minimum amount of sessions a user group needs to have to be considered valid for analysis (Google’s Guide of filters is hard to understand, so I might write an article in the future to unpack this topic — comment below if you want it so I can move it up on our blog queue).

How to go beyond Google Analytics to validate your hypotheses about your website visitors

Having these hypotheses is a great start, but we need to validate these hypotheses.

Here are a few classic methods you can use to really test your theories about your website visitors and customers.

• Surveys: The benefit of surveys is that is it one of the most cost-effective and quantitative ways to understand your customers. The drawback is that you need to not only collect more data (which takes time and resources) but also learn survey design principles to make sure the answers you are collecting are valid. This is usually harder than it sounds. You can read more about making good surveys here.

• Customer Interviews: One of the best ways to understand your customers is to just ask your customers some questions over the phone. In most cases, they will tell you insights about their needs that you didn’t even think of. However, users often don’t know what they want or why they want it, and what they say is often different from what they do. That’s why you have to consider conducting user testing or running some experiments. For more information about coming up with good customer interview questions, reference here.

• User Testing: With user testing, you can actually see them walk through your website while talking through their reactions, validating or invalidating many of your hypotheses. You can “hire” your actual target demographic to look through your website and give feedback using services such as usertesting.com. Find out more about user testing practices here

• Experimentation: Another straightforward way to test your hypotheses is to just run an experiment. If you think your user is not converting because you have a horrible shopping cart page, present multiple improved versions and see whether any of these improved versions improve your user conversion. You can use tools such as optimisly.com for running your own experiments. Learn more about A/B testing and business experiments here

Using the results of these methods, you should increase engagement with your potential core audiences by modifying your design elements, messaging, and/or marketing channels.

Track these customer segments regularly, and over time you will be able to convert them into your core audience in no time.

This piece serves as the foundation to prepare you for drilling down into specific dimensions related to your visitors such as Age, Gender, and the channels they visit your website thru.

In the next article, we are going to start talking, in detail, how to understand the demographics, geographic locations, behaviors, and devices of your users using the principle explained in this article. Stay tuned!

This article was produced by Humanlytics. Looking for more content just like this? Check us out on Twitter and Medium, and join our Analytics for Humans Facebook community to discuss more ideas and topics like this!