I think most of you will agree when I say:

Building amazing products is pretty darn hard.


Because getting insights into what your customers want is not an easy task.

But, there is help!

Today, you will learn how to use five digital tools to gain valuable customer insights.

Insights that will help you to build products your customers love.

Here are the tools we will be looking closer at:

  • Google Analytics – The old trusty GA can actually be useful
  • Google Optimize – The new kid on the block, making A/B testing looking easy
  • Hotjar – Getting hot insights with their heatmaps
  • Teston – Video killed the data-driven stars
  • Don’t hesitate – Not really a tool, but nevertheless a very important mindset

For each tool you will learn about important features, how you can use it and a few examples I did with a site called Bilkjop (used car portal).

Google Analytics

I will start off with showing you how GA (Google Analytics) can be used.

I will assume that you have some prior knowledge of GA.

Most of you probably use GA out of the box, without any customization. Maybe you didn’t even know it was possible to customize GA.

But, it certainly is, and I encourage you to do so.

We will now dig deeper into three features of GA, that will help you gain customer insights and make GA a valuable tool in a customer-driven product development environment.

  • Goals
  • Segments
  • Events

There are of course a lot more useful stuff inside GA, but I feel these three are a good starting point.


One of the first steps is to set up goals in GA properly.

Now, in order to setup goals, you first need to think hard about what you want to measure.

What are good indicators of customer engagement? A submitted form? Time spent using your product? Clicking on email links?

There are countless different events you could measure.

These goals will be the foundation to measure if changes to your product were helpful or not.

Here I will share with you my four goals I used for Bilkjop.

Number of booked test drives

The primary target of the product is to sell more used cars.

The obvious first though is to enable a buying option on the website.

And, eventually, we will get there.

But, for now, we use the option to collect hot leads from the website by allowing them to book a test drive at the dealer ship.

This is the primary goal, and we measure it closely by tracking each step of the checkout funnel.

All the features on the website are in one way or another connected to this goal.

So, when I talk about customer-driven product development, it is to gain insights into what the customers want from a used cars website to complete our end goal – sell more cars.

Number of financial leads

The second goal we measure is how many leads we generate for our financial services. It is nothing fancy, just measuring how many clicks we have on a specific button. But, it is nonetheless an important goal, since the value of securing a car loan deal with our customers is high.

Users who visited at least one car

To even be able to book a test drive, a visit to a car page it a prerequisite. Therefore, we measure the users who visited at least one car.

Users who visited more than five cars

For most people, visiting just one car isn’t enough for us to find the right fit. Potential car buyers usually check several different options. Therefore, we measure the users who visit more than five different cars.

So, there you have it, the four goals we measure for Bilkjop.

But, wait?

Isn’t this merely reporting on numbers? How does this help me gain customer insights and find out what they want?

It will help you in two ways:

  1. You can use these goals to see how your product increments affect them
  2. You can use them together with segments

And lo and behold, next topic is about segments!


Segments is sadly an often ignored and overlooked feature of GA. Even though it is pretty “in-your-face” on almost all reports. Yeah, it’s that top bar saying “All Users”.

So, why is it so often ignored and how can you use it?

It isn’t ignored all that much if we think about it. Because when you look at how a specific traffic source behaves, you have already segmented your traffic. The thing I am coming to is that there are so much more to segments, other than traffic sources. And those segments are often overlooked.

GA comes with a pretty neat list of system generated segments. Now, the deal with segments is to compare them and see how the perform against each other.

For example:

  • Does mobile traffic convert better or worse than desktop traffic?
  • Does user who did a search visit more pages than those who don’t?

By asking yourself questions like these and analyzing the answers, you will find room for improvements. If, for example, the users who searched visited more pages (and if that is important to you) you should consider encouraging more users to search. You can do so by tweaking the search design, function and location of the search bar, and thus potentially increase the number of users who search.

To go even deeper with segments, you can create your own. A good start is to create a segment for each goal you have.

In my example, I created a segment for each goal. Then I can analyze what the converting traffic for each goal does differently than the non-converters.

I have also created segments for different traffic sources grouped together. This way I can easily compare different traffic sources, in a way that is meaningful for me.

So, the basic idea behind segments is to use the default, create custom ones and compare them against each other. Based on that you will have to get creative and think why they differ and if there is something you can do about it.

When you have implemented something new, you can track the effects by looking at your goals data.


One of the most powerful and useful features of GA is the use of custom events. Hey, they even created Google Tag Manager with custom events in mind.

Events are custom information snippets sent to GA when your users do something in your product that you have specified.

A most basic example could be that when users click a specific button, we can send that button name as an event to GA. You can then use this event and connect it to a goal.

A more advanced example is shown below.

Here we track click on each step in the checkout funnel for a test drive.

And here we track the requested time for the test drives.

As you can see, we can derive any type of information from the customer’s behavior by using custom events. And with GTM (Google Tag Manager) we can do quite a lot without involving programming resources. There are even packages you can install to GTM, where you get a lot of useful events out of the box.

So, as you can see, there is a lot of power in this feature and only your imagination sets the limit on the data you can create with custom events.

There you have it.

Three areas in GA that you can use to understand your customers better. Now, we will move on to another tool from Google and see how it can be used to create products that your customers love.

Google Optimize

Next up is the A/B testing tool Optimize from Google. With this tool, you can easily create A/B test on your website.

As with all Google tools they are tightly integrated with each other. And, there is no difference between Optimize, it is nicely integrated into Google Analytics.

With that integration and a straightforward interface to create A/B test, this tool is one of my favorites.

If you don’t know what an A/B test, it is when you test two different variant and compare them against a given target.

Let me show you two examples that we did with bilkjop.no

Text on CTA button

In our first example, we made an elementary test, two different texts on a CTA button.

Avtal visning VS Booke prøvekjøring (Two variants of saying “Book test drive”). We measured which one of these who generated the most button clicks.

So, what do you think? Which on performed better? Was there a clear winner?

Well, here are the results.

Now, what can we learn from this?

As we see, there was no clear winner. So we can learn that just changing texts and colors on CTA buttons probably won’t have any significant difference.

Many argue that to get real results we need to dig deeper into our visitor’s psychology and find out what their real pain is and how our product can address those. And, this is not simply done by changing word or two on a button.

Nevertheless, let’s have a look at another A/B test with drastically different results.

We wanted to test if choosing date and time for a test drive, and then enter your contact details converted better then the other way around.

What do you think?

Our theory was that it would convert better, since choosing date and time is a lesser commitment than entering your contact details first. And once your are “in-the-funnel”, you will continue, highly inspired by the microtransaction concept.

And guess what?!

Here we have a clear winner – presenting the calendar first is without a doubt the better performer.

Entirely different results than our CTA button text A/B test!

What are our takeaways form these results?

You will probably not find a clear winner by testing small changes like colors and texts. It seems to be better to test more extensive changes to your landing pages.

Think deep, and do customer interviews to find out what your customer’s pain points are and try to address them with a new design.

And test that new design against your old one.

Then you surely will see a clear winner instead of tweaking small things.


Hotjar is a tool for creating Heatmaps.

So, what are heatmaps and what are they good for?

Heatmaps tracks where your visitors interact on your website.

This can be good since you will learn what parts of your website are the most interesting for you visitors.

By using this information, you can rearrange or promote certain elements more or less depending on how much they are interacted with. You can generate heatmaps based on clicks, mouse movement and scrolling.

Here we have our move Heatmap.

These are some insights we can learn from this:

– The pagination links at the bottom are very popular. Maybe we should replace them with endless scroll.

– The sorting options are much used. Since this is obviously something our users want, we should make sure they are as good as they can get.

– The search bar is not that popular. We could replace it with something our users would rather have in that prominent spot.

– The filter “Forhandler” is popular, should be moved up.

– The filter “Karosseri” is not popular, should be moved down.

– The favorite function are also popular and should stay where it is.

In addition to Heatmaps, Hotjar offers a few other features. Here are the ones I believe are useful:

Polls: You can easily create a poll and implement on your website.

Recruiter: You can implement a popup form on your website asking for persons who are interested in testing your site. This feature can play very well together with Teston for video and screen share testing. You will learn more about Teston next.


So far, we have used data to interpret how our visitors use and think of our products. This is all very useful, but there is one information source then beats all that.

And that is.. drumroll… observing people and listening to their thoughts as the use your product.

This has traditionally been a quite complex, costly and time-consuming activity.

But, there are now options to digitalize this process as well.

We used a service called Teston where we can get videos of actual people and their screen when using our products.

It is all digital and works like this:

  1. Create a list of activities and question you want the test person to do and answer
  2. Actual people will do your test and record themselves and their screen
  3. Watch the videos and gain customer insights

And the result will look something like this. Here we see the tester’s screen and face. The numbers in the button bar are the shortcuts to the activities/questions in our test script.

My experience with this kind of testing is outstanding. Of course, it depends on the persons doing the test, but the quality was excellent for all the testers I got.

Watching actual usage of the services helps to identify pain points and improvement to the product.

One example was the price slider, which was not very user-friendly when searching for lower priced cars accordingly to our testers. They had problems specifying the from and to price by just using the slider. And when the from and to price were close, the two round slider indicators became one.

From this, we learned that we should improve it to allow the user the enter a from and to price manually.

Don’t hesitate

“Don’t hesitate” it not a tool, but an essential mindset that you should embrace to realize all the things you will learn from the tools we have discussed.

You must not be afraid to test new ideas and concept based on the customer insights you gain.

This will lead to some successes, but also some failures.

Don’t be afraid of that.

Instead, keep track of your data, test new ideas quickly and evaluate. And, if they don’t work, go back to what you had, or try something else. Keep trying, and eventually, you will find something that your customers love.

By adopting an agile framework like Scrum or Kanban, you will enable a setup that allows for failing fast. The basic theory is to come up with an idea, plan it, develop it and release it. Then keep tracking your data to see the results.