Website optimisation | 12/04/2018

Experiments in Google Tag Manager

Posted by Lawrence Greenlee

I love Google Tag Manager. I probably bring it up about three times a week and get a groan from the Cobb Digital team. “We get it Lawrence. Google Tag Manager is your first love.”

No, optimising websites is my first love. Google Tag Manager is my second love.

Hummus is my third.

What I love about Google Tag Manager is how versatile and easy to use it is – especially when it comes to collecting data. We love data, but your standard pageview tracking for Google Analytics doesn’t always do you justice. After all, you generally want users to do more than just ‘view the page’ – how do you know they are really interacting?

So, I’ve done some experiments.

UX Experiments

There are some widely accepted facts in UX. Things that are rules of thumb – but should not sway your hand.

One rule, that I come across frequently, is that human eyes gravitate towards warmer colours. Great, you might think, let’s make all our call to actions a warm red.

Can’t I get some better data?

A simple event tag in Google Analytics will be able to tell you which buttons on your website get the most clicks. And, using a variable that pulls the colour of the CSS file (not as complicated as it sounds!), you can see how many red, green and blue buttons are getting hits on your site.

Without having to pay for A/B software.

Form values

I have a conundrum. Due to the new GDPR regulations, I must ask people to opt in to my newsletter when they submit a form on my site: I can’t just assume their consent.

This means I must have an unchecked checkbox on every form of the site – and, frustratingly, I can’t see when a user signs up to my newsletter that way, with a traditional form submit.

The newsletter has some top-notch ROI, so I want to track people signing up to it.

Using some simple custom variables with GTM, I can pull whether the newsletter checkbox has been ticked or not! Giving me some clearer data.

And those B2B forms – the ones that ask someone to select a job title or their sector, from a drop-down list. How useful would that data be to pull into Google Analytics, AdWords and Facebook?

Domain changes

A client recently changed domain – and we supported with the usual best practices for solidifying organic traffic and optimisation. However, I wondered to myself:

Google is going to continue indexing and ranking that domain for a few days, at least. How will traffic through that indexed domain appear in GA?

I assumed that the referrer would be wiped when the traffic passed through a 301. Perhaps the user would look like they came directly to the website, as opposed to via a Google search. I decided to test it.

After the site went live, I opened my Google Tag Manager preview in my browser, found that link in the Google index, and got my answer very quickly.

As if by magic, traffic through a 301 DOES remember its own referrer! Brilliant!

HTTPS to HTTP referral traffic

Our fantastic PR team have secured coverage for a client on HTTPS:// – with a great link to the client’s HTTP:// We should start seeing web traffic coming in any moment…

If your site is on HTTP (unsecure) website, then you won’t be able to track traffic that comes from a HTTPS (secure) website.

That’s ridiculous!

What about all the legitimate traffic that is coming through to our site? And converting! Can I not report back on the success of online PR, if the news website is secure?

I tested it on Google Tag Manager. I found a site we look after, and that we have set up tracking for, and I replicated the user journey: opened preview, and saw…

I saw nothing. No referrer! Traffic from a HTTPS site to a HTTP site WILL come through as direct. Those links on aren’t going to be proving their value now.

What can we do about this?

Well: have you considered upgrading to HTTPS. Search engines like it, and it’s way better for your users.

Want to learn more about the many forms of information you can pull from your website? Partner with a Google Analytics Certified teamand maximise what you can get out of your data.