Why the web metrics you’re measuring are wrong

As an agency, we spend an awful lot of time observing and then accommodating (or attempting to modify) user behaviour. Google Analytics is an invaluable tool when it comes to tracking all this – but we’ve noticed that clients often aren’t taking full advantage of the service and the web traffic-related stats it can provide you with.

Jamie Goodwin, Senior Conversion Consultant in the Conversion & Optimation team here at Code, explains how you could well be relying on all the wrong metrics – and what exactly you should be measuring instead.

Can you describe the problem as you see it, Jamie?

It’s not necessarily that clients are asking us to measure completely the wrong things, but that they misunderstand what they’re asking for in the first place.
It often feels like this misunderstanding is based on them having been misinformed by internet chatter around the subject; sadly, there are lots of people out there putting forward views on what the most valuable web metrics are, but who frankly don’t completely understand what they’re talking about!
Also, there are usually lots of things that clients don’t ask for that they probably should…

What is the most common measure that clients ask for, and why is it not always exactly what clients think it is?

Bounce rate comes up a lot, and ‘reducing the bounce rate’ is often referenced as an objective or KPI for clients’ website. But this is actually a very specific measure that I think some people misunderstand.

People talk about it as if it’s a measure of when someone visits a page and doesn’t interact with it (i.e. leaves the site), irrespective of where they’ve come from. However, if they’ve come from another page on your site then that’s actually the exit rate.

What bounce rate actually is, is a measure of the rate at which people land on your site at a particular page (i.e. it’s the first page in their journey, their landing page) and then leave without further action. It’s not necessarily a bad thing to be measuring – but it should be done for the right reasons and in the right way.

It’s useful in helping you establish the quality of your landing pages, particularly how good they are at getting new customers to stay on your site. If you take the same page and compare different traffic sources, the bounce rate metric can help you gauge the quality of your traffic too. For example, if the landing page doesn’t change but your Facebook ads’ bounce rate is much higher than your Twitter ads’, you might consider reducing your Facebook spend and investing more into Twitter as it’s better quality traffic for your site. Alternatively, you could consider creating a landing page tailored to your Facebook ads and see if that improves the situation.

Are there other measures that people frequently misunderstand or use incorrectly?

Dimensions & metrics are the core of Google Analytics but I think a lot of people struggle to grasp how they work together, particularly when they’re new to Google Analytics.

Filters and segments are a little complicated too; filters are applied permanently to your data as GA processes it, but segments are applied as you view your reports so don’t make any permanent changes to your data. Filters can be a great way to tackle sampling issues if you have lots of traffic, (see the example below), while segments are really good for comparing different types of traffic.

Sampling seems to confuse some people. Because they’re providing the service for free, Google will sample your data if you have large amounts of it. That sounds reasonable to most but then imagine you have a dedicated sub-section for landing pages and you send all sorts of traffic to it (all your advertising and pop ups, etc.); it all has a very high bounce rate, but you’re okay with that because of the nature of your campaigns. But if you’re trying to run a report on the average time on site, it’s going to be massively skewed by the data surrounding those landing pages. With filters, you could create a ‘view’ that removes these landing pages, allowing your other reports to be based only on your users who aren’t from your accepted high-bounce traffic sources.

As mentioned above, exit rate is the rate at which people leave your site on a particular page (so every page has an exit rate). But you wouldn’t really want to look at the exit rate of a traffic source because in theory, every session has to exit at some point. It’s good for telling you which pages are losing the most of your engaged (i.e. non-bounce) traffic, though.

What are the measures you think clients should be using?

Goal conversion rates are great when your goals are setup correctly. They’ll tell you about the things that matter to your business; you have to set your goals up in advance though.

Device split is a great way to see problems users are encountering on different devices; it’ll show you things like bounce rates and goal conversion rates by device.

Events are good for monitoring things that take place on your site on pages, e.g. watching videos, opening hidden content sections, browsing FAQs, etc. Google Analytics runs on each page load by default so you need to manually set up events (Google Tag Manager is excellent for this, and Code have Tag Manager experts on hand to help if you need advice).

Goal funnels, when set up correctly, can show you where people going through your funnel (e.g. checkout funnel) drop off and help you identify where to optimise.

Custom dimensions are a great way to add your own categorisation to the data. For example, if you run a blog and have authors, you can add an ‘Author’ custom dimension to see how different authors’ content performs.

If you had one piece of advice for clients what would it be?

Take the official Google Analytics course here so you really understand what you’re looking at and its value.