The ability to visualise your visitor data in map form has been possible within Google Analytics for as long as anyone in the team can remember, and we do have some veterans in our ranks. However, if you have used these reports you will understand that they don't provide you with the necessary tools to generate truly useful insight. This is surprising considering the importance for businesses to understand how their website is performing across regions, how your key metrics vary by geography.
In this post I explore the boundaries of geographical data visualisation in GA and the current limitations. Initially this is very much “in the box” thinking, providing custom solutions within the realms of GA. Later, I will cover external toolsets that can help you to realise the value of your geo GA data.
Google Analytics and Geovisualisation
Let’s begin with the basics. Google Analytics uses a user’s IP address to determine their geographical location. Within the Location report (Audience > Geo), this data is aggregated into tables and mapped. For technical and data privacy reasons, City is the most granular geographical scale available (note that towns and even villages are classified in this manner in GA). This data is visualised as a ‘Point’ map:
Sessions by City, UK
A selection of broader scales are also available for mapping: Region -> Country -> Sub continent -> Continent. For each of these, your data will be automatically visualised into a choropleth map – a type of thematic map which categorises and summarises regional trends into distinct areas through use of colour (see below).
Let’s do a quick summary:
- Continent & Sub continent: Expansive, but probably unlikely to generate actionable insight for your business in most cases. For example, you had 1,000 visits originating from Asia last month. Cool. But how are you going to use that information? You need to get more granular than this.
- Country: You might want to use this to identify which new language variants would be beneficial on your website. Unless you operate across many markets, this dimension probably isn’t going to be that useful either.
- Region: The most granular scale you can map your data to in GA. Here's what that looks like, looking at UK data in GA:
Sessions by Region, UK
In the UK, Regions are defined as countries – Scotland, Wales, N. Ireland & England, Isle of Man. Therein is the first problem. There are no options for data aggregation or visualisation between the ‘City’ and ‘Region’ dimensions. There is no sweet spot.
At this point I should explain that, if you’re lucky, you can get aggregated data by applying “Metro” as primary dimension in the Location reports, which sits at a level between Region and City. These are based on archaic UK TV regions such as “HTV West” and "Meridian", but we’ve only seen it working in some accounts. Feel free to get in touch with me if you know why. Anyway, even if you wanted to group your data into this geography, GA can’t map it.
Most of all, the customisation options are effectively nil, nada: if you are mapping by City, you only get a point map and you have no control over the grading of the data points. If you want to map by Region you are stuck with a generic choropleth map. There are actually no options for customisation at all, not even colour.
This post may quickly be devolving into an angry geographer’s rant about GA. But there is also good news if you’re looking to enhance the output of your Geo data.
1. Utilise Custom Regions
Many of our clients operate with their own custom regions which are applicable to their business. For example:
- London, North East, Midlands...
- Or: North of England, South of England...
- Or even: High value area, low value area.
These regions are of course not available in GA by default, but can be introduced and applied to enhance your GA data with Custom dimensions and the Dimension Widening feature. This is actually a relatively straightforward process and involves creating and downloading a schema from GA, matching this up with your own dataset (where each row will have a geographical label) and importing it back into GA. GA will attempt to use a common identifier – a ‘key’ – in order to match up the rows in your data. You'll be able to do this on all standard and GA360 accounts. Feel free to give us a call if you’d like some help with setting this up.
If you configure this, create some custom reports with the primary dimension (e.g. ‘Custom regions’) alongside whichever metrics are of interest to you. Alternatively, you can apply it as a secondary dimension within most of the standard reports.
2. Geographic Data Imports
Utilising the same functionality, it’s possible to incorporate third party data to enhance Google Analytics’ default geographic data, such as wealth, deprivation or home ownership.
I won’t go into too much detail since my colleague, Arran, has already written a comprehensive implementation guide. Check that out here.
As powerful and useful as this is, you still cannot map this data in GA. If ultimately you’re the type who finds more pleasure in a quality map over a table of data, you’ll want to read my next posts where I’ll take you through the options.
Don’t settle for the default.
I believe that the Web Analytics industry as a whole should be taking a more advanced approach to geospatial analysis. Google don't do enough natively in Analytics to allow you to leverage what is essentially a high quality, first party dataset. With a little creativity and a few implementation steps there's so much insight to gleam without needing to look at expensive, third-party geographical datasets.
If your geographical GA data is dormant but you feel like unlocking this potential, get in touch with us. Or drop me a line directly. I love maps.