Ahead of his speaker session at Performance Marketing Insights: London 2014, Andrew Miller, digital consultant at BlueGlass Interactive, explains how you can gain deep insight into customer behaviour with readily available software.
In the past two years analytics software has made massive strides in the levels of user data that you can accurately collect from your customers. More and more we are able to connect the dots between visits so that we are able to paint a picture of an overall customer, as opposed to clusters of visits.
Behavioural sequencing is a method of analysis where we analyse the interactions that users have with content, landing pages, marketing campaigns and channels in sequences to understand how that influences the user's propensity to convert and value as a customer.
The benefit of conducting this kind of analysis is that you can separate out and get a much better understanding of the impact of your marketing campaigns in regards to:
- New user/customer acquisition.
- Conversion & customer value influence.
- Up selling on conversions.
- Post-sale return purchases.
- Post-sale cross selling across departments.
In order to do this, it is important to understand that within Google Analytics there are 3 levels of engagement:
- A visit - a single visit on any individual page of a website.
- A session - a series of consecutive visits on pages of the property, connected by 1 cookie session.
- A User - Multiple sessions on a property, by the same user, connected by 1 cookie file.
Whilst this information may seem basic, a lot of analysis that is done on web performance still only takes account of users at a visit or a session level - effectively grouping together your most loyal customers with users unfamiliar with your brand.
The goal in this analysis is to separate out your customers into cohorts by their exposure to your brand and behaviour on your site, so if we wanted to see what kind of influence a company blog was having on user behaviour then we could set up the following segments in GA using the sequence options within the custom segments settings:
Segment 1 - New Acquisitions
Segment 1 is set to include users whose first interaction is a session where the landing page is on the blog section.
Segment 2 - Lands on the main site and then moves to blog
Segment 2 is set to exclude users whose first interaction is a session where the landing page is on the blog section, also to include users with any interaction in a session that includes a page on the blog section.
Segment 3 - No exposure to the blog
Segment 3 is set to exclude users who have any interaction with a page on the blog section.
Now by comparing these segments with each other, we can get an view of how users are influenced by the content that you produced (click to enlarge):
From this we can very quickly see that the blog is a pretty terrible source of customer acquisitions, users who come to the site via the blog do not tend to go on to make a conversion - making them much less valuable than other traffic.
The vast majority of traffic and conversions in fact come from users who have had no exposure to the blog whatsoever however users who have had some exposure to the blog at some point convert at nearly 50% higher! Presenting an entirely new opportunity for the blog as a conversion tool rather than an acquisition tool.
One of the things you cannot see above (but you can easily work out) is that not only does the conversion rate increase for users in the second segment but the average order value also goes up by 20% (£34.54 to £41.49).
So let us turn this into a forecast - if an additional 1,000 users that were initially exposed to the main site were to read the blog then:
- These 1,000 users currently have a value of £6,908 (0.2% conversion rate at an AOV of £34.54).
- Those users should start to convert at the higher conversion rate of 0.28% and at the higher order value of £41.49.
- These 1,000 users are now worth £11,617.
- This equates to additional revenue of £4,709.
This gives us quite an incentive to prioritise and look deeper into what kind of blog content is actually influencing users in this way (as opposed to simple customer acquisition). The above analysis is far too broad to give real insight so the next stage is to narrow it down and apply the principle on types of blog content that users are viewing, showing how messaging or content categories affect users and their propensity to convert in the same way as above.
Then with the results of this analysis you can start to optimise your content production with the consumer in mind, prioritising the posts that are pushed on the main site and also create new ways to market the right content pieces to users during their journey to purchase.
In order to conduct this narrower analysis your Google Analytics configuration may need some work to enable the grouping of content and campaigns properly. One of my recent posts on the BlueGlass.co.uk blog Tips to Configure GA to Show the Value of Content is a great start to go through some of the core GA features that need configuring to get this kind of data.
So there you have the general principle behind behavioural sequencing and using it to show value in different ways for your campaigns. The example that I have used above is merely scratching the surface of the different ways that you can leverage it for content, campaigns, channels and other activity but you can use it on your site to see a new way of how your blog is influencing your visitors.
There is more information to read up on sequencing and value attribution. I will be talking more about this at the Performance Marketing Insights conference in London, and will be give more examples and ideas of how you can use this principle, and also what you can do to value external campaigns in a similar way.