Call 0845 485 1219

Graeme Benge

An Introduction To Attribution Modelling With Google Analytics

23rd Aug 2013 Analytics, Analytics, Conversion Rate Optimisation | 4 Comments


Attribution ModellingYour site exists to serve a purpose and that will be to convert visitors in some form or another. With an in depth knowledge of your company’s user base and working knowledge of some very handy functionality within Google Analytics, you can get closer to the type of activity that resonates with your audience in order to do make them complete more key actions.

So let’s get started.

Credit Where Credit Is Due – Attributing Conversions To Marketing Efforts

There are likely to be several ways of considering how conversions were achieved by your marketing efforts however there are three real world conversion scenarios and we’ll start here.

  • Attributing conversions by channel
  • Multi-device attribution
  • Online to offline conversion attribution

Google Analytics offers analysis of the first scenario and I will explain how to play with the reports later on. First though, I’ll go over the different ways conversions are achieved.

Attributing Conversions By Channel

Attributing conversions by marketing channel provides you with the path to conversion labelled by visit source. For instance if someone searches using ‘ironicdomain’ and subsequently visits www.ironicdomainname.com from the SERPs, a subsequent conversion will be attributed to ‘Organic Search’.

Now the reality is that conversions are not always as straight forward as that. Conversions may have come off the back of several visits from that user by different means.

They could have found you off the back of an organic search and gone to your site. Then on a subsequent visit they remembered your site and returned by typing your domain into their browser.

Whilst carrying out some research at a later date, your reader might click on a link from within a review of your product on a third party site. At this point should the visitor convert, the Channel path will be:

‘Organic Search’ –> ‘Direct’ –> ‘Referral’

Found in the Conversion report under Multi-Channel Funnels then Top Conversion Paths, it will look a little like this:

20-08-2013 09-04-38

Direct, Organic Search, Referral, Social Network, Paid Search, Display & Other Advertising are default values. By adding customised channels you can add in extra data sets such as Brand/Non brand terms.

Let’s look a little closer at what the conversion path data can tell you:

The image above shows ‘Social Network’ as a consistent contributor to conversions, albeit without featuring in the top conversion path. This clearly suggests that ‘Social Network’ visits are a potent source. Straight away there is some reason to experiment with calls to actions, bolder triggers or different offers to see what sort of scale can be achieved.

However the most conversions were from visitors following a ‘Referral’ –> Directx2′ –> Organic Search path. Thus we can conclude:

  • The top converting path would appear to show that your content marketing efforts are being rewarded with effective links from third party results
  • Your branding is simple as people are happy navigating to your site directly
  • Whilst doing some final research, your SEO efforts (and lets admit it, probably Google’s customised search) have seen your site rank high enough to entice that click that has finally resulted in a conversion
  • Well played you

If you add in potential visits from Social Networks the combination of conversion paths grows considerably and also poses another question of which channel should receive the most credit for conversions or whether there should simply be an even spread across the board.

Multi-Device Attribution

Multi-device attribution taps into one of the big buzz’s currently: that of multi-screen content consumption and tracking the all-important conversions that occur consequently.

Smart phone and tablet adoption has been rapid. Handheld device usage to consume online media consequently has risen at a rate of knots so it’s conceivable that one person is visiting your site via a desktop PC, a tablet, a mobile phone and via their TV.

Attribution modelling here aims to assess the impact of each device during the course of the conversion path.

Online To Offline Conversion Attribution

Online marketing’s impact on offline events, such as sales or phone calls are harder to tie online participation to. But it is still important to be aware of their value as there are so many offline triggers out there that if they are not included it will skew any ROI measurement.

Coupon codes and bespoke landing pages should be mandatory for offline marketing be it newspaper advertising or direct mail in order to understand their online impact.

Model Comparison Tool

Now Google Analytics includes a Model Comparison Tool that offers a number of attribution models which provide data on an Attribution by channel basis. A pain, I know as we would love to review all three scenario’s together and in isolation. Sadly achieving that is complex and it’s own blog post!

Whilst there are six models to choose from there are only two that can be used for consistent insights (read: there are only two that really make any sense).

Linear Attribution Model

This model sees all things as equal, meaning each source on the conversion path gets an even share of the pie.

Linear Attribution Model

This is a blunt means of attributing value, however analysing the different conversion paths may throw up some useful trends that can inform your marketing strategy.

Time Decay Attribution Model

This model is far better as it distributes value through the conversion path but on a decreasing, sliding scale from the last interaction. This looks to be a better way of attributing value to the individual sources throughout the conversion path. In football terms, you’ll be better able to see who the play-makers are and who’s the fox in the box.

Time decay

Using this can highlight the channels that are consistently delivering or playing a part in conversions. It’s another way to make sense of what is assisting conversions

Conclusion

There is quite a bit to understand but an awful lot to gain from bringing conversion attribution into your marketing analysis. One opportunity it offers is a move away from the red mist of keyword targeting that is only going to get more difficult with the rise of (not provided).

Does knowing what drives your conversions matter to your business? Are you using attribution modelling to inform your marketing strategy? Which attribution models really deliver actionable insights? Let me know in the comments as I think this is going to become a vital part of any online marketing strategy if it isn’t already, especially with what Universal Analytics offers.

Image Credit:

Converting Leads To Sales via Big Stock

Share this post

About the author

Graeme Benge

With five years background in Travel, Graeme has built up strong commercial experience alongside online and offline marketing skills working with a variety of Travel Agents and Cruise Specialists. A passionate advocate of SEO and Social Media, Graeme has a strong interest in ROI and analytics in order to deliver the best level of returns.

The Practical Guide To Google Analytics

4 Comments

What do you think?

Digital Marketing Ideas Every Month

Sign up to receive our free monthly email. Including our favourite pieces of news from the digital marketing industry.

From SEO to PPC, Social Media to Brand Management and Analytics, we'll keep you informed.