Media Attribution with Google Analytics

[admin]

We’re all used to measuring campaigns using either a ‘first’ or ‘last’ click methodology. That is all conversions be it sales, quotes, online form completion etc. are tied back to the last (or first) channel that a person clicked on before converting. However, this method ignores all other touch points that a consumer had before the actual conversion which results in the undervaluation of media channels and the misallocation of marketing budgets. It also more often than not ignores organic channels such as SEO or Typed / Bookmarked that also play a significant role in the consumers conversion funnel.


By capturing all touch points a consumer has with media channels in the lead up to a conversion you can accurately identify the true value of each channel and its role in the consumer purchase cycle. For instance, some channels are great product introducers, such as unbranded paid search and banners, all are influencers to some degree (channels that sit in between the first and the last touch) and finally there are closers; ones that generally sit before the final conversion; Branded SEO, Typed/Bookmarked.
Using custom Google Analytics code we were able to record all touches a consumer had and record them in a ‘stack’. We did this using a Custom Variable and writing the first and then every subsequent channel into the one cookie – this then allows you to extract every channel sequentially prior to the final conversion event. We end up with a Google Report that looks something like this:


And an export file containing values like: sem:unbr>sem:unbr>ban>edm>seo:br>34
This indicates that 34 conversions occurred after a consumer had clicked on an Unbranded SEM term, another unbranded SEM term, a Banner, an Email and then a Branded SEO term before they converted.
There are many successful attribution models; below are some of the more straightforward ones (although still very effective).

For demonstration purposes I’ve pulled a simple version of the data using 5 channels; SEO, SEM, Referrals, Direct (Typed/Bookmarked) and Emails. However these reports can contain any granularity that you record; Branded vs. Unbranded Search, Keywords, Campaign names etc. By taking that stack data and visualizing it, you can creative meaning out of the attribution models.


Using the simplified example above we realize that SEM plays an important role both as an Introducer and a Closer, but isn’t an Influencer. (My bet is that unbranded terms would more likely be ‘Introducers’ and Branded terms would be ‘Closers’). Conversely, the importance of organic channels such as Direct and SEO have been understated in reporting to date as their role as a Influencer is proportionately very high.
As a marketer you need to know this – investment in organic channels is important, not only because these channels are generally more cost effective, but we’ve shown that they play an important role in the consumer conversion funnel; influencing purchase between the product awareness phase and the final purchase step. We can then assign an actual value to these channels – again, there are many attribution models that are both simple and effective.


Dependant on what you’re business KPI’s are you may weight conversion success towards a particular end (Introducer or Closer) or you may simply chose to assign the value evenly across all steps – either way, you end up with a more holistic approach to determining the value that each media channel plays in you overall marketing mix. For more information on Multi-touch Media Attribution with Google Analytics contact insights@datalicious.com










See and download the full gallery on posterous




Permalink | Leave a comment »



More from the Datalicious Blog...