Attribution is and no doubt always will be a hot topic in the search marketing world. Given that it is subject to unannounced but game changing updates from Google, and the fact that a healthy percentage of business owners still hold the belief that seo is hokum, proving the value of a search marketing campaign can be tough. This is where attribution modelling can play a pivotal role in determining which elements of your campaign are producing the most ROI.
Google Analytics comes with a set of attribution models already created for the user to apply to their data, with the Last Interaction model set as the default. In this model, it only takes into account the source of the last click to the site prior to a conversion (typically a sales or important goal). So in the example below of the linear model, social networks have contributed no sales.
Yet the site has received 101 visits from social media sources, which accounts for 0.5% of the total traffic.
This in itself could raise the questions within the wider context of a search marketing campaign such as, what is the current spend on social media and is it an opportunity that we are missing?
If we opt for the Last non-direct Click conversion rate model then it is shown, as in the image below, that social networks have played a role in conversions.
In this model all direct visits are ignored and instead attribution goes to the source used before the direct click. So 3 of the clicks that were previously allocated to direct had visited the site via social media and then returned to make a sale (the default time setting for prior attribution models is 30 days). This would suggest a conversion rate of 2.97%, for clicks garnered from social media sources.
In the interest of further investigating the default attribution models in Google Analytics the table below shows data from the linear model.
The Linear model gives each source that is included in the conversion path with the 30 day period equal credit. So imagine if a user initially hears about the product via social media and visits the site (but gets distracted and closes down the search), then later that day they remember what they were looking for and so use Google search (organic) to re-find the site based on the brand name. Imagine they then have a browse and decide that they want to ask their partner what they think of the potential purchase and so email it to them. In the final session, they just click on the link in the email thread that they created and so enter the site and make the purchase via a direct source. The linear model would allocate each of the 3 sources 33.33% credit. In the last interaction model, direct would receive 100% recognition and in the last non-direct model, organic would receive 100% credit. To bring it back to our example, you can see that the linear model reduces the number of conversions from social networks from 3 to 1.5 and with it the overall social media conversion rate to 1.49%.
This brings us back to the question of how attribution can benefit search marketing campaigns. While no single attribution model provides an exact picture of the conversion data, by analysing a number of models with a single question in mind, you can draw a number of pertinent conclusions.
In our example, we have identified that social media networks are only a small traffic source, but even with the linear model they attain a 1.49% conversion rate. When this data is used in conjunction with the information that the company rarely updates their social media platforms, you could make a case for utilising some of the search marketing budget, into creating tailored social campaigns focused on driving return traffic to the site (perhaps in the form of competitions).
For more information about attribution modelling within Google Analytics and how they can be used for your site, contact Digital State Marketing.