With an average of five stages before a final conversion, and the inclusion of various devices, complex consumer pathways (the combination of ways that consumers travel to your website) are established.
This is where attribution models come in. According to Google, attribution models are: "the rule, or set of rules, that determines how credit for sales and conversions is assigned to touch points in conversion paths".
Various attribution models exist which assign credit to different clicks in the pathway, and depending on the industry of your business, the model that you would use would change.
The types of models we will explore in this post will be: Time Decay, Position Based, Linear and Data Driven.
Currently, many Google AdWords users use a last-click attribution model, which places all of the credit on to the last click that was made before the conversion.
There are three branches of last-click attribution, these are 'Last Adwords Click', 'Last Non-Direct Click', and 'Last Interaction'.
The Last AdWords click is what it says on the tin, all of the credit is given towards the final click that was made on AdWords before a conversion. However, the Last Non-Direct Click doesn't consider direct traffic and instead merits the last channel the consumer was on beforehand. Finally, the Last Interaction is based on the very last point of contact with the business.
A downside to this is that it doesn't take all the complexities of the consumer pathways into account. Therefore, attribution needs to be adapted to include all touch points made across many devices to reach a full understanding.
This will allow investments to be made via proper resource allocation, which considers every click, not just the last one.
The Time Decay model gives more credit to the final interactions, which is more appropriate for those with a conservative growth strategy. In this model, the touch points in the conversion pathway which are within a few hours of the conversion or sale, are attributed more credit.
Again, this isn't always the best model to use if you want an overall picture of which strategies have been the most successful, as leads and sales can be made at any step of the marketing funnel, regardless of time.
While the Position Based model also credits the final interactions, it acknowledges that the first interactions introduce the consumer into the pathway, so need to be credited accordingly. This is best suited to those who are growth orientated and to those who may be new to the market.
The first, and last interactions are each attributed with 40% of the credit, whilst interactions in the middle are attributed with the final 20%. This is then evenly shared between those interactions made in the middle of the process by the consumer.
The Linear model simply attributes credit equally (25%) to each stage of the conversion pathway, regardless of the time or order in which the user accessed them. The pitfalls of this is that you can't pinpoint which stage is performing the most effectively, meaning that not every interaction is deserving of equal credit.
The Data Driven model differs the most different from the other models, in that it uses conversion data to calculate the credit which should be assigned to each stage of the conversion path.
This model provides the highest levels of insight, but there are minimum requirements to the amount of data you need to provide, which are 600 conversions and 1500 clicks over a 30-day period.
By changing the attribution model, you should bear in mind that the timeline for analysis should also alter; AdWords will report conversions based on earlier clicks, so using larger time frames and avoiding the inclusion of recent days is advisable (e.g. analyse days 1-30 on day 35).