Early in his career, James Turrell, the famous American light artist frequently created installations that played a trick on his viewers. He used what is called the Ganzfeld effect to fill the entire field of a user’s vision with one single color. The result is, “a loss of efficiency in detecting presence and movement” and “an increased and fluctuating state of accommodation.” Basically, with only a single color illuminating the room, the brain becomes lazy and relaxed.
CREDIT: James Turrell
Unlike single color rooms, the concept of digital marketing attribution is complex and challenging. We have addressed attribution in previous blog posts and it is a common discussion amongst analytic thought leaders. Yet our marketing analytic tools tend to over simplify our results and lull us into a single answer as to why digital users took a specific behavior. In reality, we know that human behavior is more complex than stimuli A caused behavior A. As a best practice at LEWIS Piston, we often look at the default settings of our tools and explore alternate methodologies to look at attribution data from different perspectives. Piston is a Google certified partner, so our focus in this blog post will be on Google specific tools: Google Analytics, AdWords and DoubleClick Campaign Manager. We find that by looking at different attribution windows we can prevent our very own marketing attribution Ganzfeld effect.
As an analytics department one of our most used tools is our web analytics suite, Google Analytics. Google Analytics uses a last interaction attribution as the default setting. This means that if a user clicks through from that channel or specific stimuli, the channel that drives them to the site gets credit for the conversion. The tool does allow you to see conversion behavior from other vantage points. To do this, go to the conversion section of Google Analytics and select attribution and model comparison tool. Up to three different models of attribution can be evaluated to see the impact of specific channels from different models. Moreover, beyond how Google Analytics attributes is the analysis of when. From a timing perspective Google Analytics assigns the conversion to the day when the conversion occurs.
AdWords calculates conversion differently from Google Analytics. Instead of using last interaction attribution, AdWords uses a default 30-day conversion window. This means anyone who interacted with a Google Search Ad or saw a Google Display Network ad within the 30 days of a conversion is attributed to an AdWords based conversion. Suddenly interactions such as the impact of display view-through become much more meaningful in driving conversion. From a timing perspective AdWords assigns the conversion to the day of the click on the ad.
DoubleClick relates to AdWords, but as an ad server, it is a tool used to traffic all of your digital media. DoubleClick uses floodlight tags that sit on top of AdWords and other digital partners. This means DoubleClick can see how digital channels work together and assign conversions to one specific channel based on defined business rules. The default setting in DoubleClick is defined during campaign set up and can be from one to 30 days. Unlike AdWords when the attribution windows are the same (30 days), DoubleClick assigns the conversion to the day of the conversion and not to the day of the click as used in AdWords. This often causes data discrepancies if you have longer purchase cycles or lags to conversion.
As you can see, these are three commonly used digital tools that each have a different default attribution window. It is important to look at each tool to understand how shifts in methodology can impact the way you read the results. By getting familiar with each tool one can then open the conversation of different methodologies that evolve beyond last touch. As admirers of James Turrell have said, "He instead is using the light, the illumination; it gives him a completely different pallet and allows him to present very different experiences." By looking at different attribution window durations and different methodologies one can uncover more insight around how digital channels and their stimuli impact behavior.