Interested in marketing? What a coincidence, we are too! Let us share with you everything we know.
By: Tom Murray | Managing Director, Agency
Starry Night by Vincent Van Gogh. One of the most recognizable paintings in history. What makes it so great? Some art historians love the heavy use of blues and yellows, while others rank the large, bold brushstrokes as the main reason for its allure. Critics, however, say that it lacks emotion. Who is right? Who is wrong? Who is more right or more wrong? And more importantly, what does this have to do with marketing?
This example sums up digital marketing attribution in a nutshell. Many marketers try to assign a single winner to a campaign without taking other factors into account. Depending on what systems you use, it will decide what the ultimate factor is that your campaigns are judged on. There is no one perfect system for various reasons, but often times marketers are forced to use one model and stick with it. This creates potential problems in making proper business decisions that could mean millions of dollars of lost revenue by limiting channels that are driving performance, but not always in the specific system being used.
Current Attribution Challenges
The main challenge facing marketers today is the fact that there is no attribution solution that can ingest not only all of the spend, click and conversion data, but also any view through data for ads that were seen but not interacted with. In addition, many systems over-credit their own sources while devaluing others (either on purpose or due to lack of access to other channel’s data). This means that Google’s systems tend to overvalue very down funnel conversions such as brand search queries, while devaluing Facebook campaigns. Facebook’s attribution model will overvalue Facebook conversions as it has access to not only click based data, but also users who have seen ads as well, giving it a better sense of the entire cross-device consumer journey.
Taking a step back, many marketers are still leveraging last-click attribution simply because it is something that can in theory be measured across multiple channels by capturing tracking codes in URL strings. While this allows for tracking across multiple channels, this often misses a lot of conversions, or misattributues them and forgets about the path the user took. According to Google, 1 in 3 purchases occur 30 days after their online research began, as well as 48% of users switch between brand and non brand search queries before converting. These conversion paths post some issues by only giving credit to certain parts of the path, while completely devaluing other channels.
Here are a few of the common issues when it comes to tracking a user’s path:
The above examples are just two of the various paths a consumer might take and how conversions may get unattributed. This does not mean that the channels earlier in the process did not have an impact, and in fact they likely had a bigger impact than expected as upper funnel campaign strategies help introduce users to a brand. There are some typical channels that get over-attributed and others that are typically undervalued in the most popular attribution models, which can be seen below.
How To Leverage Attribution Model Findings
No matter how you measure, attribution modeling can still provide valuable insight into the conversion path. It should just not be treated as the source of truth, as it would likely lead to improper business decisions. Take the Van Gogh example earlier — if attribution were to say that the bold brushstrokes were the key piece to making it a successful piece, then a computer system would then say to make as many paintings with that brush stroke as possible, without taking anything else in context. It assumes everything is happening in a vacuum, which is the same as how last-click models and many attribution models measure campaigns.
How Marketers Should Act
The consumer journey is not a silo, and nor should your measurement. The more models and data points you can gather, the better your decision making will be. When marketers have access to more data sets, their critical thinking ability increases greatly. While computer systems (now being powered by artificial intelligence and machine learning) can do a lot of the heavy lifting, these systems lack emotion and context, two things that can’t be replaced from a human mind. Marketers should take in as much data as possible to help their decision making and not be tied to one specific measuring system, as there will inherently be flaws no matter what system is used.