A data science colleague and marketing science professor challenged my thinking this week with a provocative statement – he said that “attribution” may have been the worst invention in marketing science.
Quick background: For many years digital media marketers have benefitted from pieces of code that follow consumers from ad exposure to behavioral response. These pieces of code enable marketers to track consumers exposed to their advertising across various websites, and then to the consumer’s visit to the brand website and then to a desired action – like “make an appointment.” And attribution methods have been established to measure performance and return on investment of those various exposures leading to the desired response.
But there are several problems with attribution. And they’re getting worse.
For starters, most attribution models have output that credits every conversion event with paid media.
In reality, a substantial portion of conversion events in any given year are caused by many variables, or a combination of variables, including previous years’ marketing efforts – paid, earned and owned. Or weather or interest rates or distribution or competitive share of voice. The trick is measuring the incremental conversions from this year’s efforts.
Secondly, walled gardens; e.g. Meta and Google, are notorious for hogging all the credit. If you’re a digital marketer, for sure you have experienced this – the sum of merit events that Google attributes with the sum of merit events that Meta attributes equates to more merit events than you actually experienced.
Nevertheless – the industry marches forth with attribution everywhere. Recently, Apple quietly announced a new Ad Attribution Kit. What does this kit actually do?
“Apple plans to introduce long-promised support for reengagement campaigns, which has been on the developer wish lists for awhile,” writes AdExchanger. Apparently the last attribution methodology was a bit of a bust – “an overly complicated tool that does less to preserve privacy than it does to undermine ad effectiveness in the name of privacy,” says Eric Seufert at Dataseat.
So what should marketers do? Here are my top three tips:
- Eliminate last-click attribution as a methodology for ad-spend decisioning. Last-click metrics include UTM codes or referrer traffic. These are the last touchpoints your consumers visited before coming to your website, but they don’t include all the previous exposures that influence that click. Last-click attribution leads to disproportionate investment in channels like search. Our modeling has revealed that some marketers are actually overinvesting in search, not measuring the synergistic impact that brand-building media is having.
- Evaluate your media efficacy over the long term because advertising exposure builds over time and, especially as positive brand messaging, lingers in the minds of consumers. Some easy ways to convey this dynamic are data visualizations correlating ad spend to key performance metrics – over time – expressed in a line chart. You can track whether the slope of the line is improving or declining. That is a much better method for evaluating the impact of ad spend. Long-term typically means by quarter, rather than by week.
- Don’t get duped by your digital walled gardens, e.g. Meta, Google, taking credit for the same conversions. Their attribution methodology says that any browser exposed to that walled garden’s ad placement, which then converts on your website, should be credited to that walled garden’s exposure. To make matters worse, this methodology is dependent on cookies, which are going away. Emerging are new algorithms that use machine learning to attribute channel performance, rather than relying on the self-measurement of walled gardens or outdated attribution methods reliant on cookies or audience identifiers.
In developing FutureSight.online™ we’ve found a way to help marketers overcome obstacles and maximize opportunities.
In summary, “attribution” isn’t bad, it’s just mostly broken. And it’s time for a new approach to attribution that is future-proof. Death to attribution – long live attribution (in its truest, incremental, sense.)
Contributed to Branding Strategy Insider by Marilois Snowman, Partner and CEO of Mediastruction.
At The Blake Project, we help clients worldwide, in all stages of development, define and articulate what makes them competitive and valuable. Please email us to learn how we can help you compete differently.
Branding Strategy Insider is a service of The Blake Project: A strategic brand consultancy specializing in Brand Research, Brand Strategy, Brand Growth and Brand Education
FREE Publications And Resources For Marketers
Post Views: 0
