Why Digital Attribution Is An Unsolvable Problem
Do you have an attribution problem? Columnist Lewis Gersh believes if you're using attribution the wrong way, then you just might.
French author Andre Gide once said, “a straight path never leads anywhere except to the objective.” Unfortunately, there is no straight path when it comes to digital advertising attribution.
Consumers are exposed to ads throughout the purchase cycle, each having an impact on the buying decision that ultimately is made. Yet despite their best efforts, most advertisers are left scratching their heads trying to figure out which ad or combination of ads drove the sale.
Was it the mobile display ad? The email offer? Or maybe the final search for the product that led them to convert?
You can wander the byways of digital attribution forever and never really arrive at your destination. You can assign credit based on last-click interactions, last non-direct click, first interaction, linear allocation, time decay or by any number of other proprietary schemes.
But despite the fact that digital overflows with data, the science of attribution remains shrouded in mystery.
Can’t Technology Solve That Mystery?
In theory, yes. In practice, no.
Ad tech gives marketers the ability to track more data than ever before. We can count every interaction and determine attribution on a one-to-one level.
But when you scale up to meet the complexities of the real world, attribution becomes a risky metric because to single out an interaction or combination of interactions as definitive is to put a subjective and oversimplified lens on the question.
The data will mislead you, not because the inputs are bad, but because any analysis has a fundamental flaw — consumer behavior is too complex to truly quantify. Short of putting a microchip in a person’s brain, we can’t really say what message or image precisely drove them to buy one car vs. another.
In fact, even with that microchip, we wouldn’t be able to get at precise attribution because we’d still have to contest with the complexities of the human mind; human decisions always come from a combination of factors.
Chasing attribution, therefore, is a lot like tilting at windmills; you can track an ever-expanding universe of data, but you’ll never reach the ultimate truth.
But I Have An Attribution Problem
Yes, you do have an attribution problem. Because you are trying to make attribution do the work for you. The problem is your mindset, not your metrics.
Smart marketers understand the need to be in front of their customers as often as possible, in every way possible. Focusing on attribution distorts reality.
Assigning sole credit to one particular aspect of your marketing, by definition, puts too great an emphasis on that particular aspect. Worse, you miss aspects of your marketing that work but aren’t attribution-worthy.
That mindset is dangerous in the context of a specific campaign, but it’s deadly when applied to your overall budget because an obsession with attribution leads to misallocation at scale.
Real Attribution Is A Function Of Teamwork And Trend Spotting
Two decades ago, there were three ad servers. The online advertising universe was so small we could easily say which click led directly to an action.
That was a big leap forward, but unfortunately, it led many in our industry to buy into the fantasy of precision attribution and zero waste. At the same time, belief in that fantasy caused marketers to abandon two things that have always worked: teamwork and trend spotting.
If the question is how to spend your budget, the worst possible answer is to look at attribution data, which is why marketers have historically looked to trends over performance.
Before the internet, marketers didn’t allocate spending because magazines were more effective than TV — they adjusted their spend based on trends that showed audiences moving from print to television.
Today’s marketers should take a page from their pre-internet predecessors. Trends like mobile adoption and second screen use can all point you in the right direction.
More importantly, trend spotting informs an omni-channel presence so marketers can better allocate budget to match consumer behavior.
One common complaint of the digital age is that teams need better coordination, which is surprising given the proliferation of collaborative tools.
Unfortunately, one byproduct of the attribution obsession is that teams are often pitted against each other. Mobile marketers and their vendors, for example, fight for budget against email and search teams.
These fights are understandable given the billions of dollars flowing into the space and the fallacy that one team really can prove its singular contribution and ROI. But the question isn’t which team has the biggest ROI — it’s which teams add value to the overall marketing mix and make it better.
Okay, But Isn’t Attribution Still Useful?
Of course it is. Right now, our industry has an unhealthy obsession with attribution. But we shouldn’t throw the baby out with the bath water.
Testing and learning are fundamental for marketing.
Attribution is a great tool for testing performance and adjusting the marketing mix when something is showing signs of effectiveness. But like all tools, attribution has a specific function, and when you use it the wrong way, you end up doing more harm than good.