RMI Contributor: Igor Vladimirovskiy
Marketers, all your data is wrong, and there is no way to make it “right”!
This is hard to hear for some, especially in 2025 when we all know that “data is king” and the “ultimate cornerstone of any marketing strategy”, that all marketing decisions are “data driven”, “powered by AI”, “based in data science” or whatever other cliché you may want to use.
The scariest part is that it’s not wrong due to some sort of collection, duplication, aging or processing errors which can be fixed with improved process or technology – although these are definitely not helping the matter. It’s fundamentally wrong. Wrong from the start. Never had a chance to be right. Wrong at its very core because of the theoretical impossibility of accurate attribution. Marketers should never expect their data to be “right”, ultimately, because there is no way to accurately attribute results (revenue, sales, leads, etc.) to marketing efforts/dollars spent. There is no way to count all engagements and touchpoints, paid or unpaid, intended or unintended, that an individual has with a brand/product/service over their lifetime. Even if there was, there is no way to assign credit to those touchpoints that is not arbitrary. There is no “right answer”, thus, all answers are wrong. Understanding that marketing data is just a set of estimates at best, the key becomes figuring out which ones are more wrong than others.
The Attribution Problem
Marketers’ holy grail is the answer to the question: “What caused this result?”. Marketers hope to trace desired outcomes like sales, leads, increased awareness or specific engagements back to their efforts and budgets. This is where the first and most insurmountable problem arises: attribution.
Consumer journeys are inherently nonlinear and multi-faceted. A single individual may encounter a brand through numerous touchpoints, over an extended period of time. Some of these touchpoints are intentional like paid digital content, or traditional paid media. Some touchpoints are serendipitous, organic and often unintentional, for example, word-of-mouth – I like the shoe because my friend has one, I like this food chain because that’s where my parents took me, etc. The journey can be one engagement, or it could take a year, or a lifetime.
Even if there was a way to record every possible engagement an individual has with a brand, product or service, there is still no “correct” way of assigning credit. What proportion of the outcome should be attributed to the Facebook ad vs a Search result vs a billboard a person saw daily for 5 years on their way to work? How much weight should be given to the email campaign vs the SEO-optimized blog post vs my dad’s opinion since he’s paying this time? Any attempt to answer these questions ultimately relies on arbitrary models. First-touch, last-touch, linear, time-decay, U-shaped, V-shaped and algorithmic attribution frameworks are all constructs based on pre-agreed rules and settings – what you put in is what you will get out – and none of them can claim to be “right.” They’re merely different flavors of wrong.
Why do I want McDonalds fries? Is it the commercial I saw this week or the fact that I got a happy meal when I was 3? Maybe it’s both, maybe neither. What about the other billion interactions – observed, unobserved, noticed, unnoticed, intended, unintended – that I had with McDonalds product, content, personnel, etc. An individual does not even know or remember all the individual interactions and often can’t say themselves as to what it was that convinced them to buy. The consumer may have their own attribution journey in their minds, but even they don’t fully understand why.
The Absence of a “Right Answer”
The core issue is that marketing exists within a complex, chaotic system in which marketing plays a very small part. Consumer behavior is influenced by a multitude of endogenous and exogenous variables – economic conditions, personal preferences, cultural trends, social norms, and random chance – majority of which are beyond the control or even the awareness of marketers. To assume that marketing data can definitively explain outcomes within such a system is to fundamentally misunderstand the nature of causality.
This brings us to the uncomfortable truth: there is no “right answer” in marketing data. Attribution models are tools of convenience, not truth. They help us make sense of the chaos, but their outputs are inherently flawed. The best of all we have is estimates from which we hope to construct frameworks that are “less wrong” than others.
What do we do now?!?
If all marketing data is wrong, does that mean we should abandon data-driven marketing? Well, of course not. Data remains invaluable for informing strategies and demonstrating worth. However, marketers must approach data with a healthy dose of skepticism.
Rather than seeking certainty, marketers should learn to operate in a sea of uncertainty. Recognize the limitations of the available data, as well as data analysis and application. Question the assumptions baked into attribution and media mix models and resist the temptation to chase perfect accuracy. Focus instead on trends, patterns, and directional insights—these are often more actionable and meaningful than precise but misleading metrics.
Conclusion
In marketing, as in life (or more importantly golf), the pursuit of perfect accuracy is often futile and frustrating. The chaos of human behavior and the complexity of modern ecosystems mean that all marketing data is, and always will be, wrong. This realization is not a cause for panic and despair but an invitation to think and approach marketing data differently. By acknowledging the limits of measurement, marketers can move beyond the pursuit of impossible precision and toward a more nuanced, creative, and effective approach.
