Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Watch here.

Complex systems beg for oversimplification through metaphor. DNA is “genetic code.” The atom is a “building block.” The brain is a “computer.” We see them all the time, usually when an expert at something is trying to explain that something to a non-expert.

It is rare, however, and arguably far less efficient, for two experts to use that metaphor with each other. Technical terms of art are more complicated, but also more specific and actionable than metaphor among those who understand them, and result in more information conveyed more accurately over less time. They are also less prone to miscommunication.

Yet we marketing professionals still go on about the “path” to purchase. What was only meant to be a conceptual metaphor for stages in a complex system has become the framework by which we conceptualize, design, and measure marketing efforts. Even to ourselves, we perpetuate a myth that the way people buy is not only entirely fathomable and marked like a well-worn trail, but also linear or even directional in nature. This doesn’t pass the common-sense test, let alone a simple web or media path analysis for nearly any product.

As an example, a study by the travel company Expedia found that people make, on average, 140 visits to various travel sites in the 45 days before making an actual booking. That’s just one channel (web). Imagine how many other influences, from social media to TV, to friends and family, also impact that one purchase decision. Even in the most poetic sense, that is no discrete “path.” Yet the entire ad industry develops campaign strategies, buys media, and tracks success based on this fiction. This weights channel preferences to those that benefit from last click attribution, which subsequently weights budgets to those media as well.

It might finally be time to throw away the “path” metaphor altogether. This is by no means the first call or attempt to do that – we have seen purchase “stages,” “loops,” “funnels,” and “sine curves,” all striving, with varying degrees of success, to put a more nuanced framework of order around the chaos that is purchase decision making. But it is precisely in that adherence to “order” where most of these frameworks break down. Most of our decision making, some experts put it as high as 95 percent, happens at a subconscious level, influenced by dozens of known and unknown variables far beyond the marketer’s control. Forcing that messiness into an order is folly. It’s time we embrace randomness.

English mathematician Karl Pearson coined the phrase “a random walk” to talk about a distribution of outcomes based on random but constrained events. Imagine a person is standing on top of a building in the center of a roof. You ask them to flip a coin and if the coin lands on heads, they turn right and take a big step. If it’s tails, they turn left and take a big step. From a probability perspective, what is the likelihood they will fall off the roof after 10 flips? After 100? A lot depends on the size of the roof (obviously). Also, are you the only agent making the person flip or are others persuading them to flip as well? Do they occasionally flip the coin when you don’t ask them to? And most importantly, could you do anything to influence the outcome of the flip based on what you know about the customer and where they are on the roof?

In the above “random walk” scenario, there are more unknowns than knowns. That may make us as experts in a field uneasy. Yet that’s a far more accurate representation of a customer experience over time than a “path.” A lot of customers, even without our touching them meaningfully, will stay close to the center of the brand. And a lot of consumers, even with multiple, carefully construed touchpoints, will end up walking right off the ledge. Random happens, though its impact can be mitigated with analysis, research, and intelligent speculation. What we need to understand and to teach as marketers is that simply because an experience has random variables doesn’t mean we can’t influence decision making or predict and measure eventual outcomes.

Entire wings of mathematics have been developed to understand probabilities from random walks, and there’s no reason we can’t apply that on behalf of our clients and brands just as well, if not better, than path analysis. Instead of spending time streamlining a path-to-purchase construct that is mostly fictional, perhaps we should be looking for ways to shift variables that we can control positively in an inherently chaotic system.

At one level, this is merely replacing one metaphor with another. But at another, it injects an authentic element of the unknown into a customer experience story that marketers have historically been unwilling to acknowledge. That feels like progress. Persuasion is messy and people are weird. We should own that.

Jason Carmel is Global Chief Data Officer at POSSIBLE.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.