Understanding marketing ROI is a lot like studying climate change or pandemics

Understanding marketing ROI is a lot like studying climate change or pandemics. As we see it, there are 3 factual principles that make this true:

  • The world is one giant web of network effects. This is true for climate change and pandemics. It’s also true for the interlocking relationships that exist between business investments in marketing and communications, audience belief and behavior, business performance, and a whole host of external factors such as the economy.
  • The unaided human brain can’t handle more than 3-4 variables with any accuracy. That’s why network effects are so challenging to see and understand.
  • To understand these interconnected and changing relationships, regression analytics are the proven GPS for scientists and business people alike.

But why is the studying of marketing ROI, climate change, or pandemics so complex?

Let’s talk about network effects first

The term “network effect” was coined to describe the idea that as more and more people use a product or service, its value increases.

In the years since, the idea of a network effect has been recognized and used more broadly to describe the cumulative impact that’s exerted on a particular outcome as the number of different independent factors or variables swells.

Many times, network effects are positive, but as we’ve all seen, they also can magnify the power of negative and even toxic factors, driving us into some very dicey places.

Climate change and pandemics are excellent examples of this web of multi-factor, multi-vector network effects that can alter the life experience of many people around the world.

When we seek to either leverage or counter a large network effect, the first thing we have to come to grips with is its extraordinary complexity.

The more factors are involved — people or otherwise — the harder it is to clearly understand what’s more or less important and what’s absolutely not relevant at all. Each factor is potentially an accelerant or retardant on each and every other factor. 

And then we have the ultimate complicator of all: Time.

Time introduces non-linearity into the network 

The concept of Time Lag is the notion that everything we do takes some amount of time in order to produce one or more effects. Time lag in a given relationship can be very quick, or it can be months or years.

And within a given period of time, there often is a chain of relationships that splinter and ultimately deliver any number of outcomes. We commonly use phrases like “ripple effect” or “knock-on effect” to describe this reality.

There’s more

Complicating matters, network effects are obscured by something called the Flywheel Effect. Jim Collins, author of “Good to Great,” explains it this way: “The Flywheel effect says that no matter how dramatic the end result… there is no single defining action, no grand program, no one killer innovation, no solitary lucky break, no miracle moment.

Rather, the process resembles relentlessly pushing a giant, heavy flywheel, turn upon turn, building momentum until a point of breakthrough, and beyond.”

Put another way, it means that something can appear to be having no effect for months or years before suddenly exerting a very large effect indeed. 

The way that pandemics spread, and the way that climate change has moved from incremental to dramatic change, are different examples of the Flywheel Effect in action, showing how huge numbers of intersecting cause and effect relationships can intersect and magnify their individual impacts on our world.

Your brain + data is not enough

Human beings love our intuition because when we’re right, we feel powerful, almost psychic. We like to think that we can look at these factors and “see” the truth.

Indeed, researchers have found that few things dump as much adrenaline and dopamine into our systems as the feeling of “being right.”

There are many situations where intuition delivers a lot of value. No experienced scientist or mathematician would argue otherwise.

But accurately understanding a network effect with “flywheel complications” isn’t something you can do by staring at data and figuring it out.

The problem is that your brain and ours can’t process more than 3 variables with any accuracy. Remember when you felt completely overwhelmed by the complexity of an urgent decision? You experienced exactly what we’re talking about.

Data scientists are famous for arguing the fine points of X at the drop of a hat. This often gives laypeople the impression that nothing is really “settled science,” particularly in mathematics or data analytics. This is where the old saying about “lies, damn lies, and statistics” has its root.

Let’s be clear: the power and accuracy of regression analytics are not in doubt. To understand network effects, regression is not only a must, it really must be considered the inarguable core of any analytics effort that seeks to understand the relationships that deliver impact, value and ROI. It’s when people decide to twist the data or the analytics to fit a particular agenda that problems are created.

Let’s look at another specific use case – Marketing and Communications

 When we think about the effects that marketing and communications have on customer experience (CX) and business performance, we’re talking about an immense array of known and unknown factors that are networked into systems of cause and effect. Some are positive, some are not. Some accelerate and multiply other factors, some slow and retard them.

As a result, no marketing or communications team can optimize the network effect that is CX without automated regression analytics. It’s not just that the network effect is simply complex.

It’s also changing at a staggeringly erratic pace that no human can keep up with, caused by far more reasons than anyone can foresee without the assistance of powerful analytics.

And as we’re seeing now, the world of marketing impact is dramatically impacted by the larger reality of the world we live in.

In the face of this challenge, there’s some really great news 

Scientists and researchers have used powerful regression analytics to better understand the many cause-and-effect relationships in the world around us.

The good news is this math is extremely well-suited to understanding the multiplicity of factors driving business performance, including their relative relationships to one another and their relative significance.

Even if they haven’t used it, many marketers and communications pros will recognize Marketing Mix Modeling, or MMM, as the application of regression and other econometric-style analytics to better understand the extent and causes of marketing’s impact, value, return on investment, and time-related escalation and decay.

Since its advent in 1990 at advertising agency J. Walter Thompson and Procter & Gamble, MMM has expanded to include many variables not originally included, such as the impact of different brand investments and longer time lags. It’s now matured into Unified Value Optimization (UVO).

Today, UVO’s mathematical accuracy and power are virtually unassailable. Its biggest challenges have been operational. Assembling the data, tech stack, talent, and domain knowledge necessary to run regression analytics at scale has been very difficult and expensive, making it hard to scale in large organizations.

Its time-consuming complexity also has meant that even experienced data science teams often struggle to deliver UVO analytics in time to affect key business or marketing decisions.

The Advent of the “Business GPS.” 

Today, the automation of regression analytics means that marketing finally has a GPS, a real “eye-in-the-sky” that understands the current state, the desired future state, the multiplicity of factors accelerating or slowing that journey, and the best ways to get there. 

All updated in minutes, not months, effectively transforming marketing from a line in the expense ledger to a powerful analytics-led multiplier of business performance.

About Mark Stouse

Mark Stouse is the CMO-turned-CEO of Proof Analytics. In recognition of his pioneering work in automated analytics, he was named to the Innovator 25 list in 2015 and 2018, and once also as Innovator of the Year.

Stouse recently was named one of the Top Ten Most Influential Analytics Leaders by Analytics Insight, the top journal serving the global data science community.

by Mark Stouse with Dr. Melissa Kovacs