By Bradley Keefer, CRO, Keen Decision Systems
In the current landscape of marketing measurement, incrementality has emerged as a holy grail metric that reveals whether our marketing efforts are truly working. Not in terms of clicks or views, but in terms of causal business impact. It’s no surprise that marketers, data scientists, and executive teams are all asking “Did our marketing actually cause this result, or would it have happened anyway?”
Incrementality is meant to answer that. But here’s the catch: most of the industry’s current tools for measuring it are outdated, narrow in scope, or strategically flawed. Even with newer tools like marketing mix modeling (MMM), the majority of marketing teams are still focused on measuring the past, not planning the future.
That has to change.
This article is a push to evolve from measurement toward decision science. Because the most strategic marketers today don’t just want to know what worked. They want to know what to do next and they want to predict that with confidence.
What Is Incrementality, Really?
At its core, incrementality measures the additional value created by a marketing action. It is about causality, not correlation. If you run a $100K campaign and see a $1 million revenue boost, incrementality asks: how much of that revenue wouldn’t have happened without the campaign?
Traditionally, marketers have used holdout testing like geo splits, user-level exclusions, and matched markets to measure this. These methods can work in theory. But in practice, they are increasingly impractical because they often require turning off marketing that might be working and they assume control groups are truly comparable. Beyond that, they ignore cross-channel effects and environmental variables, and they only answer backward-looking questions.
As Google noted in its Effectiveness Equation whitepaper, this approach is ripe for disruption. Moving away from last-click attribution toward outcome-based metrics and MMM is a necessary step. But it is not sufficient.
The Problem with Incrementality Alone
Google’s position in The Effectiveness Equation rightly emphasizes moving beyond attribution to incrementality, measuring long-term brand and business outcomes, incorporating external factors like economic shifts and competitive pressure, and respecting data privacy by leaning into aggregate modeling.
These are all important developments. But the equation remains incomplete. Because understanding what has worked still doesn’t provide a roadmap to what’s next.
This requires moving beyond incrementality into forecasting, actualization, and scenario-based planning.
That’s the realm of decision science, where the goal isn’t just measurement, but prediction.
Forecasting: The Superior Form of Measurement
Forecasting is not just a planning tool. Done right, it becomes a measurement tool. Here’s how.
You build a forecast using historical data, incorporating all relevant variables, including pricing, seasonality, competitive activity, and macroeconomic factors. Following that, you track actual results against the forecast to understand the lift or gap. Finally, you attribute deviations to marketing or external forces based on statistical inference.
When forecasts are accurate, they do more than predict, they validate. If your forecast accurately predicts sales with a certain campaign and deviates sharply without it, that delta is your incremental lift. It’s a form of virtual holdout that doesn’t require pausing marketing.
This approach accounts for all known environmental variables, reflects cross-channel effects and brand equity lift, offers continuous, scalable insight across all campaigns, and is future-focused by design.
Most importantly, it answers the question: Can I predict marketing outcomes within a reasonable margin of error?
If the answer is yes, then you’ve not only proven marketing’s value, but you’ve also made it predictable.
Beyond Measurement: The Rise of Decision Science
Decision science brings together statistical modeling, causal inference, and financial planning to help marketers make better choices before the spend happens.
MTA, incrementality and MMM tools are pioneering this shift. These tools allow marketers to forecast marginal ROI for each additional dollar spent, simulate various media mix scenarios, balance short- and long-term investments, and plan campaigns that align to financial objectives like EBITDA or NPV.
They also measure historical incrementality, and help close the loop between insight and action. It doesn’t just tell you what happened. It tells you what’s likely to happen next.
Consider a seasonal brand that traditionally spent heavily in August for its back-to-school campaign. Using conventional MMM, they validated this timing as effective after the fact.
When they layered in forecasting and scenario modeling, something shifted. They discovered that reallocating a portion of that budget to earlier months would capture demand earlier in the season and increase ROI.
The result? A $12M lift in revenue with no increase in spend.
This wasn’t just smarter measurement. It was smarter planning. And it came from modeling, forecasting, and simulating outcomes, not running a test and waiting for results.
The New Standard: Prove and Predict
Marketing measurement is no longer just about attribution or even incrementality. It’s about accountability, predictability, and actionability.
We must stop treating historical measurement as the endgame. True marketing effectiveness comes from the ability to simulate future outcomes, forecast returns, and guide strategy dynamically.
That means moving towards dynamic scenario planning, forecasting and actualization, and holistic decision systems.
The industry is ready for this shift. The tools exist. The frameworks are proven. What we need now is a mindset change from insight to foresight and from measuring to engineering results.
Google started the conversation with The Effectiveness Equation. Let’s finish it. It’s not enough to measure what happened. Marketing effectiveness is about knowing what to do next and being confident in the outcome. That’s not just measurement. That’s decision science. Let’s build for a future where marketing doesn’t just react to results. It predicts them.
