Most business owners running brands on analytics believe they’re making evidence-based decisions. They are — just not about the question that matters most.
“What does our audience respond to?” is a measurable question. You can track it, rank it, optimize it. But it’s a different question from “What do we want our audience to understand about us?” — and confusing the two is how brands drift. You end up optimizing for what worked last month without ever deciding what you’re trying to build.
Barry Schwartz, emeritus professor of psychology at Swarthmore College, has spent years making the case that good decision-making requires judgment, not calculation. His argument isn’t anti-data. It’s about where the real work of deciding happens — and where most decision frameworks skip it entirely.
What You’ll Learn
- Why judgment and calculation are different cognitive acts — and why the distinction matters for brand decisions
- What Barry Schwartz’s critique of rational choice theory reveals about how data gets misused
- How Goodhart’s Law turns brand metrics into a trap
- What contextual judgment actually looks like when applied to brand strategy
- A simple test for when you need data and when you need to think harder
What Is the Difference Between Judgment and Calculation?
Calculation finds the best answer to a problem that’s already been set up. Judgment determines whether you’ve set up the right problem. Most of the real work in a good decision happens in that second step — and most decision frameworks skip it entirely.
In a March 2026 piece for Behavioral Scientist, Schwartz puts it directly: “The real work in deciding is not in the calculation, but all the thinking that surrounds it.” He draws on Aristotle’s concept of phronesis — practical wisdom — which is the capacity to identify what the right goal is in a specific situation, at a specific moment, with specific people involved. Not which option scores highest against a predefined set of criteria. What the right criteria are.
His critique targets rational choice theory (RCT), the framework that treats good decisions as optimization problems: enumerate options, weight factors numerically, calculate expected utility, choose the highest score. Schwartz points out that framing a decision correctly is a prerequisite for any of that to work. “Framing is a prerequisite for the operation of RCT; without framing, RCT procedures can’t even get started.” RCT provides no guidance on how to frame a problem well. The calculation assumes the problem is already set up correctly — which is precisely where most decisions fail.
The failure mode to avoid: Running a precise calculation on an imprecise setup. The output is exact. The answer is wrong.
—
Judgment — according to Schwartz — requires understanding, reflectiveness, self-knowledge, and clear articulation of values. It’s not intuition. It’s deliberate thinking about what you’re actually trying to build, whether your current approach is moving toward it, and whether the goal itself is right. That kind of thinking cannot be replaced by more data or better tooling.
Why Brand Decisions Are Judgment Problems
Brand decisions belong to the category Schwartz describes as requiring judgment rather than calculation — because the core question isn’t “which option scores highest?” It’s “what do we want to mean to people?”
That question has no answer in the data. Engagement rates, click-throughs, reach, and conversion figures all measure audience behavior against content you’ve already produced. They tell you what resonated. They cannot tell you whether what resonated is what you intended, whether it builds toward something coherent over time, or whether the audience that responded is the audience you’re trying to reach.
Michael Luca (Johns Hopkins) and Amy Edmondson (Harvard Business School) documented this limitation in a 2024 Harvard Business Review piece on data-driven decision-making: experiments tend to “focus on outcomes that are easy to measure rather than on those that business leaders truly care about.” In brand terms — you can measure shares, not understanding. You can measure clicks, not trust.
The Two Questions Framework: There are two questions brand decisions require. One is measurable: What does our audience respond to? The other isn’t: What do we want our audience to understand about us? Both matter. But only one can guide what your brand should mean. Optimizing for the first while neglecting the second produces a brand that’s good at performance and unclear about purpose.
The judgment a brand decision requires is contextual: What are we trying to build? Who are we building it for? Does the signal we’re sending match the signal we intend? These aren’t calculation problems. They’re Schwartz’s practical wisdom problems — requiring sustained, honest thinking that moves between your stated values and your actual behavior, asking at each step whether they agree.
As a rule: Data tells you what performed. It doesn’t tell you what to mean. Those are different questions, and only one of them belongs in a spreadsheet.
What Happens When Brand Metrics Become Targets?
When a metric becomes a target, it stops being a useful measure. Economist Charles Goodhart made this observation in 1975: “Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.” In plain language — the moment you start optimizing for a measurement, you corrupt what it was measuring.
For brand strategy, this plays out in slow and quiet ways. Engagement rate becomes the target, so content gets optimized for engagement: shorter, more reactive, more familiar. The metric goes up. Understanding — what the audience takes away about who you are — goes down. You’ve solved the wrong problem with precision.
Schwartz’s argument is that this is a framing failure at the source. The decision-maker set up the problem as “how do we improve engagement?” rather than “what do we want this audience to understand about us?” The first is a calculation problem. The second is a judgment problem. Spend enough time solving the first, and you can make real progress on it while making the second problem harder to solve.
The warning: Optimizing brand performance without a prior judgment about what the brand should mean produces coherent numbers and incoherent meaning. The performance metrics will look healthy right up until someone asks what the brand stands for.
What Does Contextual Judgment Look Like for Brand Decisions?
Contextual judgment means bringing the full situation to bear before reaching for a framework or formula. Schwartz summarizes the answer to most hard decisions as “it depends” — not because there’s no right answer, but because the right answer requires taking the specific context seriously rather than abstracting it into a calculation.
For a brand, this translates directly. The questions data can’t answer include: Which of the things that performed last quarter actually builds the identity we’re constructing — and which just got clicks? When an audience segment responds to something off-brand, do we follow or hold the line? What does coherent mean, for us, given where we’re trying to go?
None of these have numerical answers. They require what GoPractice described in their analysis of data-driven culture: recognizing that “data reflects what exists, not what’s possible.” Strategic vision — what you’re trying to build, not what you’ve been doing — is intrinsically unmeasurable. That’s not a failure of measurement. It’s a property of the question.
Decision tree — when to use data, when to use judgment:
- The question is about past performance → use data. Which format performed better with this audience? What time of day drives highest engagement? These are calculation problems, and data answers them.
- The question is about what to build → use judgment. What do we want this audience to understand? What does this brand stand for, specifically? What meaning are we compounding over time? These require thinking, not tracking.
Most brand decisions that go wrong confuse the two. They apply data to a judgment problem — and get precise answers to the wrong question.
Conclusion
Schwartz’s argument isn’t that analysis is useless. It’s that calculation can only optimize a problem that’s already been correctly framed — and framing is a judgment, not a calculation.
For brands, the framing question is: what do we want our audience to understand about us? Not what have they responded to. Not what performed last quarter. What do we want them to understand. Brands that build durable meaning answer that question first — and then use data to evaluate whether they’re getting there.
Data tells you what performed. It does not tell you what to mean. Spend enough time optimizing for performance without answering the meaning question first, and you’ll end up with a brand that’s good at getting attention and unclear about why it deserves it.
