Optimization has a seduction problem. It offers a clean answer to a messy question: run the test, read the number, repeat what worked. That clarity feels like progress. For most business owners building a marketing strategy, the analytics dashboard has become the primary decision-making tool — not because it answers the right questions, but because it answers quickly.
Barry Schwartz and Richard Schuldenfrei, in their book Choose Wisely: Rationality, Ethics, and the Art of Decision-Making (Yale University Press), make an argument that lands harder for marketers than for most: the dominant framework for how decisions should be made is fundamentally broken. Rational choice theory — the logic beneath every A/B test, every attribution model, every click-through rate — fails not just as a description of how people decide, but as a prescription for how they should. Schwartz and Schuldenfrei are both professors emeritus at Swarthmore College, and their argument is precise where it counts.
The implication for brand strategy is uncomfortable. Optimization without judgment doesn’t produce a coherent brand. It produces a performance loop.
What You’ll Learn
- Why rational choice theory fails as a standard for marketing decisions, not just as a description of human behavior
- What gets stripped out when a brand decision is reduced to a calculation
- How optimization without a judgment layer produces results that perform in isolation but build nothing durable
- What judgment actually means in brand strategy, and why it precedes data rather than replaces it
- How to structure the relationship between judgment and optimization so each does the work it’s suited for
What Is Rational Choice Theory, and Why Does It Show Up in Every Marketing Dashboard?
Rational choice theory holds that good decisions reduce to calculation: identify your options, assign numerical values to probable outcomes, multiply, and optimize for the highest expected utility. It is the logic underlying A/B testing, conversion rate optimization, and nearly every framework that treats marketing performance as a measurement problem. Schwartz and Schuldenfrei trace this framework’s failures with precision — and their target is not the familiar finding that humans decide irrationally. Kahneman and Tversky proved that decades ago. The deeper problem is that even after behavioral research demolished rational choice theory as a description of how people think, it “left RCT as the normative standard untouched,” as Schwartz writes in Behavioral Scientist. It remained the model for how decisions should be made.
What rational choice theory requires, Schwartz argues, is closed formal framing: stripping context down to comparable metrics, eliminating the individual and situational nuances that are fundamental to decision quality. For a marketing decision, this means the framework can tell you which headline version produced more clicks last Tuesday. It cannot tell you whether you are building toward anything.
The most common mistake here is treating a performance metric as a decision framework. A metric tells you what happened. A decision framework tells you what should happen next, and why.
Why Performance Data Tells You What Worked, Not What to Build
Performance data is retrospective by design. It measures what your audience responded to given who they were, what they had already seen, and what else was competing for their attention at a specific moment in time. Used correctly, that information is valuable. Used as a substitute for judgment, it becomes a trap.
Schwartz and Schuldenfrei state the problem directly in Choose Wisely: “Rational choice theory has nothing to tell us about what your preferences among options should be, what your values should be, or what set of options you should consider.” Substitute “preferences” for “strategy” and “values” for “brand meaning,” and the marketing translation writes itself. The data operates within a frame you have already chosen. If that frame is wrong — if you are optimizing for the wrong signal, building toward the wrong audience, or reinforcing meaning you never consciously chose — the data will help you do that wrong thing more efficiently.
This is the specific failure mode Schwartz warns against: quantifying things that cannot be quantified leads to false precision. You assign a number to something without a number, then optimize for it as though the number were real. In marketing, the number is usually engagement, reach, or click-through rate. The thing without a number is meaning.
As a general rule: data can tell you how effectively you’re executing. Only judgment can tell you whether what you’re executing is worth doing.
What Optimization Without Judgment Actually Produces
The result of building a brand on performance data alone is coherence failure. Individual pieces perform. The aggregate builds nothing.
An audience responds to a certain type of post, so you produce more of that type. A different type outperforms, so you shift. Over time, the brand’s signal shifts to reflect whatever got the most recent engagement. What the brand consistently means to its audience becomes a function of last month’s analytics. That is not a signal. It is a weather vane.
The Vietnam War body-count strategy, which Schwartz and Schuldenfrei cite in Choose Wisely, is the extreme version of this failure. Military leadership chose measurable metrics over strategic judgment, optimized for those metrics without interruption, and produced outcomes that bore no relationship to the actual objective. The numbers moved in the right direction. The war was being lost. A brand built entirely on performance data runs the same structural risk at a smaller scale: the metrics trend upward while the brand’s capacity to mean something to its audience, in its category, over time, quietly erodes.
Most brands don’t suffer from a lack of content. They suffer from a lack of coherence.
What Judgment Looks Like in Brand Strategy
Judgment is not the opposite of data. It is the precondition for data being useful.
Schwartz and Schuldenfrei propose “intelligent reflection” as the proper alternative to rational choice calculation: a deliberate process that weighs values, considers individual character, and accounts for contextual meaning alongside quantifiable inputs. Unlike RCT, it lacks formal rules. That is not a weakness. It is an acknowledgment that thinking is contextually sensitive — that the meaning of a decision changes depending on who is making it, for whom, and toward what end.
For a brand, intelligent reflection looks like answering prior questions before asking the data anything. What does this brand mean to the audience it matters to? What kind of signal reinforces that meaning, and what kind of signal undermines it? What are we building toward, and over what timeframe? These are judgment questions. They require a clear perspective on the audience (not a target market — an audience with specific values, specific contexts, and specific reasons to care), and a clear perspective on what the brand means to that audience.
Narrative Branding begins here. The strategy, the language, the signals, the coherence across every touchpoint — all of it flows from a judgment about meaning made before any measurement begins. Once that judgment is in place, optimization has something real to serve. The metric is no longer the objective. It is feedback on whether the execution is working. That is a fundamentally different relationship with data.
The most reliable approach is to establish what you’re building — the meaning you’re committed to — before asking performance data which direction to walk.
When Should You Trust Data, and When Should You Exercise Judgment?
Data and judgment are not competing inputs. They operate at different levels of a decision, and confusion about which level you are on is what produces the worst outcomes.
Schwartz and Schuldenfrei acknowledge that rational choice calculation has genuine utility in specific contexts: decisions where options are clearly specified, probabilities are known, and the relevant values can be quantified without distortion. Choosing between two health insurance plans fits this model. Deciding what your brand stands for does not.
The practical heuristic: use data to evaluate execution; use judgment to set direction. When data and judgment conflict at the execution level — a well-designed piece underperforms, a less polished version outperforms — the data is probably right and the execution should change. When data and judgment conflict at the strategic level — the metrics reward a direction that doesn’t reinforce what you mean — judgment should hold. As AdExchanger observed on this distinction, none of the decisions that define the constraints within which optimization operates can be delegated to platforms or dashboards.
Brands that lead with judgment have a coherent foundation that optimization can then serve. Without that judgment layer, the data just cycles.
Conclusion
Schwartz’s argument arrives at an inconvenient moment for anyone who has structured their marketing around performance data. The data is real. The click happened. The conversion was tracked. What the data cannot tell you is whether any of it is building toward something — whether the signals your brand is sending are coherent, whether the meaning your audience is assembling from those signals matches what you intend, whether the brand you are optimizing earns anything durable.
That determination requires judgment. Not as a supplement to the data, not as a tiebreaker when the numbers are close, but as the prior commitment that makes the data meaningful in the first place.
The calculation tells you what worked last time. Judgment tells you what you are trying to build. Start there.
