The Premium Signal Trap: Why McKinsey's B2B Pricing Shift and the Upper-Middle-Class Arms Race Are the Same Problem
Homes near top-rated schools cost 78.6% more than comparable properties in the same county. They do not produce measurably better lifetime outcomes for the children who attend those schools. That gap - between what the premium signals and what it actually delivers - is not a personal finance problem. It's a decision architecture failure, and it shows up everywhere once you start looking.
Nick Maggiulli's recent analysis of the upper-middle-class trap documents the mechanics in consumer markets. McKinsey's parallel research on B2B pricing in the age of agentic AI documents the same dynamics in enterprise software. The behavioral science connecting them - the reason both traps persist despite clear evidence of their failure - comes from Guy Hochman's work on what happens to the availability heuristic under information overload.
These three datasets are about the same cognitive problem wearing different industry clothing.
The Arms Race That Produces Nothing
The upper-middle-class trap has a precise structure. It's not about overspending per se - it's about competing for positional goods whose value is relative, not absolute. When everyone in your peer group bids on the same scarce signals of quality, the premium evaporates into the competition itself.
Maggiulli's data is stark. New single-family homes shrank 11% in size between 2014 and 2024 while the cost per square foot surged 74%. Homes purchased in bidding wars deliver 6.9% lower annualized returns - you pay more and get less, not just in square footage but in financial performance. College applicants increased 78% since 2015 while elite acceptance rates collapsed, and tuition grew at twice the rate of general inflation. The premium became more expensive at exactly the moment it became harder to access.
The outcome data delivers the fatal blow: children from upper-middle-class families attending elite private schools performed no better than comparable peers at public schools. The Amex Platinum lounge, once a genuine amenity, became so crowded with premium-card holders that the benefit collapsed under its own success. In each case, the market for the signal outran any real advantage the signal once conferred.
When the Signal Becomes the Entire Product
The trap runs on a confusion between signaling and substance. The school's reputation is observable. Your child's counterfactual career trajectory from attending a different school is not. If you cannot directly observe outcomes, you proxy them with signals - and the upper-middle-class consumer market is built almost entirely on that substitution.
The bidding war premium, the private school tuition, the first-class cabin: each of these is defensible in isolation, and collectively catastrophic. What Maggiulli is documenting is not irrationality. It's a rational response to bad information architecture - and the information architecture is not improving on its own.
McKinsey's B2B Version of the Same Trap
Enterprise software built its modern pricing model around the same substitution. Seat-based licensing counts the number of people using the software as a proxy for value delivered. For a generation, this worked well enough - usage roughly tracked value, and headcount roughly tracked usage.
Agentic AI breaks that proxy completely.
When AI agents handle work that previously required human seats, the unit that justified the pricing model disappears. According to McKinsey's research, 40% of B2B buyers now cite seat reduction as their primary lever for cutting software spend. They are not renegotiating - they are eliminating the seats, which is functionally the same as deleting the pricing unit. The signal has decoupled from the substance.
The market response is already visible. Hybrid pricing models - combining subscription floors with consumption-based tiers - grew from 27% to 41% adoption between 2020 and 2025. AI-driven pricing optimization has delivered 2-6% gross margin improvements in real B2B deployments. The vendors who can demonstrate quantified outcomes are migrating toward outcome-based contracts. The ones who cannot are defending seat counts while their customers plan for a world where seats do not matter.
This is the kind of structural transition that category-level pricing research has been tracking - the widening gap between what gets priced and what gets valued, and how that gap expands during platform shifts.
The 30% Verification Problem
There's a telling asymmetry in the McKinsey data: only 30% of AI software vendors have published quantifiable ROI from actual customer deployments. The remaining 70% are still selling on capability signals - demos, feature announcements, analyst positioning - rather than documented outcomes.
That gap is not laziness. It's evidence that outcomes are either harder to measure than the features, or that the vendors have not yet had to demonstrate them. Seat-based pricing persisted because it was never tested against outcome-based alternatives at scale. The buyer accepted the signal because measuring the substance was hard. Agentic AI changed the cost structure of that measurement. Now the test is happening, involuntarily, across the entire enterprise software market.
This is the kind of pattern STI's research tracks systematically - the gap between what signals promise and what outcomes confirm, across markets and over time.
Why Smart People Stay Trapped: The Availability Heuristic's Inversion
The behavioral mechanism connecting both traps has a name, and it is getting more dangerous as information volume grows.
The availability heuristic (Tversky and Kahneman, 1973) holds that people judge probability by how easily examples come to mind. In an era of information scarcity, this was a reasonable shortcut - what you encountered frequently probably was frequent. Guy Hochman's recent analysis for BehavioralEconomics.com identifies what happened to this heuristic when information became abundant.
The original problem was over-weighting vivid, available information. The new problem is its inversion: treating the absence of expected information as evidence of impossibility. Hochman calls this "UnAvailability Bias" - if something significant happened, we should have seen proof of it. When the proof is missing, we do not conclude that the proof is hard to produce. We conclude the event was fabricated.
Applied to consumer decisions: you would have heard by now if private school were not worth it. Everyone you know who achieved something came from good schools. The absence of visible evidence for the counterfactual - people who thrived without the premium - functions as social proof that the premium is necessary. It is not. It is an artifact of which outcomes your information environment makes available.
We have written about how cognitive shortcuts drive purchasing behavior and how the relationship between visible signals and actual outcomes is being systematically exploited as data availability expands. The availability problem is getting worse, not better. More information means more vivid signals and less ability to observe the full distribution of outcomes.
The B2B Version of Invisible Counterfactuals
Enterprise software buyers face the same information failure. You would have heard by now if outcome-based pricing worked better. Your competitors are on seat licenses. The vendors you know best do not offer outcome guarantees. The absence of visible alternatives functions as evidence that seat-based pricing is the natural order.
It is not. The absence of outcome-based pricing data is evidence that vendors have not had to produce it yet - not that the outcomes are unmeasurable. The counterfactual exists. It is just not available in the information environment, which is precisely the condition the availability heuristic was built to exploit.
The availability inversion we have previously documented in AI decision-making contexts applies directly here: the absence of visible alternatives reads as impossibility rather than as an artifact of how the market was structured.
Clarity Cannot Fix a Structural Misalignment
The Branding Strategy Insider's recent piece on organizational clarity adds a third dimension to this pattern: organizations that stay trapped longest are usually the ones trying to solve structural problems with communication.
The argument is clean: teams experiencing unclear strategy respond with better decks, sharper briefs, and more aligned messaging - hoping that the feeling of clarity will follow the words. It does not. What looks like a communication problem is usually a structural one. Clarity that has to be constantly reasserted through better messaging is clarity that was never designed into the architecture.
The parallel holds. Companies do not escape the seat-license trap by communicating better about their value. Upper-middle-class families do not escape the school arms race by understanding its dynamics intellectually - Maggiulli's readers understand the trap perfectly and still bid on the houses. The trap is structural: incentives, peer comparisons, available pricing models, and social information environments all reinforce signal-seeking behavior. Understanding it changes nothing about the architecture.
Opting out requires changing what you are optimizing for. That is a design choice, not a communications choice.
What Opting Out Actually Requires
Maggiulli's prescriptions are deliberately anticlimactic: choose public schools, take economy flights, skip the bidding war. The data shows these choices do not produce worse outcomes - they produce better financial ones. The 78.6% school-proximity premium and 6.9% bidding-war return penalty are avoidable costs that buy nothing measurable.
McKinsey's prescriptions follow the same structure: price outcomes, not signals. Build quantified ROI into the contract structure. Move from consumption-as-a-proxy to outcomes-as-the-unit. The vendors building pricing architectures around documented results are not just better positioned for agentic AI - they are building the only defensible long-term model.
Both answers are versions of the same instruction: identify what you actually want to produce, measure whether you are producing it, and stop paying for signals that cannot be traced back to substance. The hard part is not the identification. It is that the available information environment, the peer comparison set, and the default pricing structures all make signal optimization feel like the rational choice.
The availability heuristic keeps the trap alive. The absence of visible, mainstream examples of the alternative makes the alternative feel impossible. It is not. The examples exist - they are just not the ones being amplified.
If you are working through what this means for how you evaluate a market, price a product, or model a consumer decision, our analysis tools at STI are designed to separate what the signal suggests from what the outcome data actually shows. The gap between those two numbers is usually where the decision lives.