A new framework — Coupled Value Theory — for understanding why interfaces fail the people who use them, and what to do about it.
Imagine someone who has spent thirty years wondering about their origins. They sign up for a service that promises to help them find their birthparents. They move through the intake form, they answer the questions, they feel — for the first time — like something might actually happen.
Then a notification arrives: we found a potential match.
They click. A screen appears. “Unlock this record — $29.99.”
That screen was designed by someone. Someone made a decision about what to put there, what words to use, where to place the button. That decision was almost certainly optimized — A/B tested, conversion-analyzed, revenue-modeled.
And it is, by almost any measure, a terrible piece of design.
Not because it charges money. Charging money is fine. Not because it appeared at a moment of emotional intensity. Moments of emotional intensity are when people make decisions.
It is terrible because it is optimizing for the wrong thing, at the wrong level of granularity, without any awareness that it is doing so. The person building this product probably thinks they have a UX problem. They don’t. They have a theory problem.
The frontend and backend of every modern application are optimizing for different things. Most design frameworks pretend this isn’t true.
The Split Nobody Talks About
Modern applications delivered over the internet have a structural divide that shapes every design decision, whether you acknowledge it or not.
The frontend — the screens, the flows, the interactions the user touches — is nominally optimized to maximize the human’s perception of value. Their experience. Their sense of progress, trust, delight, reduced anxiety.
The backend — the data, the logic, the infrastructure, the business rules — is optimized to maximize the provider’s perception of value. Revenue. Engagement. Retention. Data acquisition.
These two optimization processes are running simultaneously, on the same product, with partially competing objectives. And most design frameworks treat them as if they are the same thing, or as if one doesn’t exist.
User-centered design tells you to optimize for the human. It says almost nothing about what to do when the business model runs directly through the human’s emotional state.
Growth frameworks tell you to optimize for the provider. They are admirably honest about this, but they don’t give you a way to evaluate the human cost of each growth move.
What neither framework gives you is a way to analyze the relationship between the two. That is what Coupled Value Theory — CVT — is for.
The Framework in Three Lines
For any interface element, define:
H(e) — Net Human Perceived Value
What the human gains (functional utility, trust, reduced anxiety, pleasure, autonomy) minus what the human pays (cognitive load, time, attention, emotional friction, money). Measured at the element level. Trust ruptures locally — it does not average across a session.
P(e) — Net Provider Perceived Value
What the provider gains (revenue, data, engagement, retention) minus what the provider pays (development, infrastructure, support). Measured at the flow level. Provider utility is aggregate — a friction point that costs the human dearly may be acceptable to the provider if it lifts conversion downstream.
κ(e) — Coupling Coefficient
The degree to which the provider’s value comes from the same resource the human pays. Attention, money, data, autonomy. When κ is high, the provider’s gain is constituted by the human’s loss. When κ is low, both sides gain from the same element.
The design quality of an interface is a function of all three. High-quality design maximizes H, meets P, and minimizes extractive coupling κ.
Simple. But it has a non-obvious implication that changes how you audit your product.
H is measured at the element level because trust ruptures locally. P is measured at the flow level because revenue is aggregate. This asymmetry is where most products go wrong.
The Asymmetry That Explains Everything
Here is the most important thing CVT says, and the thing most product teams have never explicitly articulated:
The provider measures value at a different granularity than the human experiences it.
Your analytics dashboard shows you flow-level data. Conversion rates. Funnel drop-off by stage. Revenue per session. These are all aggregate measures. They smooth over what happens at individual moments.
The human does not experience your product at the flow level. They experience it one element at a time. And trust — the most valuable thing your product can build — does not accumulate gradually and fall gradually. It builds slowly and breaks instantly.
This means your dashboard can show a healthy conversion rate on a payment screen at exactly the moment that screen is destroying the trust of the 20% who don’t convert. You will never see it in your metrics until they are gone.
The birthparent search paywall is a perfect example. The provider sees: “80% of users who reach the document unlock screen convert to payment.” The provider does not see: “The 20% who did not convert left at the moment of their highest emotional investment, and they are not coming back, and they are telling people.”
CVT makes this structural blind spot visible. It demands that you put your business model and your UX flow on the same canvas, and ask: at every point where I am capturing provider value, what is the human’s H at that exact moment?
Five Coupling Patterns You Already Recognize
Once you have the vocabulary, you start seeing coupling patterns everywhere. There are five types. You have experienced all of them as a user. You have probably built some of them as a founder.
Type
What it means
You’ve seen this in...
Generative
H and P rise together. Remove the provider’s benefit and the human experience is unchanged.
A progress indicator that reduces anxiety and increases retention. A great onboarding flow that delights users and improves activation.
Extractive
Provider value is constituted by human cost. Remove the provider benefit and the human experience improves.
Results withheld behind a paywall. Notification spam optimized for engagement. Dark patterns in cancellation flows.
Transactional
Real cost, real value exchanged. But timing and framing determine whether it feels fair or exploitative.
A payment for a specific service rendered — legitimate, but dangerous at moments of peak emotional investment.
Deferred
Costs the human now, pays back later. H is negative at the element, positive across the flow.
A detailed onboarding form. Identity verification. Anything that asks for effort before delivering value.
Asymmetric Risk
Neutral on average, but high downside variance for the human. Provider bears none of the variance.
“You have a new match!” — great if true, devastating if a false positive. The provider’s P is the same either way.
The CBT parallel is intentional. Just as Cognitive Behavioral Therapy gives patients a vocabulary for naming distorted thought patterns so they can change them, CVT gives product teams a vocabulary for naming distorted coupling patterns. You cannot fix what you cannot name.
Where CVT Fits in Existing Thinking
CVT is not working in empty space. There are three bodies of prior work it builds on — and being honest about where it overlaps and where it differs is part of making it useful.
Dark patterns research (Brignull 2010, Gray et al. 2018) is the most adjacent field. It has catalogued and classified manipulative interface design with real rigor. But the literature is predominantly descriptive and taxonomic — it tells you what bad design looks like after the fact. Researchers in the field have themselves noted that it lacks a conceptual foundation for explaining why certain designs are problematic and what the underlying mechanics are. CVT is an attempt to provide that foundation.
Value Sensitive Design (Friedman, Kahn & Borning, from the early 1990s) is a theoretically grounded approach to technology design that accounts for human values throughout the design process. It acknowledges multiple stakeholders with potentially conflicting values and provides methods for surfacing those conflicts. But VSD is focused on the ethics of design process — who is involved, what values are considered, how conflicts are negotiated. It does not model the economic and cognitive mechanics of the tension between human and provider utility at the element level, nor does it produce operational design heuristics that a working PM or AI design tool can apply directly.
Principal-Agent Theory (Ross, Mitnick, Jensen & Meckling, 1970s economics) describes the structural problem of one party acting on behalf of another with misaligned interests. The frontend-backend split maps onto this framework. But principal-agent theory is an organizational economics construct. It does not produce interface design heuristics, and it does not model the cognitive and emotional dimension of what the human experiences at each moment.
Dark patterns research describes what goes wrong. Value Sensitive Design describes how to involve stakeholders ethically. CVT explains the underlying mechanics of why the tension exists — and gives designers a variable to optimize.
Three things in CVT are, as far as I can determine, genuinely new. The measurement asymmetry principle — that H must be measured at the element level while P is measured at the flow level — is not formally articulated in any adjacent framework, even though it explains the most common structural failure in product design. The coupling coefficient κ as a continuous design variable shifts the designer’s question from “is this a dark pattern?” to “what is the coupling type and magnitude of this element?” And the five-type coupling taxonomy — which includes transactional, deferred, and asymmetric risk alongside the familiar generative and extractive — gives language to the intermediate cases where most real product decisions live.
What To Actually Do: Five Heuristics
Theory without practice is just terminology. CVT produces five design heuristics that any founder or PM can apply to their own product today.
01 The Element Floor Rule
No element may produce a negative H that the human cannot recover from within the same flow.
Any screen that delivers bad news, requests payment, or creates friction must be immediately followed by acknowledgment, explanation, and a forward path. You cannot leave a negative H event unaddressed and expect trust to survive.
02 The Transparency Obligation
When transactional coupling occurs at a moment of elevated emotional state, provide: cost clarity, value specificity, and a consequence-free exit.
Not “unlock this record” — but “retrieving this document requires a researcher. The cost is $X. This is the 1947 Cook County birth record that may contain your birth mother’s name. Not now — your search is saved and you can return at any time.” All three components. Every time.
03 The Variance Test
Design for the worst case, not the average case.
Show your design to someone and ask them to imagine the most painful plausible outcome. A false positive match delivered to someone who has been searching for decades. A payment screen appearing when someone can’t afford it. If the interface is not humane in the worst case, it is not ready.
04 The Flow-Level P Audit
Map every point where you capture provider value against the human’s H state at that exact moment.
This is the structural audit that CVT demands. Put your business model and your UX flow on the same canvas. Flag every row where P capture coincides with negative or vulnerable H. These are your highest-risk design decisions — they require explicit resolution, not just good intentions.
05 The Generative First Principle
For every point where you must capture P, ask first whether a generative alternative exists.
Before defaulting to a paywall, a data request, or an engagement nudge — ask whether there is a design that achieves the same provider value while raising human value simultaneously. You will not always find one. But you will find them more often than you expect, and they will build better products.
How to Audit Your Own Product Right Now
You do not need a research lab or a new tool. You need thirty minutes and honesty about your business model.
Take your most important user flow. Map it step by step. For each step, ask three questions:
What is the human feeling at this exact moment? Not what you hope they are feeling. What are they actually feeling?
Are you capturing any provider value at this step — revenue, data, engagement, attention?
If you removed the provider benefit from this step, would the human experience improve? If yes, you have extractive coupling. Name it.
The audit will show you things your metrics never will. It will show you the moments where you are asking humans to pay — in attention, in data, in money, in trust — without giving them the information they need to make that payment freely.
It will also show you something surprising: how many of your highest-value product opportunities are sitting inside moments of human need that you have been treating as monetization targets rather than design challenges.
The same emotional intensity that makes a product moment dangerous for extraction makes it powerful for generative design. That is where your best product work lives.
The Person on the Other Side of the Screen
CVT is not an argument against building profitable products. It is an argument for building products whose profitability and whose human value are aligned — because that alignment is what makes products last.
The birthparent search example is extreme by design. Most of your users are not arriving at peak emotional vulnerability. But all of them arrive with something — a job to do, a problem to solve, a hope that your product will help them. Every one of them is making a moment-by-moment assessment of whether you are on their side.
The framework gives you a way to be honest about when you are and when you are not. That honesty is not just ethical. It is strategic. Trust compounds slowly and ruptures instantly, and the products that understand this build the kind of loyalty that no growth hack can manufacture.
The person on the other side of the screen is not a conversion rate. They are a human being with a specific emotional state at a specific moment. Design for that person, at that moment, and the business will follow.
Coupled Value Theory (CVT) was developed as the design framework for Kindred Search, an application helping people find their birthparents. It builds on dark patterns research (Brignull, Gray et al.), Value Sensitive Design (Friedman, Kahn & Borning), and principal-agent theory, while contributing a measurement asymmetry principle, a continuous coupling coefficient, and a five-type coupling taxonomy that are, as far as the author can determine, not present in adjacent literature. The full CVT design specification — including the complete coupling taxonomy, all five heuristics with worked examples, and a flow audit template — is available as an open design document.
If this framework is useful to your product work, the author welcomes responses and applications.