A longer look at the idea from an Auckland University Open Day lecture that’s been rattling around my head since — and why it explains, better than any feature list could, why HappyOrNot exists.
In the last weekend of August, I sat in on an Economics 101 taster session at Auckland Uni’s Open Day. The lecturer opened with a line that’s stuck with me since: economics only exists because we’re different. If every person, firm, or system held identical skills, resources and wants, there would be nothing to trade and nothing to gain from working together. Value gets created precisely at the points where we differ.
It’s a tidy idea for a 30-minute taster lecture. But the more I sat with it against 12 years of helping organisations measure customer and employee experience, the more it stopped feeling like a classroom device and started feeling like a genuinely useful lens — one that explains something most CX conversations skip past entirely: not how to measure feedback better, but who gets to be measured at all.
So, I began asking questions from an economics lens, and used Claude to help me find relevant articles, theories and summarise them for me so that I can digest and test my thoughst against. I have included the links to the articles that were identifed, if you want to further deeper dive like I did.
This is what I landed upon and it just shows how Economics plays a pivotal role in the world.
The problem economics actually describes here
Most customer feedback tools, the platforms tracking NPS, CSAT, and full journey maps share a structural starting point: they need to know who you are before they can ask you anything. An email address, a booking reference, a loyalty number. That requirement does two things at once, and both are worth naming properly rather than waving at.
First, it filters the population down to people who were identified as customers in the first place, which quietly excludes anyone who walked in, looked around, and left; anyone who almost bought and didn’t; anyone interacting with a service anonymously (a hospital corridor, a public library, a council counter). Second, of the people who were identified, only a fraction bother to respond and it’s rarely a random fraction.
This isn’t a vague hunch. It’s one of the best-documented problems in applied economics and statistics, going by the name sample selection bias. Economist James Heckman won a Nobel Prize in 2000 largely for formalising it: when the people who show up in your sample aren’t a random draw from the population you actually care about, analysing that sample as if it were representative produces systematically wrong conclusions — no matter how sophisticated the analysis layered on top of it gets. Heckman was writing about labour markets and wages, but the mathematics doesn’t care what you’re measuring. A CSAT score built entirely from people who opted to reply is the textbook case.
Why it’s always the extremes who reply
There’s a second, older piece of economics that explains why the people who do respond skew toward the emotionally loud. Mancur Olson’s 1965 book The Logic of Collective Action asked why some groups organise and act while others, even much larger ones, stay quiet. His answer: people act when their individual stake is high enough to be worth the effort, and stay passive when it isn’t. A furious customer or a delighted one has a high enough personal stake to spend two minutes on a form. Someone whose experience was simply fine has almost none — the effort of replying outweighs any benefit they’d get from doing so. Olson was writing about unions and lobbying, but the free-rider logic transfers cleanly: the quiet, satisfied-enough majority isn’t being deliberately excluded from a CSAT score. They’re behaving exactly as rational-actor theory predicts they will — by not bothering.
Anthony Downs made a related point about voters in An Economic Theory of Democracy (1957): most people rationally choose not to gather information or participate in a system where their individual input barely moves the outcome. Substitute “leave feedback” for “vote,” and you get the same quiet-majority problem showing up in an entirely different domain, decades before “customer experience” was a job title.
The knowledge problem, not just the response problem
Friedrich Hayek’s 1945 essay The Use of Knowledge in Society is usually read as an argument about central planning versus markets, but its actual claim is broader and more useful here: valuable knowledge is inherently dispersed across huge numbers of individuals, each holding a small, specific, often tacit piece of it, and no single central observer can gather it all through one channel. Hayek’s answer for markets was the price mechanism, a decentralised signal that aggregates dispersed information without needing anyone to explain themselves in words.
Customer experience has its own version of this problem. The “why” behind an organisation’s performance isn’t sitting with the ten people who filed a complaint, it’s dispersed across thousands of small, unremarkable moments that never get reported anywhere, because no one asked, and even if someone had, most of those people wouldn’t have felt it was worth stopping for. A system that only listens through one identified, effortful channel is trying to solve Hayek’s knowledge problem with exactly the kind of single central observer he argued couldn’t work.
Why more features doesn’t fix a different problem
Here’s where the economics stops being background reading and starts explaining a genuine strategic choice. Most of the well-established platforms in the CX space compete with each other on the same axis: more journey-mapping stages, more integrations, richer dashboards, better analytics layered on top of the responses they collect. That’s a real and valuable kind of competition, but it’s all downstream of the same original sample, the identified and sufficiently-motivated respondents Heckman’s and Olson’s work describes.
Ricardo’s theory of comparative advantage, first laid out in On the Principles of Political Economy and Taxation (1817), is normally taught as a trade story between countries, but its underlying logic is simpler than that: gains come from specialising in what you’re relatively better at, and trading for the rest, rather than trying to do everything yourself. HappyOrNot’s position in the CX ecosystem reads the same way. It doesn’t try to out-feature the journey-mapping platforms on journey mapping, or the enterprise analytics suites on dashboards. It specialises in the one part of the chain none of them were built to solve, a near-zero-effort way to hear from literally anyone who moved through a space or a transaction, customer or not, happy, indifferent, or annoyed. Then feeds that signal into whatever systems an organisation already runs for the rest.
That specialisation is also, not coincidentally, the reason HappyOrNot doesn’t require knowing who you are at all. Ronald Coase’s work on transaction costs (The Nature of the Firm, 1937) is about why firms exist rather than every transaction happening on the open market. The core insight generalises: reducing the cost of participating in an exchange changes who participates in it. Requiring an email address and a follow-up click is a transaction cost. Removing it doesn’t just get “more” responses it changes whose responses you get, pulling in the Olson-quiet majority who would never have paid the cost of a full survey.
What this means for an organisation that already has NPS or CSAT running
TNone of this is an argument against NPS, CSAT, or journey mapping. They answer a real question well: what do the people who chose to speak up think, in detail? The point Heckman, Olson, Downs and Hayek collectively make is that this is a genuinely different question from “what does everyone think,” and no amount of refinement on the first question closes the gap to the second. An organisation with a highly mature CX platform still has exactly the same unaddressed gap sitting underneath it, because that platform was never built to answer the second question in the first place.
That’s really the case for something like HappyOrNot, not instead of what’s already running, but alongside it, because the gap doesn’t close on its own no matter how good the rest of the stack gets.
Where this leaves me
I didn’t expect a Saturday morning at a university open day to hand me a better vocabulary for something I already believed. But the economics is oddly precise about it: difference isn’t just a nice-sounding value, it’s the reason feedback systems built around identified, motivated respondents will always describe a real but partial population and why closing that gap needs a fundamentally different starting point, not a better version of the same one.