Why Citizen Satisfaction Scores Mislead Public Agencies

A citizen renews her passport in eleven minutes, with a clean counter and a clerk who smiles. She ticks "very satisfied" on the exit survey. She would have ticked the same box after a ninety-minute wait, because the alternative to a slow passport office was never a better passport office — it was no passport office at all. That single fact should worry every government measuring itself by citizen satisfaction scores.

Citizen satisfaction surveys largely measure relief, not quality, wherever citizens have no alternative provider to compare against. A more honest measurement system for public services tracks four separate signals — effort required, resolution achieved, trust earned, and dignity preserved — captured at the touchpoint level rather than bolted onto the end of a transaction as a single happy-face question. Satisfaction alone tells an agency almost nothing about where the service actually breaks.

Why do citizen satisfaction scores mislead public agencies?

Because the market conditions that make satisfaction meaningful in the private sector don't exist in most public services. A retail bank losing customers to a rival feels it in churn within a quarter. A passport authority, a land registry, or a municipal licensing office rarely has that pressure — the citizen cannot switch providers, and often cannot opt out of the transaction at all. Economists call this a captive market, and it distorts self-reported satisfaction in a very specific way: people rate a monopoly service against their fear of the alternative, not against an achievable standard.

This isn't a fringe observation. The American Customer Satisfaction Index, which has tracked cross-sector satisfaction in the United States since the early 1990s, has for years placed the federal government sector near the bottom of its rankings — even in a market where citizens have no exit option and every incentive to rate leniently. If satisfaction is depressed even under captive conditions, the honest conclusion is that the underlying experience is worse than the scores in most agencies' own dashboards suggest. Governments that only track their own internal satisfaction metric, without a cross-sector benchmark, are grading themselves on a curve nobody else can see.

The structural problem compounds a measurement problem: a single post-transaction satisfaction question collapses an entire journey — booking, waiting, document checks, payment, follow-up — into one number. It cannot tell a service owner whether the frustration happened at intake, at the counter, or in a backend process the citizen never saw. That is the same blind spot private-sector CX teams solved years ago by moving from a single NPS or CSAT score to full journey-level mapping — a practice public agencies have been slower to adopt, largely because "citizen journey" still sounds like a private-sector import rather than a civic obligation.

What does behavioral economics reveal about inflated public-sector ratings?

Two mechanisms explain why citizens rate captive services more generously than the experience deserves, and both come straight from behavioral economics rather than public-administration theory.

The first is loss aversion, the finding from Daniel Kahneman and Amos Tversky's prospect theory that people weigh potential losses roughly twice as heavily as equivalent gains. Applied to public services, a citizen doesn't rate the passport office against an ideal service; they rate it against the terror of it getting worse, closing, or becoming even less accessible. Anything that clears the low bar of "it worked and didn't get worse" reads as satisfactory, even when the objective effort involved — the forms, the queue, the follow-up call three weeks later — was substantial. Status quo bias reinforces this: an established, if imperfect, service centre feels safer to endorse than to criticise, because criticism carries an implicit risk of change the citizen didn't ask for.

The second is the peak-end rule, also from Kahneman's work on experienced utility: people judge an entire experience mainly by its most intense moment and how it ends, not by the average of every moment along the way. A forty-minute wait followed by a warm, competent final interaction at the counter can produce a "very satisfied" rating that erases the queue entirely — a pattern the Nielsen Norman Group has documented extensively in its research on the peak-end rule in service and digital experience. That is not a flaw in citizens' judgment; it is how memory works. But it means a satisfaction score collected at the exit door is structurally biased toward whatever happened in the final ninety seconds, and blind to the eighty-nine minutes before it.

Put the two together and you get the exact pattern seen across most citizen satisfaction dashboards: consistently high scores, low variance, and almost no correlation with independently observed service quality. The score isn't lying. It's answering a different question than the one the agency thinks it's asking.

What should governments measure instead of satisfaction?

Satisfaction shouldn't disappear from the dashboard — it should stop being the only entry on it. A more diagnostic model tracks four distinct signals, each answering a question satisfaction cannot.

• Effort — how much work, time, and cognitive load did the citizen have to absorb to complete the transaction? This is the public-sector equivalent of the Customer Effort Score, the metric Matthew Dixon, Karen Freeman, and Nicholas Toman championed in their influential 2010 Harvard Business Review article "Stop Trying to Delight Your Customers", which found effort reduction a stronger predictor of loyalty than delight across the service interactions they studied. In government, effort is measurable in concrete terms: number of documents requested more than once, number of physical visits required, number of different departments a citizen had to contact to resolve one issue.
• Resolution — was the citizen's actual problem solved on this contact, or did it bounce to a second visit, a callback, or an appeal? First-contact resolution is one of the few metrics in public services that correlates directly with operational cost, because every unresolved case becomes a repeat transaction somewhere in the system.
• Trust — does the citizen believe the agency will do what it says, treat their case fairly, and protect their data? Trust is a slower-moving, higher-stakes signal than satisfaction, and it is the one the OECD tracks explicitly in its biennial Government at a Glance report, which benchmarks citizen trust in public institutions across member countries as a distinct measure from service satisfaction.
• Dignity — was the citizen treated as a capable adult with a legitimate claim, or as a suspect to be processed? This is rarely on a survey at all, yet it is the variable most often cited in qualitative research on why citizens disengage from public services entirely rather than complain about them.

None of these four signals is inherently more "correct" than satisfaction. The point is that satisfaction alone cannot distinguish between a service that is genuinely good and a service that citizens have simply stopped expecting anything from. Effort, resolution, trust, and dignity, tracked together, can.

How can public agencies build a better citizen measurement system?

Replacing a single satisfaction question with a proper measurement system is a service-design exercise, not a survey redesign. It works best as a sequence.

1. Map the journey before you measure it. Break the service into its actual stages — awareness, booking, preparation, the transaction itself, and any follow-up — and identify every touchpoint a citizen passes through, digital and physical. Without this map, any metric you collect is anchored to an arbitrary moment, usually the exit interaction, and inherits the peak-end distortion described above.
2. Attach a different question to each stage, not one question to the whole journey. Ask about effort at the point of preparation, ask about clarity at intake, ask about resolution only after the outcome is confirmed. This is the same discipline private-sector customer experience teams use when building touchpoint-level scoring rather than a single end-of-journey NPS.
3. Instrument the operational data you already have. Repeat visits, average handling time, appeal rates, and callback volume are frequently sitting in case-management systems already, unconnected to the satisfaction survey. Pairing operational data with citizen-reported data is what turns a satisfaction score into a diagnosis.
4. Segment by citizen archetype, not just by service line. A first-time applicant, an elderly citizen without digital access, and a business owner renewing a licence for the fifth time have entirely different effort thresholds and trust baselines. Treating them as one undifferentiated respondent pool flattens the exact variation that matters most for equity and accessibility.
5. Close the loop visibly. Nothing degrades trust in a feedback system faster than a citizen reporting a problem and receiving no evidence it changed anything. Publishing what was fixed as a result of feedback — even a short quarterly note — converts a survey from a compliance exercise into a functioning voice-of-customer mechanism citizens have a reason to keep using honestly.

Agencies that skip step one and jump straight to a new survey instrument almost always end up with a better-worded version of the same single, end-of-journey question — a nicer flashlight pointed at the same narrow spot.

What role does digital government play in fixing this?

Digital channels make touchpoint-level measurement possible at a cost and speed that paper-based service centres never allowed, but only if agencies use that capability to disaggregate feedback rather than simply digitise the same single satisfaction question. The most instructive regional example is Dubai's Happiness Meter, a network of kiosks and digital prompts deployed across government service centres by Smart Dubai (now Digital Dubai) from 2016 onward, capturing a real-time reaction — happy, neutral, unhappy — at the moment a specific service interaction ends, rather than through a delayed post-visit survey. Whatever its limitations as a single-item measure, the design insight is sound: capture the reaction close to the moment it happened, at the specific touchpoint, not in a generic survey sent days later covering an entire transaction the citizen has half-forgotten.

The deeper opportunity in digital government is combining that real-time capture with the operational data digital systems generate automatically — time-to-complete, form abandonment, repeat logins, resubmissions after rejected documents. A citizen who abandons an online licence renewal three times before calling a hotline has told the agency more about effort and friction than any five-point satisfaction scale ever could. Digital transformation programmes that treat this operational trail as a feedback channel in its own right — not just as system logs for IT — get a far richer picture than one built on satisfaction surveys alone, and it's a discipline that sits squarely inside broader digital transformation work rather than being a side project for the comms team.

This is also where the behavioral lens pays for itself twice over. Because loss aversion inflates satisfaction under captive conditions, digital measurement should be designed to surface friction rather than confirm relief — asking "what took the most effort in that process?" rather than "how satisfied were you?" produces answers less contaminated by the fear of losing access to the service altogether. Choice architecture matters here too: a satisfaction prompt that defaults to a neutral midpoint, rather than a pre-selected "satisfied," removes a small but real thumb on the scale.

Does citizen trust matter more than satisfaction?

Trust is the metric that predicts whether citizens comply voluntarily, pay on time, and engage with a service proactively rather than only when forced to. Satisfaction predicts almost none of that. A citizen can rate a single transaction as satisfactory — the counter clerk was pleasant, the wait was tolerable — while holding deep distrust of the institution behind it, formed over years of inconsistent rulings, opaque appeals, or data mishandled once and never explained. That gap between transactional satisfaction and institutional trust is precisely why the OECD tracks them as separate constructs in its cross-country government reporting rather than treating trust as a downstream consequence of satisfied transactions.

For a service leader inside government, the practical implication is to stop treating a rising satisfaction average as evidence the institution is winning. It may simply mean citizens have adjusted their expectations downward — the same adaptive process that keeps captive-market ratings artificially high. Trust, by contrast, moves only when citizens see consistent, fair, and transparent behaviour repeated across many interactions, which is exactly why it belongs in CX governance conversations at the leadership level, not left as a line item on a quarterly survey report nobody outside the contact centre reads.

Where citizen measurement goes from here

The agencies that get this right in the next few years won't be the ones with the highest satisfaction scores. They'll be the ones that can say, precisely, where in a citizen's journey effort spikes, where resolution fails, and where trust quietly erodes — and can show a paper trail of what they changed as a result. A satisfaction score with nowhere to point is a number in search of a decision. Measurement that respects how citizens actually experience a captive service, rather than how a survey designer wishes they would, is the difference between a government that reports satisfaction and one that earns it.

Renascence works with public-sector organisations across the region on exactly this shift — from single-metric satisfaction tracking to full journey measurement, citizen feedback management, and the governance structures that turn feedback into service redesign. If your agency wants a clearer read on where it stands, the CX Maturity Assessment is a useful starting point, and our work in public services experience goes deeper into the operating models that make it stick. For related reading, see how agencies are moving from listening to action with VoC data and our companion piece on measuring satisfaction with public services.