Finding and Fixing Moments of Truth in the Customer Journey

A bank's mobile app scores 4.8 stars on the app store. Net Promoter Score sits comfortably above the industry benchmark. And yet customers still leave in droves the moment they try to close an account by phone. Nobody measured that call. Nobody blueprinted it. It sat outside every dashboard the executive team trusted — and it was quietly deciding who stayed and who walked.

That call is a moment of truth: the point where a customer's opinion of you is made or unmade in seconds, largely irrespective of everything that came before it. Most organisations think they know where theirs are. Most are wrong, because they go looking for moments of truth with the wrong instrument — surveys and gut feel — instead of a structured service blueprint that follows the customer, the frontline, and the back office at the same time.

Here is the argument this piece will make and defend: a moment of truth is not the touchpoint that generates the most complaints — it's the touchpoint that carries the most consequence for the relationship, and those two things are rarely the same. Finding the real ones requires blueprinting the full service system, not just the customer-facing surface. Fixing them requires prioritising by emotional weight, not average satisfaction. And keeping them fixed requires accepting that moments of truth migrate as journeys change — a solved moment of truth is a temporary state, not a permanent one.

What is a "moment of truth" in customer experience, exactly?

The term comes from Jan Carlzon, the chief executive who turned around Scandinavian Airlines System (SAS) in the 1980s and later wrote about it in his book Moments of Truth (1987). Carlzon's insight, built on service-management thinking developed with the strategist Richard Normann, was blunt: every one of SAS's roughly 10 million customers came into contact with about five SAS employees for an average of 15 seconds each. The company's reputation, he argued, was decided in those 50 million quick encounters — not in the boardroom, not in the marketing campaign, and not in the annual report.

A moment of truth, then, is any interaction where a customer forms or revises a judgement about you — often quickly, often emotionally, and often on evidence that has nothing to do with your core product. The mechanism behind why these moments carry disproportionate weight is what behavioural economists call the affect heuristic: under time pressure or uncertainty, people substitute a feeling ("that was easy" or "that felt like a fight") for a reasoned evaluation. Customers don't average their forty interactions with you. They remember a handful of charged ones and build a story from those.

Why do journey maps keep missing the moments that actually matter?

Because most teams find moments of truth by asking, rather than measuring — and the loudest answer is rarely the most consequential one. Ask a contact centre manager where the pain is and you'll hear about the queue. Ask customers in a survey and you'll hear about the last thing that annoyed them, which the availability heuristic guarantees will be recent and vivid rather than representative. Neither method surfaces the touchpoint that quietly drives cancellations six weeks later — the one nobody complains about because dissatisfied customers don't file a ticket, they just leave.

This is the gap between a perceived moment of truth and a measured one. A perceived MoT is whatever the organisation currently believes matters, usually inherited from whoever shouted loudest in the last steering committee. A measured MoT is identified by tracking where emotional variance — the swing between a customer's expectation and their experience — is largest across the actual journey, using a genuine service blueprint rather than a tidied-up journey map drawn from memory. The Nielsen Norman Group has documented this exact failure mode in its guidance on journey mapping: maps built from internal assumption rather than observed behaviour routinely misplace the moments that matter most.

The distinction matters commercially. A telecom's billing-error call might generate ten times the complaint volume of its SIM-activation failure — but if activation failure is what causes silent churn before the relationship even starts, fixing billing complaints first is optimising for the wrong variable entirely.

How do you actually find the moments of truth that move the needle?

You find them by building a full service blueprint — the method Lynn Shostack introduced in her 1984 Harvard Business Review article "Designing Services That Deliver" — and then layering emotional and operational data onto it, rather than relying on a single-lane customer journey map. The blueprint forces you to see the front-stage interaction and the backstage process failing it at the same moment, which is where most moments of truth are actually manufactured.

1. Blueprint the full system, not just the customer's view. Map the journey across four lanes: customer actions, frontline/onstage actions, backstage support actions, and the underlying processes and systems. A "moment of truth" at the counter is frequently the visible symptom of a backstage failure three steps upstream — a stock system that lags reality, an approval workflow with no SLA, a policy nobody frontline can override.
2. Layer every evidence source you have onto the same blueprint. Voice of customer verbatims, CSAT and Customer Effort Score by touchpoint, complaint volume, abandonment and drop-off data, and frontline escalation logs all belong on the same map. A moment of truth reveals itself where several of these sources spike at the same touchpoint, not where any single one does.
3. Plot the emotional arc across the sequence. Score each touchpoint for the gap between what the customer expected and what they experienced, then chart it as a line across the journey. The peaks, the troughs, and — critically — the ending are your candidate moments of truth. A flat line with one violent dip is a completely different diagnosis than a slow, steady erosion of goodwill, and each demands a different fix.
4. Validate with the people living it, on both sides. Sit with frontline staff and watch a real interaction, not a described one; frontline teams routinely know exactly where the process breaks but have never been asked in a format that reaches decision-makers. Pair that with a handful of real customer interviews at the candidate moments — five well-chosen conversations will tell you more than five hundred survey responses about why a moment breaks trust.
5. Score for consequence, not just intensity. A touchpoint that generates fury but has no bearing on renewal or referral is an irritant, not a moment of truth. Cross-reference each candidate against downstream behaviour — churn, upsell, referral, complaint escalation — before you commit resource to fixing it.

This is, in effect, the same discipline as hunting for operational bottlenecks — the two exercises share a method even though they answer slightly different questions, as we've argued in more detail in Finding the Bottlenecks That Actually Hurt Customers. A well-run service design engagement typically runs these two lenses — bottleneck and moment-of-truth — side by side on the same blueprint, because friction and emotional consequence don't always live at the same touchpoint.

Why does the peak-end rule change which moments you fix first?

Because it tells you, with some precision, which moments the customer's memory will actually retain — and it is memory, not the lived average, that drives repeat behaviour. In their landmark study When More Pain Is Preferred to Less: Adding a Better End, published in Psychological Science in 1993, Daniel Kahneman, Barbara Fredrickson, Charles Schreiber and Donald Redelmeier found that patients undergoing colonoscopies rated a longer procedure with a gentler ending as less unpleasant overall than a shorter one that ended abruptly at peak discomfort — even though the longer procedure involved strictly more total pain. Memory doesn't average an experience. It anchors on the most intense moment and the final moment, and largely discards the rest.

Applied to a service journey, this is liberating and disciplining in equal measure. You do not need to smooth every touchpoint to a uniform, pleasant hum — that's an expensive and largely wasted pursuit. You need to identify the peak (the most emotionally intense moment, positive or negative) and the end (however the interaction closes), and put your redesign budget there first. A telecom that fixes the tone of its cancellation call and adds one memorable gesture at the close of a service recovery will move perception further than a telecom that shaves twenty seconds off average handle time across every call in the queue.

The peak-end rule doesn't ask you to fix everything. It asks you to fix the right two things — and stop pretending the other forty are equally important.

This is also where loss aversion earns its place in the diagnosis. A negative moment of truth is weighted roughly twice as heavily in memory as an equivalent positive one — the foundational finding behind Kahneman and Amos Tversky's prospect theory. Practically, that means a single badly handled complaint can erase the goodwill built by several genuinely good interactions, which is precisely why service recovery at the moment of failure deserves a disproportionate share of your design attention.

Why does fixing a moment of truth once not keep it fixed?

Because moments of truth migrate, and most organisations design as though they don't. The classic version of this failure: a retailer spends eighteen months perfecting its in-store return experience, only to discover that the moment now deciding loyalty has quietly moved to the WhatsApp support thread customers open instead of visiting the store at all. The store team never sees the new moment of truth. Nobody re-blueprints the journey to check.

Three forces drive this migration, and a service design practice should expect all three, permanently:

• Channel shift. Every time a self-service option launches, some proportion of customers route around your carefully redesigned human touchpoint entirely, creating a new moment of truth in the channel you designed last — usually with less oversight, because nobody thought to treat a chatbot flow with the same rigour as a branch counter.
• Expectation inflation. Once you fix a moment of truth, customers recalibrate what "good" means and the bar rises. The goal-gradient effect works against you here too: customers who feel close to a resolved outcome become more sensitive to any remaining friction, not less.
• Competitive reset. A moment of truth is partly relative. If a competitor removes a friction point in the category, the equivalent point in your journey becomes newly painful by comparison, even though nothing in your process changed.

The practical implication is that moment-of-truth discovery is not a project with an end date — it is a standing capability, ideally reviewed on the same cadence as your journey mapping and voice-of-customer programmes, not a one-off workshop that produces a poster for the office wall.

What does fixing a moment of truth actually look like?

It rarely looks like a single grand redesign. It usually looks like a precise intervention at one of four levers, applied to the specific touchpoint the blueprint identified — not a broad "improve the experience" initiative that touches everything and moves nothing.

• Policy. Remove the rule that forces a frontline agent to say no when a reasonable customer expects yes — often the single highest-leverage fix, because it costs nothing to implement and everything to leave in place.
• Process. Redesign the backstage sequence feeding the moment, using the blueprint to find where handoffs, approvals, or missing data are the true cause of the front-stage failure.
• People. Give frontline staff the authority and the script latitude to close the moment well — a scripted apology reads as sludge; a genuinely empowered one reads as care. This is where frontline enablement and moment-of-truth design are really the same project wearing two names.
• Proof. Add a small, visible signal at the moment itself — a confirmation, a status update, a human acknowledgement — that closes the loop the customer opened. Uncertainty is often what makes a moment feel like a moment of truth in the first place; resolving the uncertainty is sometimes the entire fix.

Whichever lever you pull, sequence the fix so the customer feels resolution before they feel the process behind it — the mechanism, not the moment, should stay backstage. And build the improvement into a blueprint that gets revisited, not a one-time patch: an organisation with a mature CX governance model treats moment-of-truth review as a standing agenda item, with clear ownership for re-scoring the journey every time a channel, policy, or competitor moves.

Where should you start if you've never mapped this rigorously?

Start narrow and start honest. Pick one journey with real commercial weight — onboarding, renewal, or complaint resolution are usually the highest-yield candidates — and build one proper blueprint, all four lanes, before you attempt to scale the method across the portfolio. Running a structured CX maturity assessment first will tell you whether your organisation even has the data infrastructure to score touchpoints by consequence rather than volume, which is the precondition for doing any of this well rather than just redrawing the same assumptions in a nicer template.

The organisations that get this right treat moment-of-truth discovery the way a good pilot treats a pre-flight checklist: not a creative exercise, but a disciplined, repeatable inspection of the system, run often enough that nothing gets to hide in it for long.

Carlzon's SAS turnaround worked not because someone identified moments of truth once, but because the organisation built the habit of looking for them continuously, at every level, as the business itself kept changing. Four decades on, the discipline hasn't dated. Only the channels have.