Letter No. 47 · May 2026 · An essay

Customer obsession is an operating model.

Most CX programmes stall at the dashboard. The score is up on the wall, the steering committee meets monthly, the board asks how it's moving, and it does not move. The truth is quieter: there is no operating model underneath the dashboard, and so nothing converts attention into action.

Silvia Montero Collado executive advisor on customer experience
14 min read 1,850 words Published May 12, 2026

A CX programme is two years in. The score is up. The board has been told the trend three times. The CEO has retold it once. And then somebody in the operating committee, usually the CFO, occasionally a regional MD, asks the question that breaks the meeting open: is anything actually different on the ground? The honest answer is rarely yes. The score moved, the business did not, and nobody around the table can quite say why.

This is the failure mode the practice has watched, again and again, across retail banks, B2B software companies, and consumer-services groups. It is not a measurement failure. It is not a culture failure. It is a quieter thing: the absence of an operating model beneath the dashboard. Customer obsession, properly understood, is not a value or a posture or a quarterly campaign. It is a model that converts a piece of customer attention into a decision, a decision into an action, an action into a change the customer actually feels.

If that conversion chain is missing or broken, no amount of score-watching will rescue it.

The dashboard is the output of an operating model. Not the model itself, and never its substitute. from the practice, on the operating model

The five questions.

What follows is the small set of questions a serious operator (a chief customer officer, a head of service, a managing director who has inherited the CX file) should be able to answer in writing, before the next steering meeting. The questions are deliberately boring. The honest answers, rarely.

  1. What signal are we listening for, and from whom? Not "what survey are we running." That is method, not signal. Signal is the precise piece of customer experience we believe carries strategic information: the post-onboarding call, the first invoice query, the early-termination interview.
  2. Where is that signal heard, and by which role? A signal heard only by the survey team is not heard by the business. Name the role, name the meeting, name the date.
  3. What decisions does the signal change? If the answer is "we'll review it," there is no operating model. A signal that changes nothing is a metric, not a model.
  4. Who acts, and by when? Action without an owner and a date is a slogan.
  5. How does the customer know we acted? The closing of the loop is the experience. It is also, almost always, the part that is skipped.

Five questions. Most CX programmes can answer the first one. A good number can answer the second. The third is where the wheels come off, because the third question forces the programme to declare what it is for, and most programmes have never been asked to do that out loud.

3/5

The third question, what decisions does the signal change?, is the one most programmes cannot answer in a sentence. Until they can, the rest is theatre.

Why the dashboard survives.

The dashboard is comforting because it has the shape of a model without being one. It shows a number, the number moves, and movement looks like progress. Boards read it. CFOs tolerate it. CEOs cite it in earnings calls. And underneath the comforting movement, the operating model, the actual conversion chain from signal to action, may not exist at all.

This is not malicious. It is the natural drift of a programme that was set up to track rather than to operate. Trackers do not need an operating model; they need a steady source of numbers and a willing audience. Operators do, and the cost of pretending to be one while only being the other is the lost decade many CX programmes will look back on.


What changes, if the model is there.

Three small things, mostly. None of them photograph well in a deck.

None of this requires more dashboards. Most of it requires fewer. The discipline is in the model, not the display.

A note on scale.

The five questions are scale-invariant. A two-hundred-person business answers them in a paragraph. A multinational answers them in a thicker book. The shape is the same: signal, hearing, decision, action, feedback to the customer. If your version has more than five questions, you have added decoration. If it has fewer, you have skipped something the customer will notice.

The practice keeps the one-pager (five questions, a worked example, two empty columns for next steering) printed on the back of every engagement folder. It is the most-borrowed thing in the room.


Customer obsession, properly understood, is not louder than other operating disciplines. It is quieter. It runs on writing, not on slogans. It survives change in leadership because it is written down. And it produces the only outcome that has ever mattered in this work: the customer notices that something is different, and the business is the reason.

The next letter will sit with the third question for longer, and look at how a steering committee should read it. If you would like it in your inbox the day it is published, the subscription is below.

CX operating model Frameworks Board credibility Steering committees
working with Silvia

The advisory room.

If the five questions are useful and you'd like to think about them at the level of your own programme, the practice takes a small number of executive engagements each quarter.

subscribe

The next letter, on the third question.

Arrives on the second Tuesday of June. No marketing, no more than that.

More from the operating-model series

See all 14 →
No. 44

Channel programmes the country actually uses.

Essay · 13 min · Jan 2026
No. 39

When the score moves and the business doesn't.

Essay · 10 min · Aug 2025
No. 35

The operating-model diagnostic: twelve questions.

Framework · 8 min · Apr 2025
← previous letter Local readiness, not localisation. next letter → The service desk after AI.