9 min read · 1,815 words
The cost of poor customer experience is real, but it is almost never the number in the headline. Qualtrics puts nearly three trillion dollars of global sales at risk in 2026. That figure will not survive one question from a CFO. Your own arithmetic will, and it is smaller.
Some years ago, in an industrial equipment business, I watched a long-standing account go quiet. No complaint, no escalation, no angry email. Orders thinned over roughly two quarters and stopped. The sales note recorded the reason as price.
It wasn’t price. The trigger was a documentation failure on a shipment, the kind of thing closed as an operations ticket inside a day, which never appears in a satisfaction score or on any slide with the word experience on it. By the time anyone joined the two facts together, the account had been buying elsewhere for a year. In eight years running marketing for Asia, the Middle East, Africa and Turkey at CNH Industrial, I saw some version of that in four markets.
That is the gap worth writing about. Our industry publishes enormous numbers about what bad experience costs. None help you find the account you are about to lose.
What do the headline numbers actually say?
They say the aggregate is huge, and that it keeps moving. Qualtrics XM Institute puts nearly $3 trillion of global sales at risk in 2026, split into $2.1 trillion of reduced spending and $865 billion abandoned outright, from 20,001 consumers across 14 countries surveyed in Q3 2025.
The mechanism is simple. After a negative experience, 34% of consumers reduce spending with that company and 13% stop altogether: one in two damaged relationships producing a revenue consequence.
Now look at the same programme four years earlier. In 2022 it put $3.1 trillion at risk, framed as 6.7% of revenue, with 50% of consumers reducing or cutting spend. Between those editions the world economy grew considerably. The at-risk number went down.
Why the trillion-dollar number collapses in a budget meeting
Because it is a macro extrapolation from self-reported intent, and every CFO I have presented to knows the difference between what people say they will do and what they do.
Emplifi is the cleanest illustration. In 2025 it reported that 70% of consumers will abandon a brand after two negative experiences, roughly a quarter after one. In 2022 the same firm reported 86% after two to three. Sixteen points in three years, on a hypothetical. In that 2022 study, 49% said they had actually left a brand they were loyal to in the previous twelve months. Stated intent 86%, revealed behaviour 49%. Only one of those is a business input.
Retention economics have the same problem. The famous line that acquiring a customer costs five to 25 times more than retaining one comes from a 2014 Harvard Business Review piece citing Bain. The companion claim, that a 5% lift in retention raises profits 25% to 95%, traces to Fred Reichheld, whose own Bain paper says something narrower: in financial services, a 5% increase in retention produces more than a 25% increase in profit. A sector finding from 2001, quoted at steel plants and freight companies in 2026 as though it were a law of physics.
The trillion-dollar headline proves customer experience matters somewhere in the world economy, and proves nothing at all about your business.
The three inputs your own cost-of-failure number needs
A defensible number has three inputs, all from systems you already own.
| Input | Where the number comes from | Worked example |
|---|---|---|
| Failure exposure | Your complaint log, credit notes, quality claims and delivery-miss reports. Accounts hitting at least one customer-visible failure in twelve months | 120 of 400 active accounts, or 30% |
| Defection rate on failure | Cohort your own billing data. What those accounts spent in the twelve months after the failure against the twelve before | 13% stopped entirely, 34% cut spend by about a fifth |
| Margin at risk per account | Annual gross margin per account, never revenue, times the years it would otherwise have stayed | Rs 43.2 lakh a year, three-year horizon |
The third input is where most marketers give the argument away. Quoting revenue makes the number look bigger and you look like someone who does not read a P and L.
What the arithmetic looks like when you actually run it
Take a mid-size industrial supplier with 400 active accounts averaging Rs 2.4 crore a year. That is Rs 960 crore of revenue at an 18% gross margin: Rs 172.8 crore of margin, Rs 43.2 lakh per account per year.
| Line | Arithmetic | Gross margin lost, year one |
|---|---|---|
| Accounts lost outright | 120 failures x 13% = 16 accounts x Rs 43.2 lakh | Rs 6.9 crore |
| Accounts that cut back | 120 failures x 34% = 41 accounts x Rs 43.2 lakh x 20% | Rs 3.5 crore |
| Total exposure | Sum of the two lines above | Rs 10.4 crore |
| Share of total gross margin | Rs 10.4 crore of Rs 172.8 crore | 6.0% |
| Share of revenue | Rs 10.4 crore of Rs 960 crore | 1.1% |
Here is the part that matters. The bottom-up number is 1.1% of revenue. The headline research says 6.7%. The honest figure is roughly a sixth of the one our industry quotes, and it is the only one a finance director will fund against, because every line traces to a document someone in the building signed.
Then add replacement cost. Sixteen lost accounts must be won back as new business at whatever your acquisition cost is. If the five-to-25x range is even directionally right, recovery spend dwarfs prevention spend. That is the budget argument. Not the trillions.
What does the cost of poor customer experience look like in B2B?
Concentrated, delayed and silent. A consumer brand losing one shopper loses a rounding error. An industrial supplier losing one account can lose a percentage point of margin, because in most B2B books the top twenty accounts carry a disproportionate share.
The delay is the second problem. B2B contracts run on annual or multi-year cycles, so a March failure surfaces as a December non-renewal, by which point it is no longer the story anyone tells.
The silence is the third and worst. Fewer than one in three consumers give any feedback to the company, and B2B buyers are more reticent still, because complaining carries political cost inside their own organisation. This is the blind spot behind what NPS never tells you about a B2B account. A nine from a procurement contact who is quietly running a tender is not a signal, it is noise with a decimal point.
The Indian evidence is specific and recent. ServiceNow’s third India CX report, published in March 2026 from over 5,000 Indian consumers and 425 service professionals, found Indians lose 10.8 hours a year resolving service issues, 15 billion hours collectively, with 44% prepared to switch brands when dissatisfied. Zendesk’s 2025 work puts it higher, at 70% of Indian consumers saying they would switch after a single poor experience.
Is the experience gap closing or widening?
Widening, and Asia is the worst of it. Forrester’s 2025 Global CX Index found 21% of brands declined and only 6% improved across 275,000 customers and 469 brands. In Asia Pacific, 37% of brand scores fell.
The perception gap is worse than the performance gap. PwC’s 2025 survey of 5,511 consumers and 406 executives found 89% of executives believe customer loyalty has grown. Only 40% of consumers agree. In the same study, 52% had stopped buying from a brand after a bad product or service experience.
That 49-point gap between what leadership believes and what customers report is, to me, the most useful statistic in the field. It explains why the CX budget conversation is so hard. You are not arguing about money. You are arguing with a leadership team that sincerely believes the problem is solved.
What to do with this on Monday
Stop quoting the trillions. Build the three-input number instead, in this order.
- Pull twelve months of failure records from operations, quality and logistics. Not the CX team’s data. The operational data, where real failures are logged as tickets and credit notes.
- Match them to account codes and pull billing for twelve months either side of each failure. Two days with a finance analyst, and it is the entire argument.
- Convert to gross margin, not revenue. Apply your own margin rate per segment.
- Present one number on one page. Margin lost to logged operational failure in twelve months, and the prevention cost against it.
- Name the three failure types behind most of the loss. In industrial businesses that is usually delivery reliability, documentation and after-sales response.
Step five is the same discipline as building journey maps that survive the buying committee, and it feeds the budget conversation I set out in what a B2B company should actually spend on brand. The version that worked best for me is described in a CX programme built around the accounts that mattered.
One warning. Your number will be smaller than the headline, and someone will say so. Agree, and say it is smaller because it is real.
Frequently asked questions
What is the cost of poor customer experience?
Globally, Qualtrics XM Institute estimates nearly three trillion dollars of sales at risk in 2026. For one company the useful figure is far smaller and specific: the gross margin of accounts that reduced or stopped spending in the twelve months after a logged service failure, measured against your own billing data.
How do you calculate the cost of a customer service failure?
Multiply three numbers you already hold. The count of accounts that hit a customer-visible failure in twelve months, the share of those accounts that cut or stopped spending afterwards, and the annual gross margin per account. Use margin rather than revenue, or finance will discount the entire case.
Is it really five to 25 times cheaper to retain a customer than acquire one?
That range comes from a 2014 Harvard Business Review article citing Bain, and it is directional rather than precise. The related claim about a five percent retention lift was originally a financial services finding from 2001. Both are useful as arguments and unsafe as forecasts in an industrial category.
Why is poor customer experience more expensive in B2B than in consumer markets?
Three reasons. Revenue is concentrated, so one account can be worth hundreds of consumer relationships. Contract cycles delay the consequence by months, breaking the link between cause and loss. And B2B buyers rarely complain, because raising a formal issue carries political cost inside their own organisation.
How many customers leave after a single bad experience?
Estimates vary widely by market and method. Emplifi found roughly a quarter of consumers stop after one bad experience, while Zendesk reported seventy percent of Indian consumers saying they would switch after one. Stated intent consistently overstates behaviour, so treat these as upper bounds rather than planning inputs.
The most expensive customer experience problems in B2B are not the ones customers complain about. They are the ones operations closed as tickets, correctly, on the day they happened, with nobody asking what the account did next. That data already exists in your building.
So here is my question. If you matched your last twelve months of delivery failures to account billing, what would the number be, and are you sure you want to know?
