What does bad customer service really cost you? The hidden costs

What does poor customer service really cost you? The hidden costs in 2026

TL;DR: the essentials in 30 seconds

• Biggest cost block: not support staff costs, but the revenue dissatisfied customers quietly spend elsewhere.
• The figure for 2026: the Qualtrics XM Institute estimates nearly $3 trillion in revenue at risk worldwide due to bad customer experiences.
• Silent churn: according to Zendesk benchmark data, more than half of consumers switch to a competitor after just one bad experience – many without complaining first.
• Rule of thumb: annual revenue × share of bad experiences × share with reduced spending = revenue at risk.
• Fastest lever: resolve standard requests immediately and correctly, cut waiting times and ensure a clean handover to people.

Poor customer service mainly costs you revenue you never book as a loss. According to the Qualtrics XM Institute, customers cut their spending after 47% of all bad experiences – and in online retail, after as many as 58%. For a shop with €3.6 million in annual revenue, that's around €230,000 in revenue at risk on paper, before a single complaint lands in the inbox.

Definition:
The costs of poor customer service are all direct and indirect losses that arise when customer requests are resolved too slowly, incorrectly or not at all. They include reduced spending and churn among existing customers, additional follow-up contacts in support, higher acquisition costs for replacement customers, and strain and turnover in the service team.

What are the costs of poor customer service?

The costs of poor customer service are made up of visible and hidden items – and the hidden ones weigh much more.

Most companies only measure the visible part: staff costs, ticket volume, handling time. The more expensive part lies below the surface. It only shows up months later in falling repurchase rates and a weaker customer lifetime value, i.e. the total revenue a customer generates over the entire business relationship.

How much revenue is at stake due to poor customer service?

In 2026, nearly $3 trillion in revenue is at risk worldwide because companies deliver bad customer experiences. That's the result of the latest analysis by the Qualtrics XM Institute, based on a survey of over 20,000 consumers in 14 countries in the third quarter of 2025. Germany is part of the sample.

The study provides two figures you can apply directly to your company. First, 11% of all customer experiences are bad. Second, 47% of these bad experiences lead customers to cut their spending. In online retail, this figure is 58% – well above average. Only after bad experiences with fast food do consumers hold back on purchases more often.

A second data point comes from the PwC Customer Experience Survey 2025. There, 29% of respondents say they've stopped buying from a brand online or in-store because of a bad customer experience. At the same time, 70% of the executives surveyed say their customers' expectations are growing faster than their company can adapt.

Which costs don't show up in any support report?

The most expensive consequences of poor service don't appear in any helpdesk dashboard. They're spread across revenue, marketing budget and staff, and are therefore rarely attributed to customer service.

Silent churn: customers leave without complaining

Most dissatisfied customers won't tell you they're dissatisfied. According to the Zendesk customer service statistics 2026, more than half of consumers switch to a competitor after a single bad experience, and after several bad experiences it's 73% (Zendesk benchmark data). A 2023 Coveo study cited there also shows that 56% of consumers rarely complain about a negative experience and switch quietly instead. For you, this means: a low complaint rate is no proof of satisfied customers.

Smaller baskets: customers stay but buy less

Not every disappointed customer cancels or disappears. Many order less often, choose cheaper products or buy complementary items elsewhere. Qualtrics puts this effect at around $2.1 trillion in reduced spending worldwide, compared with $865 billion that customers withdraw entirely. Smaller baskets are the bigger item – but also the hardest to measure.

Follow-up contacts: every unresolved request comes back

A request that isn't resolved at first contact generates at least one more contact. Customers write a second email, switch to the phone or follow up via chat. Each of these follow-up contacts takes up handling time again without creating any additional value. The metric that makes this visible is first contact resolution, i.e. the share of requests fully resolved at first contact.

Replacement acquisition: new customers cost more than retained ones

Every customer who leaves has to be replaced with marketing budget if revenue is to stay stable. For Bain & Company, Fred Reichheld showed that even small improvements in customer retention have a disproportionate effect on profit. His often-cited analysis, on which the Bain article Prescription for cutting costs is also based, found that if the customer retention rate rises by 5%, profit rises by 25 to 95% depending on the industry. The data is older, but the direction still holds today: retention is the cheaper growth lever.

Strain on the team: overload creates the next bad contact

When queues grow, pressure on your service team rises. Staff answer the same standard questions hundreds of times, have less time for complex cases and make more mistakes under time pressure. This creates a vicious circle: poor service creates more contacts, more contacts create more pressure, and more pressure creates poorer service.

How do you calculate the costs of poor customer service for your company?

The easiest way to calculate your revenue at risk is the Qualtrics XM Institute method: annual revenue × share of bad experiences × share of bad experiences after which customers cut their spending. The result isn't a certain loss but the revenue put at risk by poor service.

An example for a mid-sized online shop: the shop has 20,000 active customers with average annual revenue of €180 each, so €3.6 million in total revenue. With the global average of 11% bad experiences and the e-commerce figure of 58%, you get: €3,600,000 × 0.11 × 0.58 = €229,680 in revenue at risk per year. That's around 6.4% of annual revenue.

On top come the operational costs of follow-up contacts. As an example assumption, take 3,000 tickets a month, 15% of which are follow-up contacts on requests already made, at €5 handling cost per ticket. That's 450 additional tickets and €2,250 a month, or €27,000 a year. Replace these example values with your own figures from the helpdesk, as follow-up rates and cost per ticket vary widely by industry and team structure.

What are the most common causes of poor customer service?

The most common cause of bad customer experiences is problems with service delivery – cases where what was promised doesn't arrive, or doesn't arrive on time. According to the Qualtrics XM Institute, this is the most frequently cited reason globally. In customer service, it shows up in three recurring patterns.

Long waiting times arise when ticket volume grows faster than the team, for example in peak season around Black Friday or after campaigns. Repeated explanations arise when customers have to describe their issue anew every time they switch channels, because chat, email and phone don't share the same context. Unresolved requests arise when answers are fast but incomplete, so the customer gets in touch again.

All three patterns share a common root: standard requests such as order status, returns or delivery times tie up capacity that's missing for complex cases.

Does AI in customer service make the costs smaller or bigger?

AI only reduces the costs of poor customer service if it actually resolves requests rather than just answering them. A poorly set-up chatbot merely shifts the problem and can itself become a source of bad experiences.

The Qualtrics Consumer Experience Trends Report 2026 shows this too: almost one in five consumers who used AI in customer support saw no benefit at all. According to Qualtrics, this failure rate is almost four times higher than for other AI applications. At the same time, according to PwC, 86% of consumers consider human interaction moderately or very important in their brand experience.

This leads to three requirements for using AI in a way that cuts costs instead of creating them. The AI agent needs access to real system data from the shop, ERP and helpdesk so it can answer order status or return questions individually. It needs clear limits at which it hands over to a person, including the conversation so far. And it needs ongoing monitoring of metrics such as the ticket deflection rate, i.e. the share of requests resolved without human intervention.

There are different approaches in the DACH market. Helpdesk providers like Zendesk and Freshdesk bring their own AI features, which are mainly suited to teams working entirely within their ecosystem. Platforms like Cognigy tend to target large enterprises with a voice focus. Providers like moinAI or melibo rely on AI agents for chat and email with hosting in Germany and connections to shop systems such as Shopware and Shopify. Which approach fits depends on your ticket volume, your systems and your data protection requirements.

Which levers reduce the costs of poor customer service fastest?

The fastest lever is to fully resolve the most common standard requests at first contact. This reduces waiting times, follow-up contacts and team strain at the same time.

Start with a ticket analysis. Break down the requests from the last three months by topic. In many online shops, a few categories such as order status, returns and delivery times make up most of the volume. These categories are your automation potential.

Measure what customers don't say. Add repurchase rate, follow-up rate and first contact resolution to the complaint rate. Only these metrics make silent churn visible.

Automate with system access. An AI agent that only knows FAQ texts hardly resolves any individual request. Only the connection to shop, ERP and helpdesk turns an answer into a solution.

Plan the handover to people deliberately. Define for which topics, moods or customer groups the AI agent hands over to your team immediately. That way, you prevent automation itself from becoming a bad experience.

Online retailer Deuba shows what this looks like in practice: the company says it saves over €5,000 a month in customer service with melibo and has 40% fewer tickets, because recurring requests are answered automatically.

Frequently asked questions about the costs of poor customer service

How much revenue does a company lose due to poor customer service?

That depends on the industry and customer base. The Qualtrics XM Institute estimates global revenue at risk in 2026 at nearly $3 trillion. For individual companies, the figure can be estimated using the formula annual revenue × share of bad experiences × share with reduced spending. In the e-commerce example with €3.6 million in revenue, that comes to around €230,000.

How many customers switch after a bad service experience?

According to Zendesk benchmark data (as of 2026), more than half of consumers switch to a competitor after a single bad experience, and after several bad experiences it's 73%. In the PwC Customer Experience Survey 2025, 29% say they've stopped buying from a brand because of a bad customer experience.

Why don't dissatisfied customers complain?

Many customers consider complaining more effort than switching, especially in online retail with many alternatives. According to a Coveo study cited by Zendesk, 56% of consumers rarely complain and switch quietly. That's why repurchase rate and follow-up rate are more meaningful than the complaint rate alone.

Can an AI chatbot cause poor customer service?

Yes. A chatbot without access to order data and without handover to people creates frustration and follow-up contacts. According to Qualtrics, almost one in five consumers who used AI in support saw no benefit in it. AI only cuts costs if it actually resolves requests and hands complex cases over cleanly to the service team.

Which metrics show the costs of poor customer service?

The most important metrics are repurchase rate, customer lifetime value, first contact resolution, follow-up rate and average waiting time. Together, they show whether customers come back after a service contact and how much capacity unresolved requests tie up.

Conclusion

Only a small part of the costs of poor customer service arises in support itself. The larger part shows up in smaller baskets, silent churn and expensive replacement acquisition – exactly where hardly any company attributes it to service. With the Qualtrics XM Institute formula, you can estimate this amount for your company in a few minutes. In the e-commerce example, it's over 6% of annual revenue.

So if you want to reduce costs, don't start by cutting staff – fully resolve standard requests at first contact and organize the handover to people properly. AI only helps if it works with real system data and knows its limits. Measure repurchase rate and follow-up contacts as well as the complaint rate, because the most expensive customers are the ones who leave without saying anything.

‍