Address directly

Customer loyalty hinges on effective problem resolution. Yet when multiple people share potential responsibility, accountability often dissolves—no one steps forward. The critical questions are: How is responsibility distributed within groups? When does shared responsibility result in inaction? And what does the evidence reveal about this phenomenon?

Studies

The Seizure Experiment

John Darley and Bibb Latané conducted the most famous bystander effect experiment in 1968. They invited 59 psychology students to what was presented as a discussion about university problems. Each sat alone in a booth with an intercom connection. Suddenly, everyone heard a participant stammering, calling for help, and simulating an epileptic seizure. When students believed they were the only listener, 85% helped within 60 seconds. However, when they believed four others were also listening, only 31% helped—and they took three times longer. Remarkably, everyone heard the identical emergency, yet the mere presence of others paralyzed action.

The Smoke Experiment

In 1968, Latané and Darley asked Columbia University students to complete a harmless questionnaire. After two minutes, white smoke began streaming through a wall vent—enough to fill the entire room with haze. When subjects sat alone, 75% reported the smoke within two minutes. In groups of three, only 38% reacted. The most striking finding: when two confederate actors were present who ignored the smoke, only 10% reported the incident—even as they coughed and could barely see. The passivity of others defined the situation as "normal," despite the looming danger.

The Three Mechanisms of Inaction

Fischer and colleagues analyzed over 105 studies involving more than 7,700 participants in 2011 and identified three critical processes: pluralistic ignorance—when nobody reacts, everyone assumes "it's probably nothing"; diffusion of responsibility—the more witnesses present, the less responsible each individual feels; and evaluation apprehension—the fear of embarrassing oneself. The alarming result: in dangerous emergencies, willingness to help decreased by an average of 38% when others were present. Nevertheless, most people believe they would be more likely to help in groups, not less.

Principle

Which principle for Customer Experience Design can be derived from this? People are more likely to help and act faster when they feel personally responsible and are directly addressed. In customer experience, this means vague appeals to "all customers" or impersonal mass communication often fall flat, while targeted, individual approaches achieve significantly higher response rates. This approach is particularly effective for feedback requests, community activities, and referral programs that require customers to perform specific actions. However, direct responsibility assignment only works optimally when the desired action is clearly defined and feasible for the customer. The following guidelines show how to implement this principle in practice.

Guidelines

Address community questions directly

In community forums, questions should be actively directed to specific experts or power users. Asking '@Max, you're familiar with X – can you help?' is more effective than posing a question to everyone. Unanswered questions foster pluralistic ignorance. The following examples illustrate this guideline:

• Stack Overflow: The bounty system addresses the bystander effect: An explicit incentive for a specific helper breaks through the diffusion of responsibility.
• Apple Support Communities: Apple marks 'Helpful' answers and honors 'Top Contributors' – this creates identifiable experts who feel accountable.

Escalations with personal owner

Critical incidents require a designated Incident Owner. Not 'the team is working on it,' but 'Lisa is coordinating the resolution.' The Owner must be empowered and visible – both internally and externally. The following examples illustrate this guideline:

• Amazon: The 'Single Threaded Owner' principle: Every initiative has one owner who is fully responsible. No committees, no shared responsibility.
• GitLab: For incidents, an 'Incident Manager' is immediately designated who communicates publicly: 'I am [Name], I am coordinating the resolution.' This creates trust and clarity.

Create relationship instead of anonymity

Reduce anonymity through personal connections: Use the recipient's name, reference known details about their situation, and assign specific contact persons rather than anonymous teams. Handwritten elements signal special effort and trigger reciprocity. Personal connections raise the psychological cost of dishonesty. The following examples illustrate this guideline:

• Kleinanzeigen (eBay Kleinanzeigen): Profiles with names, ratings, response time. The visible identity reduces fraud because dishonesty is no longer anonymous.
• Airbnb: Mutual reviews and profile photos. The personal connection between host and guest reduces damages and no-shows.