Voice of the customer (VoC) is a program for capturing what customers need and how they experience a company, then turning it into decisions. It combines feedback customers give when asked, such as surveys, with feedback they give unprompted in calls, chats, emails, complaints, and reviews.
The term comes from product development. In a 1993 Marketing Science paper, Abbie Griffin and John Hauser described Quality Function Deployment. In that process, the "voice of the customer" is carried through R&D, engineering, and manufacturing. A later encyclopedia entry by both authors and colleagues notes that everyday use is broader and can cover customer feedback in any form.
Solicited and unsolicited feedback
Solicited feedback answers questions the company asks. Customer satisfaction score asks how satisfied the customer was with an interaction. Customer effort score asks how easy it was to get something done.
Net Promoter Score (NPS) asks how likely the customer is to recommend the company to a friend or colleague. Answers use a 0 to 10 scale. Bain & Company subtracts the share of detractors (0 to 6) from the share of promoters (9 to 10).
Surveys give scores that compare over time. They cover only the customers who answer, and a score usually shows how customers feel more clearly than why.
Unsolicited feedback is what customers say without being asked: calls, chats, emails, complaints, app store reviews, and social posts. It comes in the customer's own words and in larger volumes. It is also unstructured, so it needs customer feedback analysis to turn it into themes.
Contact center conversations as VoC data
Every call and chat records why a customer got in touch, in their own words. For many companies, it is one of the largest sources of unsolicited feedback they hold.
Contact reasons. The share of contacts per reason, and how it moves week to week, shows where products, bills, or policies confuse people.
Effort signals. Phrases such as "this is my third call" or "nobody called me back" point at repeat contacts and broken promises.
Sentiment and complaints. Sentiment analysis and complaint tagging help rank themes by how much they hurt.
Conversations with AI agents. Requests an AI agent could not complete, and the reasons customers asked for a person, are feedback too.
Calls reach VoC through transcription and speech analytics, so recognition errors carry through. It also helps to remember who calls. Customers with problems tend to be over-represented, and satisfied customers phone less often.
Closing the loop
With the customer. A low score or a complaint triggers a follow-up, often a callback within a set time. The customer hears that someone listened, and the company learns the detail behind the score.
In the business. Recurring themes go to the team that can fix the cause: product, billing, policy, or the website. Each theme needs an owner, a fix date, and a measure, such as contact volume on that reason before and after.
Many programs hold a monthly review where CX, operations, and product go through the top themes and their owners.
Governance and GDPR for recordings
Call recordings and transcripts that identify a customer are personal data. Under the GDPR, data collected for specified purposes may not be further processed in a way incompatible with them (Article 5(1)(b)).
Recordings made for quality or compliance and later used for VoC analysis raise that question. Article 6(4) lists what a controller weighs. The factors are the link between purposes, the context of collection, the nature of the data, possible consequences, and safeguards. Encryption and pseudonymization are its examples of safeguards.
In practice, it helps to name analysis in the privacy notice and to keep recordings no longer than needed (Article 5(1)(e)). Working on redacted transcripts (see PII redaction) and reporting themes in aggregate keep exposure low. Emotion detection has its own rules under the EU AI Act, covered under speech analytics.
Frequently asked questions
What is the difference between voice of the customer and customer feedback analysis?
VoC is the program: what feedback is collected, who reviews it, and how decisions follow. Customer feedback analysis is the method that turns raw feedback into themes and trends. A VoC program uses it as one of its steps.
What are examples of voice of the customer data?
Common examples are survey answers, such as CSAT, CES, and NPS, with their comments. Call recordings and transcripts, chat and email threads, complaints, online reviews, social posts, and customer interviews count as well.
How do you measure a voice of the customer program?
Survey scores show how customers feel over time. The effect of fixes shows whether the program works: fewer contacts on a fixed reason, fewer repeat contacts, and better first contact resolution. Counting how many themes reached an owner and a fix shows whether the loop closes.
Can call recordings be used for voice of the customer under the GDPR?
Often, with care. The use has to fit the purposes customers were told about, or pass the compatibility test in Article 6(4). Redacted transcripts, aggregate reporting, and clear retention limits reduce the risk. Your data protection officer can confirm the legal basis for your setup.
How we approach this at Zowie
Zowie's AI agents work on voice, chat, and email, with over 100 million conversations a year across our clients. The language model speaks. A separate, deterministic Decision Engine returns every amount, rate, offer, disclosure, and escalation from your rules, with an audit log. To talk about your conversation data, you are welcome to book a demo.
Explore: Customer feedback analysis, Speech analytics, Customer satisfaction score (CSAT), Customer effort score (CES), Sentiment analysis, GDPR compliance