Voice-of-customer research for ad copy: a practical workflow

Voice-of-customer research for ad copy is the work of understanding how people describe a problem, the alternatives they considered, and the conditions that make a solution useful to them. Its output should be a brief a writer can use, with source context still attached. A folder of striking quotations is only an intermediate step.

The practical goal is to improve your decisions about audience moments, objections, and language. A small set of conversations can reveal questions worth exploring; it cannot by itself establish what an entire market believes.

Define the decision the research must inform

Choose one campaign question before collecting material. For example: "What causes a small design agency to replace its client approval process?" This question is more actionable than "What do agencies care about?"

Write down the audience boundary and the current uncertainty. You may know that prospects dislike scattered feedback but not whether they care more about finding approvals, explaining revisions, or bringing clients into another tool.

State what would change after the research. If the answer would not affect the angle, offer explanation, or proof you show, narrow the question until it can.

Gather sources with different perspectives

Useful starting points include customer interviews, support conversations, sales questions, product feedback, and public reviews you can legitimately access. Use only material you are authorized to work with, and remove unnecessary personal details from your working brief.

Each source has a context. A support ticket captures someone who needed help. A sales call reflects someone considering a purchase. A review may have been written long after onboarding. Do not flatten these into one undifferentiated customer voice.

Build a source register with a simple ID, source location, date, audience context, and relevant product version or plan when known. Record an excerpt location so another reviewer can find the original passage.

Ask about an event rather than a slogan

For interviews, focus on a recent experience. Questions such as "Do you value efficiency?" invite agreeable but vague answers. Ask instead:

1. What were you trying to finish when the problem became noticeable?
2. What did you do next, and what did you use?
3. What made the current approach difficult?
4. Which alternative did you seriously consider?
5. What almost stopped you from trying something different?
6. How would you describe the problem to a colleague?

Avoid inserting your preferred answer into the question. If you ask whether email chaos is frustrating, you have already supplied both the problem and the emotion. Let the person describe the event before testing your interpretation.

Separate quotation, interpretation, and implication

Create one evidence card for each useful observation. Give it three distinct parts: original language, your interpretation, and a possible campaign implication.

Here is a fictional example created for this guide:

Original language: "I knew someone had approved it. I just couldn't find which reply said so."

Interpretation: The person describes difficulty locating a decision, rather than difficulty producing the design.

Campaign implication: Explore an angle about making approval status visible beside the work.

The interpretation could be wrong or incomplete. Keeping it separate makes that uncertainty visible. Never present this invented example as customer research, and never turn a paraphrase of a real customer's words into a quotation.

Look for patterns and disagreements

Group cards by trigger, workaround, desired outcome, objection, and evaluation criterion. Keep contradictory observations in the same view. An agency that prefers email and one that wants a portal may need different messages or may represent different stages of readiness.

Count mentions if that helps you navigate the material, but label the counts as occurrences in the collected sources. Repeated comments from one person are not independent customers. A frequently discussed issue in your sample is not automatically the market's largest pain point.

Choose candidate angles by considering relevance to the audience, fit with the actual offer, and available evidence. A dramatic complaint that your product does not address is a poor foundation for an advertisement.

Turn the findings into a one-page writing brief

Summarize the selected audience moment, the current workaround, the desired change, the leading objection, and the product behavior that may help. Attach a few representative source cards and explain why you chose them.

Include an "unresolved" section. Perhaps the research suggests prospects worry about inviting clients into another system, but you have not established how onboarding works in your own product. The copywriter needs that gap as much as the attractive customer phrase.

The next step is to develop candidate angles, not to paste all the research into an ad. The review-to-angle workflow shows how to translate a particular observation into a message and proof requirement.

Keep the evidence available during revision

When a draft changes, revisit the source card. A cautious observation can become an absolute claim after several rounds of polishing. "I could find the approval" should not quietly become "Never lose an approval again."

MediaWriter supports research, campaign angles, and copy development within a promo. Keep the research brief and source material with that project so you can refer back to them while revising. Review the original sources yourself; an AI summary is a useful working aid, not independent evidence that the interpretation is correct.