If your respondents don’t actually match the audience, nothing downstream can save the study.

Most Gen Z research fails before a single question gets asked. A sample gets bought, a quota gets filled, and the respondents are technically the right age and functionally nothing like the people the client is trying to reach. The numbers come back clean. They are also wrong.

We keep a standing panel of 900+ U.S. high school and college students and handpick from it for every project, so the sample matches the client’s target audience instead of merely matching an age bracket. It is the least glamorous part of the work. It decides whether the other 8,000+ responses we have collected mean anything at all.

What bad recruiting actually costs you:

• Findings that describe a demographic instead of your customer

• Confident percentages built on people who would never buy the thing

• Regional splits that flatten the exact differences you needed to see

• A deck that survives the meeting and dies on contact with the market

None of that surfaces as an error. It surfaces as a strategy that quietly underperforms two quarters later, with no obvious cause anyone can point at.

How we recruit instead:

• Define the audience with the client before we touch the panel

• Select respondents by hand against that definition, every project

• Staff the team from 8 partner universities so one campus’s worldview isn’t the reading

• Publish the composition — age, gender, region — so clients can judge the sample themselves

We report who we talked to because a finding cannot be evaluated without it. On the Warner Bros. work our interview base skewed 18–23 at 45.8%, was 55.1% female and 58.5% students, and we said so in the report — the recommendation only holds for the people it came from.

See how we build a panel →