Data quality claims start sounding identical after so many vendor calls. The seven questions in this guide change that.

Each one is written in plain procurement language, designed to be asked verbatim, and built so a vague answer becomes obvious the moment you hear it. You'll know what a strong answer sounds like before the call starts.

Every question follows the same five-part structure. Read the parts in order; each one is a separate check on the answer you get. Every answer in this guide traces back to a number we measured and published ourselves in our 2026 Data Quality Benchmark Report. 

We put our own fieldwork through the IA's GDQ framework and published the results in the open, with the formula, denominator, and time period behind each figure. This guide gives you the questions. The above report is where we show our work, in full.

This guide and our 2026 Data Quality Benchmark Report go together. The guide is the questions worth asking; the report is where we answer every one of them ourselves, with the methodology disclosed. Read either, and the other is the natural next step.

Who is this guide for?

• Insights leaders running a vendor search, RFP, or annual supplier review

• Researchers who need a defensible answer for how sample was sourced and cleaned

• Procurement and vendor management teams who need the methodology and security questions covered

• Governance and IT stakeholders reviewing AI in a research pipeline against ICC/ESOMAR 7(e)