Choosing an approach
Sample frame vs panel vs expert network: which one actually covers your population?
If you’re researching a hard-to-reach professional population, the question isn’t “which panel is biggest?” — it’s “what proportion of the real population can I reach, and can I prove it?” A built sample frame answers that with a number, a denominator, and an evidence URL per record.
The three approaches, plainly
Consumer / online panel (Cint, Dynata, Prolific)
A standing pool of opted-in respondents profiled against attributes. Good for: fast, cost-effective quantitative work on populations that exist in the panel at workable incidence. Struggles with: low-incidence professional populations — the panel either can’t fill the quota, or fills it with people who claim to qualify under pressure. You get completes; you don’t get a defined, screened population, and you can’t state what share of it you reached.
Expert network (GLG, AlphaSights, Guidepoint)
A roster of vetted senior professionals available for paid, hour-long consultations. Good for: depth — a handful of deep conversations. Struggles with: breadth, representativeness, and unit economics at scale. The roster reflects who signed up to consult, not the structure of the population — so it is not a sampling instrument.
A built sample frame (SampleQuick)
A constructed, screened list of the actual members of a defined population, built for your study. Each record carries an evidence URL, the frame is measured for coverage against an authoritative registry, and it ships with a GDPR compliance pack. Good for: reaching defined, hard-to-reach, low-incidence professional populations and proving how much of them you reached. It is not lead generation — it’s a research recruitment instrument.
Side by side
| Dimension | Panel | Expert network | Built frame |
|---|---|---|---|
| What you get | Completes vs quotas | Paid expert hours | A screened, named frame |
| Coverage of a population | Unknown | N/A | Measured vs registry denominator |
| Representativeness | Self-selection; weak for niche | Who signs up to consult | Built to mirror the registry; gaps reported |
| Provenance / audit | Usually none | Per-engagement | Evidence URL per record |
| Compliance docs | Panel-level | Engagement-level | Per-frame GDPR pack |
| Cost model | Per complete | Per hour | Per frame / project |
| Best for | High-incidence quant | A few deep interviews | Hard-to-reach, defined populations |
When a panel is the right call
We’ll tell you when you don’t need us. Use a panel when your population exists in it at workable incidence, you need large N quickly and cheaply, you don’t need record-level provenance, or the work is repeated tracking on a broad audience. If that’s your study, a good panel will serve you better and cost you less.
When you need a sample frame instead
- Low incidence — panels return thin, unreliable, or unaffordable samples.
- Defined population — a profession, a register, an accreditation; “close enough” won’t do.
- Coverage or census needs — you must state what share you reached, with a denominator anyone can check.
- Accreditation / registry mapping — the frame must map cleanly onto a body or register.
- Defensible provenance — every record must answer “where did this come from?”
If two or more describe your project, a panel will quietly undermine it and an expert network won’t scale to it. That’s the gap a built frame fills.