Pre-flight research: how to screen 20 ideas synthetically before you pay for a panel
Aviation doesn't skip the pre-flight checklist, and neither should product decisions. Here's how to screen many ideas synthetically, then spend real research budget only on the few that survive.
By VibeIQ Research · Product research team
Key takeaways
- Pre-flight research means screening many ideas synthetically, then validating only the top few with real people.
- It flips the economics: you explore broadly for the cost of time, and spend panel budget only where it counts.
- The workflow is a funnel, diverge to ~20, screen synthetically, cut to 3, validate with humans, decide.
- It plays to synthetic's strength (comparative, directional signal) and covers its weakness (absolute numbers) with a human check.
No pilot skips the pre-flight checklist. It's fast, it's cheap, and it catches the problems that are ruinous later. Product and marketing decisions deserve the same discipline, but traditional research is too slow and expensive to run on every early idea, so most ideas get filtered by opinion in a meeting instead. Pre-flight research fixes that: screen broadly and synthetically first, then spend real budget only on what survives.
What pre-flight research is
Pre-flight research is a two-stage funnel. Stage one runs your candidate ideas, concepts, messages, price points, features, against a synthetic audience to get fast, comparative signal. Stage two takes only the handful that rise to the top into real human validation. You're not replacing panels; you're making sure they only ever see your best three ideas instead of your first twenty.
20 → 3 → 1
diverge broadly, screen, then validate
Hours
to first directional signal, not weeks
80-95%
synthetic-vs-real correlation on directional questions
Top 3
ideas that ever reach a human panel
The pre-flight funnel, step by step
1. Diverge, generate ~20 candidates
Resist narrowing too early. List every serious variant: 20 concepts, 12 price points, 8 value-prop angles. The whole advantage of pre-flight is that breadth is nearly free, so the cost of adding one more idea to test is minutes, not a new recruit.
2. Screen synthetically
Run all candidates against a synthetic audience that mirrors your real ICP, asking comparative questions, appeal, clarity, preference, willingness to pay. You're after the ranking and the reasons behind it, not a precise percentage.
3. Rank and cut
Read the results by segment, not just the average, an idea can win overall while losing your beachhead segment. Cut ruthlessly to the top three, and keep the synthetic reasoning (the "why") as hypotheses to probe with humans.
4. Validate the finalists with humans
Now spend the real budget. Put the surviving three in front of a real panel or in real interviews, focused on the specific risks synthetic can't settle: absolute demand, pricing thresholds, sensitive reactions. This is where you buy certainty.
5. Decide
You reach the decision having explored ten times more ground than a purely human process could afford, with a documented trail from twenty ideas down to one, and a human check on the number that actually matters.
Why the funnel works
It matches the method to the job. Synthetic research is strongest at comparative, directional questions and weakest at absolute point estimates, so you use it to rank and cut, and use humans to confirm the final magnitude. Neither method is asked to do what it's bad at. (More on that split in the evidence on synthetic accuracy.)
A worked example: 12 price points in an afternoon
A team debating a pricing change lists 12 plausible price-and-packaging combinations. Screening them synthetically against their ICP surfaces a clear top tier and, crucially, a segment split: smaller customers welcome usage-based pricing while enterprise buyers want a committed floor. They carry the three strongest options into real buyer conversations to nail the exact thresholds, a week of work instead of a quarter, and no option shipped on a hunch.
What to screen synthetically, and what not to
Good pre-flight candidates
- Concept, message, creative and positioning comparisons where you need a ranking.
- Early price and packaging exploration to narrow the range.
- Feature prioritisation and value-prop angles across segments.
Keep these for humans
- Final, absolute demand or pricing thresholds a P&L depends on.
- Sensitive, identity-charged or highly regulated topics.
- Anything where a small, unrepresented sub-population is the decision.
Wire it into your cadence and pre-flight stops being a one-off, it becomes the default first step before any idea earns real research budget. For a continuous version of this, see building a continuous discovery loop with synthetic users.
Frequently asked questions
What is pre-flight research?
Pre-flight research is a workflow where you screen many ideas, concepts, messages, prices, features, against a synthetic audience to rank them quickly, then validate only the top few with real humans before deciding.
Does pre-flight research replace human panels?
No. It filters ideas so your human panel only ever sees your strongest few, and it hands the final, high-stakes numbers to humans. It makes real research more focused, not obsolete.
How many ideas should I screen?
As many serious variants as you have, the marginal cost of one more is minutes. Teams commonly screen 10-20 concepts or price points, then cut to about three finalists for human validation.