One person stalls at the setup screen and another gets through it and never comes back. They are two different problems, and a funnel chart shows them as one number.
The interview opens in the product at the moment someone stops, so you are asking while they are still looking at the thing that stopped them.
Which step of onboarding to fix first
Cancellation forms come back saying price. Price is the polite answer; the real one is usually a follow-up question away, and a form has no follow-up.
A study can start the moment someone cancels, within minutes, while the reason is still specific rather than rehearsed.
What to change, and what to stop worrying about
A release lands and the numbers barely move. Nobody can tell whether it was not wanted, not found, or not understood, and the fix is different for each.
You can ask the people who used it and the people who never did, separately, each at the point where it mattered.
What to build next, what to fix, and what to simply explain better
Someone upgrades and someone else looks at the same page and leaves. What separates them is rarely the number itself.
People say more about money to a machine than to a salesperson, in their own words rather than a rating out of ten.
The price, the packaging and what goes in which plan
You describe the product one way; people who have used it describe it another. The gap between the two is the copy you should be writing.
Interviews run in 18 languages, so the answer is not limited to the market that happens to speak yours.
The words on the site, in onboarding, and in the pitch
A campaign is signed off by people who already know what it means. The audience sees it once, out of context, and takes something else from it entirely.
Hundreds of people in the target audience, reacting to the real creative in their own words, in the days before it goes live rather than the week after.
Run it, change it, or drop it before the spend
A brand known for one product wants to be known for a category. The only people who can say whether that has happened are the ones who were not in the room.
Awkward subjects get straighter answers when the interviewer is not a person, which is most of what makes this kind of study worth running.
Whether to commit to the new direction or keep the old one
An idea everyone internally likes meets the one detail that decides it: a size, a texture, a price, a step nobody wants to take.
Your own customers if you have them, or a recruited audience if you do not, interviewed by voice or text in the language they think in.
Go, change it, or drop it before you build
A satisfaction score tells you something is wrong and nothing about what. The same score can hide a late delivery, an unanswered phone and a confusing invoice.
Interviews can go out after the event by link or email, with an incentive when you need the response rate, in the customer's own language.
Which part of the service to fix, and where
The annual engagement survey comes back positive and vague. The useful version is a conversation, which is exactly what nobody has time to run a hundred of.
Nobody is sitting across the table, which is the whole reason people answer this one honestly.
What to change internally, and what people actually want
A product that works at home lands differently somewhere else: a habit, a payment method, a word that means something slightly wrong.
Every interview runs in the participant's own language, so a market is not skipped because translating a study was too expensive.
Whether to enter, and what to lead with