Your users will tell you why.They just won't book a call.

askedwhy interviews them instead — a voice conversation on their phone, six minutes, in their own words. Then it tells you what they all said, not what one of them said.

Start freeSee it interview someone

10 interviews free · no credit card required · your first study takes ten minutes to set up

askedwhy heard somethingILLUSTRATIVE

signed up in the last 60 days · never activated

31 of 43 said the same thing

They never found where to bring their own data in.

“I got to the empty dashboard and I just sat there. I assumed there was an import somewhere but I wasn't going to go hunting for it.”

DO THIS Put import on the empty state, not in settings.
heard from 43 interviews

Illustrative. No customer's interviews appear anywhere here, and none ever will without their say-so.

The bottleneck was never the analysis.It was getting anyone to talk to you.

The problem

You're shipping to people you've never heard from. And it's not because you didn't try.

You sent the emails. A few replied, fewer booked, one showed up. Three weeks gone and one person's opinion to show for it — so the next decision gets made on instinct, and you find out six weeks later whether it was right.

  • Write the guidean afternoon
  • Chase the invitesa week of emails
  • Schedule around themanother week
  • Run each call45 min
  • Write it upnever happens
  • What you actually learnedn = 2
Watch it work

It doesn't take “it was fine” for an answer.

This is where a survey dies. Someone gives you a vague answer and there's nobody in the room to push. askedwhy pushes — it names the vague thing, asks what actually happened, and keeps going until the answer is specific enough to act on.

Activation study · interview 27 of 43ILLUSTRATIVE
  1. askedwhyYou've been on the plan about two months. What made you sign up in the first place?
  2. themUh, I think a friend mentioned it? It looked useful.VAGUE — DON'T ACCEPT IT
  3. askedwhyWhat did they say it would do for you — was it the reporting, or something else?NAME IT
  4. themThe reporting, yeah. I wanted the weekly thing for my team.
  5. askedwhyDid you end up sending one? Walk me through the last time you tried.USAGE CHECK · NOT WHAT THEY MEANT TO DO
  6. themHonestly no. I built it once and it pulled the wrong date range, and I couldn't work out where to change that, so I just went back to the spreadsheet.
6m 12s · they finished at 22:47 on a Tuesday

Every probe label above is a technique in the interview script, not a name invented for this page — the same ones the code runs.

  • It asks one thing at a time. The reason people answer properly: no three-part question, no leading, no list to work through.
  • It knows when to stop. Guide covered, it thanks them and ends. Nobody is held hostage for 40 minutes, which is most of why they finish.
  • It never puts words in their mouth. It asks what happened, not whether they liked it. A leading question produces agreement, and agreement isn't data.
What you get

Not 43 transcripts. One thing 31 of them said.

A count, their exact sentence, and what to change.

Every frame here is written for the page. We have not run these interviews yet.

ILLUSTRATIVE
31 of 43

They never found where to import their data

DO THIS Put import on the empty state.

“I assumed there was an import somewhere but I wasn't going to go hunting for it.”

activation study · 43 interviews, this month
ILLUSTRATIVE
9 of 12 who said “too expensive”

It isn't the price — it's who has to approve it

“It's not that it's a lot. I just can't put it on my own card and I don't want to ask.”

DO THIS Add a plan one person can expense.
ILLUSTRATIVE
14 of 61

Export, weekly email, Slack — all one request

“I end up screenshotting it into Slack every Monday, which feels stupid.”

DO THIS Scheduled send, not another export button.
ILLUSTRATIVE
7 of 19 who churned

They didn't leave for a competitor

“The launch was over. I forgot we were still paying for it, honestly.”

DO THIS Offer a pause, not a save-offer discount.
How it works

One link out. One answer back.

Set it up once. However many people click, they all get a real conversation — and you get one thing they said, not a folder of recordings. Then it asks you what you still don't know.

  1. 01

    Say what you need to know

    Describe it in a sentence and askedwhy drafts the guide. Edit any question you want.

  2. 02

    Share one link

    Email it, put it in your app, drop it in a support reply. Nothing to install.

  3. 03

    They all talk at once

    Six minutes each, on their phone, whenever suits them. Forty at a time if forty click.

  4. 04

    It comes back with one answer

    The theme, the count, their exact words — then the next question you should ask.

Instead of

Depth or scale. You've always had to pick one.

A survey

Scale, no depth

You get a 3 out of 5 with nothing attached. Nobody fills in the free-text box, and a radio button can't be asked why.

People reached
hundreds
Follow-up questions
none
Time to answers
days

Good for measuring something you already understand.

An agency

Depth, no scale

Real interviews by a real researcher, and a deck at the end. Then a follow-up question in month two is a new engagement.

People reached
8–12
Cost
$15–30k
Time to answers
6 weeks

Worth it for the thing we don't do: a senior human deciding what to study.

askedwhy

Both, or it isn't worth building

Real conversations with follow-up questions, at the scale of a survey, because the interviewer costs nothing to run a fortieth time.

People reached
everyone who clicks
Follow-up questions
every turn
Time to answers
this week

Not for you if you have no users yet. There's nobody to interview.

Survey depth at survey scale is the thing that hasn't existed. Not because nobody wanted it — because until very recently a machine couldn't hold a conversation well enough that people would finish one.

Your users, not ours

These are your customers. We treat them like it.

They're doing you a favour, and nothing about this works if they feel tricked. So four things are true before a microphone ever opens.

  • They're told it's an AIUp front, in the first sentence. Hiding it is the only version of this that fails — and it makes people blunter, not shyer.
  • They're told what's recordedAudio and a transcript, who sees it, and what it's for. Before they agree, not in a footer afterwards.
  • They can stop, any timeOne tap, no guilt, no progress bar shaming them into finishing. Most finish anyway, which is the point.
  • Your interviews stay yoursIsolated to your account, never shown to another customer, and never used to train a model for anyone else.
What it costs

Priced against one interview, not against a research budget.

The honest comparison isn't software — it's what a single user conversation actually costs you today, once you count the recruiting, the scheduling and the hour of writing it up that never happens.

One interview, by hand~2 hrs of yours

Recruit, schedule, reschedule, run it, write it up. Times however many you need before two people agreeing means anything.

An agency study$15–30k · 6 weeks

Eight to twelve participants, a senior researcher, and a deck. Excellent, and not a thing you can do four times a year.

askedwhy10 free · then per interview

Ten interviews on the free plan, no credit card required. After that you pay for conversations that happened — not for seats.

Run your first study on the free plan and see whether the answers are worth anything before we ever talk about a number.

Start

Find out what your users actually think.

Write one guide, share one link, and read what comes back. Ten interviews on us — enough for a theme to show up more than once.

Start freeI already have an account

No credit card required · one link to share · your interviews stay yours

Want a hand pointing your first study at the right question? Ask us — we read every email.

Field notes

Why customer discovery quietly stops happening

Every founder knows they should. Almost nobody does it twice.

User interviews are the most recommended and least practised habit in software. The reason isn't discipline — it's that the cost is front-loaded and invisible. Writing a discussion guide takes an afternoon. Recruiting takes a week of emails for a handful of replies. Scheduling takes another, and half the people who agreed won't appear. By the time you've run two conversations you've spent three weeks and learned two people's opinions, which is not enough to change a roadmap and is more than enough to make you never do it again.

So teams fall back on what's cheap: analytics, which tell you what happened and never why, and surveys, which get you a number nobody can act on. Both are measurement. Neither is discovery. The questions that actually decide what to build — why they signed up, what they expected, where they gave up, why they cancelled — only have answers if someone asks a person and then asks a second question.

The second question is the whole thing

Nobody's first answer is useful. “It was fine”, “too expensive”, “I got busy” — these are politeness, not information. What matters is what comes out when someone says fine is doing some work there, what made you pause? That's the difference between a churn reason of “price” and finding out that nobody objected to the number, they just couldn't expense it without asking their manager. One of those is a discount. The other is a new plan.

It's also the difference between an anecdote and evidence. One person saying onboarding is confusing is a story; thirty-one of forty-three saying they couldn't find the import is a decision, and the only way to know which one you have is to ask enough people that a pattern can show up. That's what askedwhy automates — not the analysis, the asking. Continuously, in the background, and it reports back when there's something worth your attention.