Things that scale give you numbers. Things that do not scale give you understanding.
A survey of a thousand users tells you what happened. Sitting with one user for an hour tells you why.
The why is where the product comes from and there is no shortcut to it.
The Form Gives You The Wrong Answer
The moment you turn a conversation into a form, you get the answers people think you want.
There is a well-known example of this from Intuit. They used to ask people whether they would recommend the product, and the scores were good.
But when they watched people actually use it, they found something the survey never caught. Users were completing the task and feeling terrible while doing it.
So the number said one thing and the room said another. And you only get the room by being in it.
You have to hear the person hesitate. Watch them get stuck. Listen to the thing they say right before they change the subject.
None of that arrives through a text field.
Airbnb Went Door-to-Door
In 2009, Airbnb was making about two hundred dollars a week and had been at it for a year.
Paul Graham told them to go to New York and meet their users in person. So the founders flew there most weekends, knocked on doors, and stayed with hosts.
They found the listings had bad photographs. Hence they rented a camera and photographed the apartments themselves, one at a time.
Now notice what that fix required. It could not be found in the data, because the data showed listings that were not booking and gave no reason. And it could not be delegated, because the founders had to be in the apartments to see it.
The Inefficiency Is A Filter
Here is the part most people miss.
This work is slow, unmeasurable, and looks like it is not working for weeks. Therefore most founders will not do it.
They will run a survey instead, because a survey produces a chart and a chart can be shown in a meeting.
So the ones who do the slow version end up knowing something their competitors do not. And that knowledge is not available any other way, at any price.
The thing that makes it a bad use of your time is the same thing that makes it valuable. Nobody else is willing to spend the time either.
Do It While You Still Can
There is a window on this, and it closes.
You do it early, when you have few enough customers that talking to all of them is possible. Later you will have too many, and by then you should already know what you learned.
The mistake is thinking you will get to it once things settle down. Things do not settle down. You go from forty customers to four hundred, and the conversations become a research programme with a budget instead of a Tuesday afternoon.
So the founders who skipped it early are the ones running surveys later. They are asking a thousand people a question they could have answered by sitting with ten.
What This Actually Costs You
Look at your last product decision. Then ask where the input came from.
If it came from a dashboard, you know what people did. You do not know why, and you have been guessing at the why this whole time without saying so.
Go and book five calls this week. No script, no form, and no product manager sitting in for you.
You will learn something in the first hour that your analytics could never have told you. And the reason nobody else has it is that they thought an hour was too expensive.