Before a format travels from one country to the next, someone has already played it out. TikTok ran an anime drawing challenge in Japan; the challenge later spread to anime communities across Southeast Asia. The path looks like a platform decision made in a meeting room, but its starting point was a group of people who felt the itch earlier than the market — demand not yet formed, and they were already building their own fixes. In 2005, von Hippel turned this into a theory: Democratizing Innovation, published by MIT Press. The claim is direct. Lead users perceive demand pain points earlier than ordinary users, and the practical experience they hold is called sticky information. Sticky is the key word: what sticks is hard to move, and what is hard to move is valuable precisely because competitors cannot move it either.
Surveys Cannot Ask About Next Year
Some say insight comes from market research: scatter enough questionnaires and demand will surface on its own. Questionnaires measure averages, and averages contain no demand that has not yet taken shape. People answering a survey report what has already happened to them; what is happening right now at the frontier of demand is out of the questionnaire’s reach. Research finds today’s users. Lead users are next year’s. They stand at the very front of demand: the pain lands on them first, and the fix grows in their hands first. Take experience from them and what arrives is a live report. Look for demand inside a sample and what arrives is memory. The two sources differ by an entire time gap. Research counts what has happened well and portrays what has not poorly; the value of lead users sits entirely on the not-yet side. The question then becomes: how does the live report get collected.
Sticky Information Cannot Be Bought, Only Received Back
Others say lead users are niche extremists, and following them drags a product off course. The niche part is true. Sticky information is hard to transfer precisely because it lives inside real usage scenes — nothing on the market sells it, and no procurement process can order it. This knowledge was never written into documents or packed into manuals; it grows on the users’ own hands, tangled into their daily lives. The right move was never to copy extreme demands. It was to distill: turn local preferences into a replicable format. The anime challenge that started in Japan and extended to Southeast Asian anime communities followed exactly that path — distill the usage scene, do not transplant the extreme taste. Copying wrecks the product; distilling makes it travel. Take the wrong fork here and every insight collected earlier goes to waste: copy the extremes and the product narrows until only hobbyists remain; distill the scenes and what walks out of the narrow door is what most people will want tomorrow.
Data Sees Behavior, Not Motives
A third line says the platform has data, and algorithms understand demand better than users do. Data sees behavior; it does not see motive. The creation and participation signals of lead users carry both at once: they are not only using the product, they are modifying it, patching the missing piece with their own hands. A person who keeps rebuilding a tool is announcing two things — the tool has not fed them, and they know exactly where the gap is. Algorithms are good at amplifying signals that have already been validated. Validating them first is a human job. Outsource validation to the algorithm and the car is being driven by the rearview mirror. By the time the data reveals a trend, the trend is no longer new. An algorithm is an amplifier, not a detector — and detection always belongs to the person in front.
The Last Mile of R&D Sits on the User’s Side
Democratizing innovation does not mean handing R&D down to the crowd. It means admitting something else: the intelligence from the frontier of demand already sits in users’ hands. The last mile of R&D is receiving sticky information back from lead users — receive it, distill it, replicate it. That is the whole path by which a weather forecast of demand becomes a product. This approach has a boundary. In mature, stable categories where demand barely moves for a decade, users have nothing to teach, and the shortcut does not exist. Everywhere else, one question remains: the people who first played the format into being — where are they working on it right now?
