The companies that changed consumer behavior built new habits and coined new terms of how people are interacting with their phone: "Shazam it." "Swipe right." "Send a snap." Each one started as a physical action on a phone and became a verb that entered the culture. So now the gesture, the brand and what happens when doing it collapsed into one word.

"Screenshot" already has the first two. It's a word everyone knows, an action everyone does, a habit that runs billions of times a day across every phone in the world. It's one of the most powerful actions since it sits above every app, at the hardware level, completely platform-agnostic.

It's just missing its destination.

You screenshot something. It makes a sound. Maybe you share it with someone immediately. But latest after that it disappears into your camera roll. The habit ends in a void. No destination or action that follows after.

That's a big white space. You have a well established gesture but no real destination and brand that owns it.

I've spent weeks studying what people are screenshotting. What I found: your screenshots knew what you were going to buy or do on average 39 days before you did. The shape and process of a decision is fully visible in the data, weeks before the person consciously made it.

This is the largest pre-decision intent dataset that exists.

What the gesture actually contains

Most people think of screenshots as this messy thing that is polluting your camera roll. A bunch of unaesthetic stuff that has no real use. But the data tells a different story.

Working with Pool, we studied millions of screenshots across thousands of users. We saw one thing clearly: screenshots are a picture of active, private, pre-conscious desires.

In 1938, economist Paul Samuelson introduced the concept of revealed preference: what people actually choose reveals their true preferences more reliably than anything they say. Stated preferences are contaminated by social desirability, by self-image, by the desire to appear rational. Actual choices are honest because they have consequences.

The screenshot is revealed preference at its most raw.

Bursztyn et al. at University of Chicago showed what we already intuitively know: remove the social audience entirely and you get honest behavior. The ballot box effect, applied to a camera roll.

The beauty of screenshots is that you can capture the whole process of a person doing something before rationalization begins and then seeing how the "do I really need this" kicks in. Kahneman's System 1, the fast, intuitive, emotional mind, generates the desire. The screenshot captures it before System 2 kicks in with "do I really need this" and "I'll think about it later." It is the raw want, before the editing process begins.

Fishbein and Ajzen's research found that stated purchase intent predicts actual behavior only 34% of the time. The screenshot sidesteps the problem entirely. Nobody screenshots things to signal intent to others. It's private, frictionless, involuntary almost. The purest implicit signal available at scale.

Why social media isn't the same thing

The obvious objection: don't Instagram and TikTok already know all of this? They track everything I am doing on their platforms. If anyone understands pre-decision behavior, it's them.

True. But the signals aren't the same thing.

Social platforms track mostly passive pre-intent. The algorithm watches what your thumb does and infers what you might want. You didn't tell it anything. Your behavioral residue was harvested in the background, and optimized to keep you in the feed longer and sell you more ads. They don't care about what you actually want, they care about optimizing time spent on device.

The screenshot is an explicit, active decision. Two buttons, pressed on purpose. With an intent or action behind it.

On top of that, social data is siloed. Meta knows your Meta behavior. TikTok knows TikTok. Neither sees that you screenshotted the product on Instagram, found a review of it on a blog three weeks later, then screenshotted a competitor, then came back to the original 47 days after that. The full arc of the decision and how it transforms over time is invisible to any single platform.

The camera roll sees all of it. The screenshot goes to the same place regardless of where it was taken.

And then the most important difference: social platforms optimize your signal against you. Your intent data is used to capture more of your attention, serve more ads, deepen the loop. The data exists to serve the platform.

The screenshot belongs to the user. It was made for them, not about them. That's what makes it a different kind of dataset and a different kind of opportunity.

The gap: 22 to 78 days

We measured the distance between desire and action.

We mapped the gap between when a user first screenshots content in a category and when a confirmable post-decision signal (a receipt, a booking confirmation, a purchase record) appears in the same history.

Instagram saves are the longest-burning intent in the dataset. Someone screenshots something from Instagram and on average acts on it 78 days later.

Travel converts fastest. 22 days, because travel has more often a concrete desire and logistics attached to it.

Shopping sits in the middle, at 39 days. But that average hides enormous variance. The fastest 25% act within 9 days: they see it, want it, buy it. Only 22 screenshots accumulate in the gap. The slowest 25% take over five months, generating 241 screenshots during the gap. Months of circling the same decision, before pulling the trigger.

The most indecisive buyer produces the richest dataset. The harder the decision, the more legible the intent.

The decision has a heartbeat

When I mapped shopping burst sessions (clusters of 5+ screenshots within 10 minutes) against proximity to a purchase event, a pattern appeared:

Burst intensity peaks 2 to 4 weeks before the purchase. By the final week the decision is essentially made.

The peak happens earlier, when the person is still in active comparison mode. Pulling in more product shots, checking alternatives, building conviction. That's also when they're still persuadable.

The escalation is visible in the data 2-3 weeks before the person consciously knows they've made up their mind. The camera roll isn't recording a decision in progress. It's running ahead of it.

The camera roll is upstream. The inbox is downstream.

Every major AI assistant being built today is trained on the inbox.

Email is clean, structured, attributable. It feels like intent data because it arrives at the moment of a decision. The booking confirmation, the shipping notification, the receipt.

But email is a post-decision signal. By the time a confirmation lands, the decision has been made. The email is the receipt of a process that was interesting much earlier and is now over.

In Pool's dataset, post-decision content — receipts, invoices, booking confirmations, plane tickets — represents 2.9% of all screenshots. Pre-decision content — shopping, restaurants, travel inspiration, design, social discovery — represents 41%. The ratio is 14:1.

And that's the reality of an actual consumer experience. For every purchase confirmation, there are fourteen moments of pre-decision activity. For every signal that arrives after, there were weeks of private intent that arrived before. So while the confirmation lands in the mail, the camera roll already knows what's gonna happen in the next 4 weeks.

In the screenshot dataset, preferences are still being constructed. Ariely's research on coherent arbitrariness showed that consumer preferences aren't fixed — they're built in the moment, shaped by framing, timing, available information. The person who has been circling a purchase for 39 days hasn't made up their mind yet. They're still building it. That's the window that's actionable and where it's possible to intervene.

By the time the email arrives, the window is closed.

Intent that doesn't die

Not all pre-decision intent resolves quickly. Some of it persists for months. For years.

In Pool's data, there are 11,014 items that the same user screenshotted across multiple separate days — returned to across visits separated by days, weeks, sometimes years.

An Amazon cart screenshotted four times over months. An IKEA cabinet captured twice, 1,421 days apart. These are unresolved decisions, things the person can't act on, can't let go of, or just can't make a final decision on.

Behavioral science calls this the Zeigarnik effect — incomplete tasks occupy more mental space than completed ones. Every unresolved save is an open loop. Over time, the accumulation creates a low-grade pressure of having hundreds of open loops you are not following through.

These 11,014 returning items are the highest-confidence signals in the dataset. Things someone has demonstrably wanted, repeatedly, for a long time, without resolving.

Collectors collect everywhere

The clearest predictor of how much intent someone generates is how many platforms they screenshot from.

Users who save from a single platform have a handful of screenshots on average. Users who save from 10 different platforms average nearly 800. Users who save from 15 distinct platform types average over 5,000.

This is collector psychology: people who curate in one domain curate across all domains. A Pinterest user in Pool's data generates 4.4x more burst sessions across every category.

These are people with an active relationship to wanting things: they know what they like, they're looking for more of it, across every app on their phone.

The last unowned gesture

Every gesture that got claimed — the search bar, the tap, the swipe, the double tap — followed the same pattern. The behavior existed first, as a raw reflex with no brand attached. Then someone built a reward loop around it. The reflex became a ritual. The ritual became a platform.

The screenshot is at that inflection point now.

The behavior exists at massive scale. Burst sessions have increased 9x in two years driven by behavioral acceleration from other platforms. TikTok restructured how a generation consumes content: see it, want it, capture it, move on. The screenshot is the natural exhaust of that culture.

What's missing is the reward that closes the loop.

Right now, the screenshot is a one-way gesture. You press two buttons. Something is saved. Nothing comes back. There's no variable reward, no resolution, no follow-up action that makes something out of that save. The behavior exists at scale. What's missing is the thing that transforms this behavior into something smart.

BJ Fogg's behavioral research is clear: a habit without a reward doesn't compound. It plateaus. The screenshot, as it exists today, is the most common intentional action on a phone that returns the least value to the person who made it.

The gap and the opportunity are the same thing.

When screenshots create something actionable, and become a place where desire gets resolved rather than stored, the attached meaning changes. It's no longer a save. It's the beginning of something.

That habit loop, once closed, is extraordinarily hard to break. Because it's not attached to an app. It's attached to a button that already exists on every phone, pressed billions of times a day, by people who have no idea what they're generating.

The screenshot button is the last major gesture in consumer tech that nobody owns.

The dataset inside it is the largest pre-decision intent signal that exists. More honest than search, earlier than email, more explicit than anything any platform has ever pulled passively.

The behavior is already there. The data is already there. And Pool builds the reward loop that makes the gesture mean something.