Users Want Now, Not Later

TL;DR: Present bias makes current friction louder than future value. If users have to do homework before they get anything useful, most of them disappear. Give them the good part first and ask for commitment later.

You put the value at the end. Behind the setup screens, the welcome email, the tour of features nobody asked for. Users would get through it, see what the product could do, and stick around.

They didn’t.

And the thing is, this keeps happening. Teams run the post-mortem, tweak the copy, shorten the flow by one step, relaunch. Same result. Because the diagnosis is wrong. It was never a retention problem. It was a timing problem. You designed for a person who evaluates future benefits with the same calm logic they bring to decisions happening right now. That person is not your user. That person does not exist.

The gap between what someone plans to do and what they do is not a willpower problem. Researchers have been documenting it across contexts and populations for decades. The pattern holds in health behavior, in financial decisions, in software adoption. People intend to do things and then don’t. Product teams keep designing as if intentions are enough to carry someone through friction. They are not. They never were.

When later becomes never

In 1997, David Laibson published a paper on hyperbolic discounting that described something strange about how people value time. Not irrational people. All people. The idea is this: the closer a reward gets, the more you overweight it. The further away it is, the more you discount it. And the rate of that discount is not steady. It bends hard near the present. So a person can mean something in the future and then, when the future arrives, find a convincing reason not to do it. Not laziness. Math. The calculation changes as time passes.

Ted O’Donoghue and Matthew Rabin gave it a name in “Doing It Now or Later,” published in the American Economic Review in 1999. They called it present-biased preferences, and they showed it wasn’t a personality trait. As they wrote: “present-biased preferences give stronger relative weight to the earlier moment as it gets closer.” Every time. For everyone.

Which means the user who signed up with every intention of finishing setup is not lying when they say they’ll come back. At the moment they said it, they meant it. The problem is that moment and the next moment are not the same calculation. Present bias fills whatever gap sits between good intentions and actual use. A long onboarding flow is a series of gaps.

Kahneman and Tversky’s prospect theory adds something here too. Losses feel roughly twice as heavy as equivalent gains. In onboarding terms: the friction your user hits right now, the fields, the wait, the “we just need a few things from you,” lands heavier than the payoff waiting on the other side. The payoff isn’t real yet. The friction is. And that is where most users decide to leave.

The free door that everyone walked through

When Spotify launched in 2008 with a free tier, the music industry thought they’d lost their minds. If you can get the music for free, why would you ever pay?

Spotify was betting on something the labels weren’t thinking about. Not that people would feel generous out of nowhere. That the hardest moment in any conversion is getting someone to feel the product before they’ve committed to it. Once you have music running through your commute, your evenings, your work day, the question stops being whether the product is worth trying. It becomes whether the ads are worth tolerating. That’s a different question. A much smaller one. The moment of decision shifted from the product itself to one of its designed-in annoyances.

By 2014, CEO Daniel Ek said in public that 80% of paid subscribers had started on the free tier. By Q1 2023 Spotify had 210 million paid subscribers out of 515 million monthly active users, and in its annual report to the SEC that year the company described the free tier as “a funnel for our paid subscriptions.” Not a gift. A funnel.

Apple Music launched in 2015 with a three-month free trial that required a credit card upfront. Apple had every advantage: installed on every iPhone, one of the most valuable companies in history behind it, catalog on par with Spotify. Apple Music today sits at roughly 100 million subscribers. Spotify has more than twice that. The difference wasn’t the music or the algorithm or the brand. It was when each company asked users to decide.

Spotify designed around hyperbolic discounting without ever using the term. Value now. Commitment later. Future-you can deal with the subscription.

One question worth asking before you ship

When does your user actually feel value? And when do you ask them to act?

If effort comes before value, most won’t make it. Not because your tutorial is unclear or your copy is weak. Because the sequence is wrong. You are asking for a real, present-moment cost in exchange for a future payoff they haven’t touched yet. That bet loses more often than teams want to admit.

The fix is not removing friction from the system. It’s moving value in front of it. Let people feel something worth staying for before you ask for anything in return. Call it Value-Before-Ask if you need a name. In practice it means finding the first moment something useful actually happens for a new user, not the first completed step but the first moment the product did something for them, and then looking hard at what sits between them and that moment.

A credit card field before the trial. A five-step setup before the core screen. A feature walkthrough before anyone has used the product. These aren’t inconveniences. They’re bets on your user’s ability to carry future intentions through present obstacles.

Some users make it through anyway. Most don’t. That’s not a bug in your users. It’s a feature of every human brain that ever evolved under conditions where right now mattered more than next week.

Your product is competing with right now. It is always competing with right now.