Progress bars reward you at 100% — motivation quits at 72%

There's a strange thing that happens on a website when a progress bar appears. People lean in. They keep going. And then, somewhere past the halfway mark, a lot of them quietly leave — even though they were clearly interested enough to start. The number that keeps surfacing in usability testing and behavioural research is somewhere around 72%, which is a suspiciously specific place for motivation to die.

The goal gradient has a cliff edge

Back in the 1930s, a psychologist named Clark Hull noticed that rats running a maze sped up as they got closer to the food at the end. The closer the reward, the faster the behaviour. Later researchers — most famously in a 2006 paper by Kivetz, Urminsky and Zheng — showed that humans do the same thing with loyalty cards. Give someone a coffee card with ten stamps and they'll buy coffee faster as they approach the tenth stamp. Give them a twelve-stamp card with two stamps already filled in, and they behave almost identically to the ten-stamp group. It's not the absolute number that matters. It's the perceived distance to the finish line.

This is the goal gradient effect, and it's one of the more reliable findings in behavioural psychology. It should mean that a progress bar is a perfect motivational tool: the closer you get to 100%, the harder you push.

Except it doesn't always work that way.

What Hull and Kivetz were measuring was effort within a task that people had already committed to. What most websites are dealing with is something different: a person who is deciding, moment by moment, whether to keep going at all. And that's where the 72% figure starts to make sense.

Why the last stretch is the hardest stretch

There's a well-documented phenomenon in decision-making called the peak-end rule, described by Daniel Kahneman and colleagues. People don't remember an experience by averaging it — they remember the most intense moment and how it ended. A long, pleasant experience with a frustrating ending gets remembered as frustrating.

Now think about what a progress bar actually promises. It says: stay with me, and at 100% you'll get something. That's a reward loop, and it's a variable one if the payoff is vague. But here's the problem — most progress bars don't deliver a clean, satisfying reward at 100%. They deliver a form, a confirmation email, a "we'll be in touch", or worse, a page that loads slowly afterwards.

So the user has been running on the promise of a reward, and somewhere around three-quarters of the way through, they start to do a cost-benefit calculation. Kahneman's work on loss aversion is relevant here too: people weight losses roughly twice as heavily as equivalent gains. The remaining 28% of the bar isn't just "a bit more effort" — it's the moment where the user starts asking whether the reward at the end is actually worth what they've already spent getting there.

The sunk cost fallacy should keep them going. Often it doesn't, because sunk cost reasoning is weaker than pop psychology suggests. What actually keeps people going is anticipation of a specific, concrete reward.

The multi-step form problem

Here's a concrete example that anyone who's built a business website will recognise. A Sydney accounting firm redesigned its client onboarding form as a five-step wizard with a progress bar. Completion rates went up in the first three steps, then fell off a cliff at step four. Analytics showed the drop-off concentrated around the 70–75% mark.

The team assumed the fourth step was too long. They shortened it. Drop-off barely moved.

What was actually happening: step four was where users had to upload documents. The progress bar said they were 72% done, but the remaining work wasn't 28% — it was the hardest single task in the whole process. The bar was lying to them about effort, not about distance.

Once the firm changed the bar to reflect estimated effort remaining rather than steps completed, and moved the document upload to a point where users had already received something of value (a draft summary of their situation), completion improved. The bar didn't motivate anyone. It just stopped demotivating them.

What actually keeps people going past the dip

The research on reward loops — particularly the work on variable-ratio reinforcement that B.F. Skinner did with pigeons — suggests that unpredictable rewards sustain behaviour longer than predictable ones. But that's a dangerous lesson to take literally into web design. You don't want your users to feel like they're pulling a lever hoping for a pellet.

What you want is micro-rewards that arrive before the 72% mark, so that the user has already received something concrete before they hit the hard stretch. A few things that work:

Show value delivered, not just distance covered

Instead of "72% complete", try "You've saved your details — here's what we've got so far." The user sees a partial result. The remaining work now feels like finishing something, not starting something.

Break the bar into meaningful chunks

A single bar from 0 to 100 is a long, abstract journey. Four or five labelled segments — each with its own tiny completion moment — give the user multiple finish lines. This is the same principle behind the coffee card with two stamps already filled in. The perceived distance shrinks.

Put the hardest step first, not last

This is counterintuitive but it works. If the difficult task is going to cause drop-off, better it happens at 15% than at 75%. Users who get past it are more committed, and the rest of the process feels like downhill.

Never let the bar hit 100% before the reward does

If the bar completes and then there's a loading screen, a "check your email" message, or a payment page, you've broken the implicit contract. The reward must arrive at the moment of completion, or the bar is just a lie with a nice gradient.

Where this is heading

The interesting frontier here isn't better progress bars. It's adaptive ones. If you know from analytics that your users tend to stall at 72%, you can design the experience to deliver a small, genuine reward right before that point — a preview, a confirmation, a piece of useful information — so the dip never happens.

That's not manipulation. It's just paying attention to how motivation actually works: not as a straight line from start to finish, but as a series of small decisions where the user is constantly asking "is this still worth it?" Your job is to make sure the answer stays yes, all the way to the end.