Why a progress bar can make you quit
The goal-gradient effect is real, but it cuts both ways: if people speed up near the end, they are slowest at the start — which is exactly where a bar sits near zero and reads like a verdict. What to measure instead.
Progress bars have become the default way to show someone how a long piece of work is going. They are on fitness apps, language apps, savings accounts, reading trackers, and almost every goal tool including this one. The assumption underneath them is that seeing progress produces more progress.
That assumption is half right, and the other half is where people quit.
The finding everyone cites
The goal-gradient effect goes back to Clark Hull in 1932, who observed that rats ran faster the closer they got to the food at the end of a runway. The idea sat in the psychology literature for decades before Kivetz, Urminsky and Zheng brought it to consumer behaviour in 2006 with a coffee-shop card study. Customers were given loyalty cards; some needed ten stamps for a free coffee, others needed twelve but arrived with two stamps already applied. Both groups faced the same actual work — ten purchases. The group holding the twelve-stamp card with a head start completed it faster, and across the study people bought coffee at shorter intervals as they got nearer to the reward.
It is a genuinely useful result, and it is worth being precise about what it is not.
- It is a field experiment about a small, immediate, certain reward — a free coffee. A goal like changing careers offers nothing comparable in structure.
- The effect sizes are modest. The intervals between purchases shortened; nobody was transformed.
- It measures behaviour on a reward schedule, not the effect of displaying a bar. "People move faster near the end" and "showing someone a percentage helps them" are two different claims, and the second one does not follow from the first.
- It is one line of research in a field going through a serious replication reckoning. Treat it as a plausible tendency, not a law.
The half nobody quotes
If motivation rises as you approach the end, then by the same curve it is at its lowest at the beginning. The goal-gradient effect is not a story about how progress bars help. It is a story about a slope, and the flattest, coldest, least motivating part of that slope is the first stretch — which is also the longest part of any goal worth having, and the part where nearly all abandonment happens.
A bar makes that stretch visible. Three weeks of honest effort on a year-long goal produces a bar that has barely left the left-hand edge. The bar is not lying. That is exactly the problem: it is an accurate report that reads like a verdict. You did the work, and the interface responded with a sliver of colour that looks indistinguishable from having done nothing.
Endowed progress, and why you should not use it on yourself
The head start in the coffee-card study has a name: endowed progress. Give someone two free stamps and the same ten purchases feel like finishing something already begun rather than starting something from zero. Every app that shows you "step 1 of 5 complete" for having typed your email address is running this.
It works, within limits, and it is a manipulation. That is fine when a shop is trying to sell you coffee and you can see what is happening. It is worse when you do it to yourself, because you are both the manipulator and the mark. A plan that counts "chose a goal" and "downloaded the app" as completed steps is a plan that has spent its opening credibility on nothing, and you will notice. The moment a tracker gives you progress you did not earn, every number it shows you afterwards becomes slightly less believable — including the ones that were real.
When the bar tells you something you did not want to know
The second failure is the discouragement effect. You had a rough sense that you were most of the way through. The bar says thirty-one percent. It was not wrong before — you simply had never quantified it, and vagueness was doing quiet motivational work.
Sometimes that correction is exactly what you need, and a bar that stops you deceiving yourself has earned its place. But if the gap it reveals is large enough to look unbridgeable, the usual response is not to work harder. It is to close the app. There is a threshold past which accurate information about how far you have to go stops being useful and starts being a reason to stop, and no tracker knows where your threshold is.
Percentages of unknowable totals are fiction
A bar needs a denominator. Some goals have one: read forty-five of sixty chapters, save eight thousand of twenty thousand, complete nine of twelve sessions. Those bars are honest.
Most real goals do not. What percentage of the way through learning a language are you? Through changing career? Through recovering from something? Any bar answering those questions has invented its denominator — usually by counting how many of your written steps are ticked, which measures your planning, not your progress. Tick every step in a plan that was wrong and the bar reads one hundred percent while the goal is untouched.
Outcome bars punish real work
A bar that only moves when something finishes is flat during exactly the periods that matter most: the fortnight of reading before you can write anything, the month of practice before anything sounds different, the weeks when the code does not run yet. These stretches are not failure. They are what the work looks like from the inside, and an outcome-only indicator reports them as nothing happening.
What to use instead
Phases, so something actually finishes
Cut the goal into two to four stretches of weeks, each with an end you would recognise. Then you complete something every few weeks rather than once a year, and the goal-gradient slope — whatever its real size — gets to work near the end of each phase instead of once, at the very end, when you are no longer there. Northstar structures goals this way for this reason: phases sit between the goal and the daily steps, in order, each one a finish line you can actually reach.
Count inputs alongside outcomes
Days logged. Steps done. Weeks in which you turned up at all. These are the things you control, and unlike an outcome bar they move on the days when the work is genuinely hard and produces nothing visible. Northstar's seven- and thirty-day review is built around this: it reports your completion average, how many days you logged anything, and what share of your logged time was linked to a goal — three numbers that describe your behaviour rather than one number that grades your position.
Ask whether you are behind, not what percentage you are
A percentage on its own is meaningless without time. Forty percent is excellent in month two and alarming in month eleven. The useful reading combines the two — how far along you are against how much time is left — and returns something you can act on rather than a number you can feel bad about. Northstar computes a goal-health reading of ok, warning or risk from days remaining against progress, which is deliberately coarse. Three states are enough to tell you whether to continue, adjust or cut the goal, and coarse states are harder to obsess over than a decimal.
Show nothing rather than something made up
If there is not enough data to say anything true, the honest display is an empty state. Northstar's alignment score returns nothing at all when you have not logged enough — no zero, no encouraging placeholder. A fabricated number is worse than a blank, because a blank tells you to go and do something and a fabricated number tells you a lie you will later have to unlearn.
On streaks, briefly
Streak counters are progress bars with a cliff. They work by the same mechanism and fail more sharply, because losing thirty days of accumulated evidence to one missed Tuesday makes the missed Tuesday feel like it destroyed something. It did not. In Northstar, breaking a streak resets the count and nothing else — every logged day, every completed step and every note stays exactly where it was. Worth knowing, because the belief that you have lost your record is what turns one missed day into a fortnight off.
The honest summary
Measurement is not neutral. Whatever you display becomes what you optimise, and whatever sits near zero for a long time becomes a reason to stop looking. That is not an argument for tracking nothing — untracked goals fail quietly and you find out much later. It is an argument for measuring the things that move when you do the work, keeping a coarse honest reading of whether you are behind, and structuring the goal so that something real finishes every few weeks.
A bar that has not moved in three weeks is telling you the truth. Just make sure it is telling you the truth about a period, not delivering a verdict on a person.
Frequently asked questions
Is the goal-gradient effect real?
The tendency to accelerate near a finish line has been observed since Hull's 1932 rat studies and in Kivetz and colleagues' coffee-card field experiment. But the effect sizes are modest, the consumer work concerns small immediate rewards, and it shows people speeding up near a goal — not that displaying a progress bar helps. Treat it as a plausible tendency rather than a law.
Why do I feel worse after checking my progress?
Usually because the bar has quantified a gap you had only ever felt vaguely, or because it sits near zero during the long early stretch where the same goal-gradient curve says motivation is lowest anyway. The reading is accurate; the discouragement comes from an accurate report arriving in a format that looks like a judgement.
Should I give myself a head start to stay motivated?
Endowed progress works in shops, but doing it to yourself corrodes every number your tracker shows afterwards, because you know which ones you handed yourself. The legitimate version is to make the first phase genuinely short, so that real completion arrives early.
What should I track instead of a percentage?
Inputs you control — days logged, steps completed, weeks you turned up — alongside a coarse reading of whether you are behind relative to the time left. Northstar's review reports completion average, days logged and goal-linkage share, and its goal-health reading returns ok, warning or risk rather than a decimal to obsess over.
Try it on a real goal
Northstar cuts a goal into phases and phases into steps you write yourself, then names one next action a day. Its statistics come from your own entries, and it shows an empty state rather than a placeholder number. Everything stays on the device — no account, no sync, no analytics, no notifications. Free, with no in-app purchases.