Goal gradient

The goal gradient is the oldest finding in this corner of the vocabulary and the least obvious to design around. Clark Hull described it in the 1930s from animals running a maze: the closer the food, the faster the run. People do the same thing with tasks, and Ran Kivetz, Oleg Urminsky and Yuhuang Zheng found it on coffee loyalty cards in 2006, where the gap between one purchase and the next gets shorter as the card fills up. Effort is not flat across a task. It rises as the end comes into view, and it rises because the end is visible, which is the part an interface actually controls.

So a design that wants the effect reports the remainder rather than the total. “3 of 10 collected” is a statement about the past, and it grows. “7 to go” is a statement about the distance left, and it shrinks, which is the number a person is actually running toward. The intensification that follows is real work rather than decoration: near the goal the label sharpens, the fill earns emphasis, and the untaken stretch is drawn as a small named quantity instead of as empty track. Far from the goal the same track should be quiet, because a bar that shouts at 2 of 10 has nothing left to say at 9.

Three neighbours sit close enough to be worth separating. Steps left is the same instinct applied to a bounded flow, naming how many screens remain so the task can be priced before it is begun. A progress bar is the component the gradient is usually drawn on, and it is indifferent to where the fill sits: it reports a fraction, and the goal gradient is a claim about what the reader does as that fraction approaches one. Endowed progress is the other half of the pair: the goal gradient is about effort rising as the end comes into view, while endowed progress is about where the meter is allowed to start.

The effect is easy to counterfeit and the counterfeit is worth naming, because it spends the only thing that makes the track work. Keep the units even: a final segment drawn shorter than the others is a lie about distance rather than a report of it. Do not manufacture nearness by splitting the last step into three, and do not park a bar at 95 percent while the real work continues. Both buy one burst of effort, once, from a reader who will read the next bar differently.