Definition
Function is what an object is designed to do. Use is what people actually do with it, in real conditions, with their own habits and goals. The two often diverge. A chair's function is to support a seated body; its uses include standing on it to reach a shelf, hanging a coat on it, and propping a door open. Designers control function directly but influence use only indirectly, through the cues the object provides. This distinction matters because evaluating a design against its intended function alone will miss most of what happens to it. Studying use reveals mismatches, unexpected needs, and failure modes that a specification document never anticipates.
Why It Exists
The separation of function from use became an explicit design concern once objects were made for strangers. A blacksmith making a tool for a known farmer could ask how it would be used and adjust. A factory producing a million units has no such conversation, so assumptions about use become embedded in the object and are only tested after release. Twentieth-century human-factors research showed that these assumptions were often wrong, and that people were reliably doing things designers had not planned for. Rather than treat this as misuse, designers learned to treat it as information. Understanding actual use became the justification for observation, field study and, eventually, the whole apparatus of design research.
Examples
- →A flat door plate signals push; a vertical bar signals pull, and a bar on a push door produces constant error.
- →A kettle lid that doubles as a measuring cup, an unintended use designers later formalized into the product.
- →Worn paths across grass showing where people actually walk rather than where the paving was laid.
- →A phone used as a flashlight, spirit level and mirror — functions the hardware allows but was not designed around.
History
The psychologist James J. Gibson introduced the concept of affordances in the 1970s to describe what an environment offers an animal: a surface of a certain height and rigidity affords sitting. Don Norman brought the idea into design in The Psychology of Everyday Things, published in 1988 and later reissued as The Design of Everyday Things, arguing that objects should signal their possible actions. His example of doors that invite the wrong action became so well known that the term Norman door entered common use. Norman later refined the vocabulary, distinguishing affordances — what is actually possible — from signifiers, the perceivable cues that tell a person what to do. That distinction remains central.
In Modern Design
Contemporary practice studies use rather than assuming it. Ethnographic observation, contextual inquiry and diary studies were adopted by consultancies such as IDEO and by in-house teams to watch people work in their own environments. Connected products add telemetry, so designers can see aggregate use patterns directly, which is powerful but partial: data shows what people did, not why. The same framework has crossed fully into software, where a feature's intended function and its observed use routinely diverge. The ethical dimension has also sharpened, since designers can now deliberately shape use through defaults and feedback loops, which makes influencing behavior a responsibility rather than a side effect.
Real-World Example
Norman's door is the canonical case. A glass door fitted with identical vertical handles on both faces affords both pushing and pulling, and signals neither. People approach, choose wrongly about half the time, and blame themselves. The fix is not a larger sign reading PUSH, because a sign is an admission that the form failed. The fix is to change the signifier: a flat plate where the hand should push, a graspable bar only where pulling works. The hardware then communicates the correct action without language, to anyone, in any script. This is the whole argument in one object — the design decides what people will do, and error is the designer's evidence.
Key Principles
- →Intended function and observed use are different; only one of them can be measured after release.
- →Affordances are what an object makes possible; signifiers are the cues that tell people what to do.
- →Unexpected use is information about unmet needs, not user error.
- →If a product needs an instruction label to be operated correctly, the form has already failed.
- →Feedback matters: people need to know an action registered, or they will repeat it.
- →Design constrains behavior whether or not the designer intended to, so that influence should be examined deliberately.
Why it matters
Designing for function alone produces objects that are technically correct and practically frustrating. Attending to use is what closes that gap, and it is the direct ancestor of user research in digital design. The framework also reassigns blame. When people consistently fail at a task, the honest conclusion is that the object misled them, which changes what a team does next: it investigates the signal rather than writing better instructions. In safety-critical settings — medical devices, vehicles, industrial equipment — that reassignment has saved lives. And because designed objects shape habits at scale, understanding use is also how designers become accountable for the behavior their work encourages.
Then vs Now
Then
A maker who knew the person using a tool could ask how it performed and adjust it, so function and use stayed close together.
Now
Products are made for strangers at scale, so use must be studied deliberately through observation, testing and telemetry, and designers are accountable for the behavior their defaults encourage.
Try it yourself
Pick a public object used by strangers: a ticket machine, a coffee machine, a public door, a recycling bin. Watch it for fifteen minutes without helping anyone. Record every hesitation, wrong action, repeated press, and moment where someone asked another person for help. Note the exact cue that misled them. Then write two lists: what the object was designed to do, and what people actually did. For each mismatch, propose a change to the object's form, labeling or feedback — not to the instructions — that would remove the error, and say how you would check whether it worked.