011 min
The stove problem behind natural mapping
A cook stands at a four-burner stove with a pot of water on the back-left burner. The cook turns the second knob from the left. Nothing happens under the pot. A minute later the front-left burner is hot, with nothing on it.
The cook was not careless. The burners sit in a square of two rows, but the four knobs sit in one straight line along the front. Nothing about where a knob sits says which burner it heats. So the cook guesses, or bends down to read a small printed diagram beside each knob.
Four knobs in a row can be linked to four burners in 24 different ways (4 × 3 × 2 × 1). Each maker can choose a different one. A person who cooks in two kitchens has to learn two different answers. When they are in a hurry, their hand goes to the answer they learned first.
The fault is not in the knob or in the burner. It is in the link between them, which this stove does not show. Designers call that link a mapping: the relationship between a control and the thing it changes.
022 min
Where natural mapping comes from
The cook's mistake had been measured in a laboratory before anyone treated it as a design problem.
Before the name: stimulus-response compatibility
In a 1953 test by Paul Fitts and C. M. Seeger, people pressed a left or right button when a light came on to their left or right. People took longer to respond when the light and the button were on opposite sides. Researchers call this effect stimulus-response compatibility. A signal (the stimulus) and the action it asks for (the response) are compatible when they share a place or a direction. Compatible pairs are faster and cause fewer errors.
The first stove experiment
Alphonse Chapanis and L. E. Lindenbaum took the idea into the kitchen in 1959. They built a model stove with coloured plastic discs that lit up as burners. Then they timed people as they turned on or off the burner they were asked for, with the knobs linked to the burners in several different ways. Their clearest result came from moving the burners, not the knobs. When the burners were shifted slightly so that each one lined up with its own knob, people agreed strongly on which knob went with which burner.
Norman names it
Don Norman gave the idea the name most designers use today. He wrote about it in his 1988 book The Psychology of Everyday Things, later published as The Design of Everyday Things. He used the stove as his main example. He also wrote a short test that design courses still repeat: "If a design depends upon labels, it may be faulty." Norman's argument was that good mappings cut down how many labels a product needs. A label is still useful. But a control that needs one has a layout that does not explain itself. The stove became the standard teaching case, and many later writers reused it, sometimes without giving Norman credit.
033 min
How natural mapping works
Norman's test explains what the cook was missing. On a well-mapped stove, the answer to "which knob?" is in the layout, so the cook reads it with their eyes instead of recalling it. Norman called this putting knowledge in the world instead of in the head. Each step of translation that the layout removes is one less step where a person can go wrong.
Norman traced natural mappings to two sources: physical analogies and cultural standards. Jan Borchers' design course at RWTH Aachen University splits the physical side in two. That gives three kinds:
- Spatial analogy. The controls are arranged the way the things they control are arranged. The left switch turns on the left lamp. A car stereo's fader slides forward to send sound to the front speakers. Direction counts too: push a side-mirror control up, and the mirror tilts up.
- Perceptual analogy. The control looks like the thing it moves. A seat control shaped like a small seat lets a driver push the backrest part to tilt the backrest.
- Biological or cultural analogy. People share expectations that do not come from the object at all. A rising level means more, and a falling level means less.
Researchers who build tangible interfaces, which are physical objects used to control software, rely on the same spatial skills. People already know how objects in a room relate to each other, so a layout that copies the room needs no teaching.
Mapping is also about time. When a lamp comes on the moment its switch is pressed, the person links the two because the effect is visible at once. If the effect is delayed or hidden, people cannot tell which control caused it, and they press again.
Which quantities can go "up"
The rule that up means more has a limit. It works for quantities that add up: an amount of water, heat, volume, brightness, line thickness, weight. It does not work for qualities that replace one another, such as colour, taste or the pitch of a sound. A slider that runs from red to blue has no "more" end, so its direction has to be learned. The RWTH course notes list pitch in this second group, which surprises many people.
Position, not the hand
Fitts and Seeger also asked why the same side was faster. One idea was that signals take longer when they cross from one half of the brain to the other. To test it, they had people cross their arms, so the left hand pressed the right button. The difference in speed between crossed and normal arms was too small to matter. What counted was where the button sat, not which hand pressed it. For design, this means a control should sit where its effect is, even when that is not where the user's stronger hand rests.
042 min
What the stove studies found about natural mapping
If the fix has been known since 1959, the next question is why stoves kept the row of knobs. Later studies gave several answers.
The layout nobody expected
In 1978, David Shinar and M. B. Acton asked 222 people which knob they would turn for each burner of a drawn stove. No single answer won: people split across at least four different linkages. But 98% agreed on one thing. The right knobs should work the right burners, and the left knobs the left burners. The same study looked at 49 stove models then on sale and found five different layouts. The most common one was chosen by none of the 222 people, and it broke the left-right rule they agreed on.
Preference is not performance
Asking people which layout feels natural is not the same as watching them use it. Hoffmann and Chan's review of the stove research found gaps between the layouts people said were best and the layouts they made the fewest errors with. A survey can suggest a candidate. Only a timed test on real controls shows which layout people use correctly.
The same stove in Taiwan
In 1993, Sheng-Hsiung Hsu and Yu Peng repeated the stove tests in Taiwan, with a paper questionnaire and a computer simulation. Their Chinese participants expected a different knob-to-burner linkage from the one American participants had preferred in earlier studies. A layout that one population reads at a glance can mislead another. A product sold in several countries needs testing with people from more than one of them.
The test itself can steer the answer
Hsu and Peng found a second problem. The letters, signs or numbers used to mark the burners on a test sheet pushed some people toward particular answers without their noticing. The codes on the test, not only the layout, shaped the result. So a mapping test should show the layout as plainly as possible, ideally without letters or numbers that suggest an order.
Fix the layout, not the label
Errol Hoffmann and Alan Chan tried a different route in 2011. Instead of looking for the best linkage for a square of burners, they let one group of participants place burners and another group place knobs. Participants kept choosing layouts in which each burner lined up with its own knob. Chapanis's model stove had shown the same thing in 1959: the most reliable fix is to move the parts until the layout explains itself.
052 min
How natural mapping shows up in real interfaces
Stoves are hard because the knobs cannot sit on the burners. Many products have more freedom, and in the revised 2013 edition of his book Norman ranked the options from best to third best.
- Put the control on the object. A sensor tap has no handle. A person holds their hands under the spout and the water starts. There is nothing to map, because the control and the effect are the same thing.
- Put the control right next to the object. When each control sits beside the thing it changes, people can tell which goes with which from closeness alone.
- Arrange the controls in the same pattern as the objects. The window switches on a car's driver door copy the layout of the windows seen from above. The front pair works the front windows and the back pair works the back windows, each on its own side.
The car seat shows good and poor placement side by side. Mercedes-Benz puts its seat controls on the door, in the shape of a small seat: push the backrest part back, and the real backrest tilts back. Many other cars put plain levers and buttons on the side of the seat. The driver cannot see them while seated, so adjusting the seat becomes trial and error.
Natural mapping also works on screens. Te Papa, the national museum of New Zealand, built a touchscreen for its Anzac Print Gallery. It held 10 videos, one for each of the 10 prints on the gallery's four walls. The menu showed the videos in the same order as the prints hung on the walls, spread across four menu pages. A visitor could find the video for a print by where the print hung, not by its title.
Everyday screen controls follow the same rule. A volume slider that moves up makes the sound louder. A back arrow points left in languages that read left to right, because earlier pages sit to the left. The rule breaks when controls are separated from what they change. A toolbar of identical icons, where each one needs a hover to explain itself, is one example. A setting that sits three screens away from the feature it affects is another.
062 min
Natural scrolling: when natural mapping depends on the device
A volume slider works because nearly everyone shares the idea that up means more. Scrolling shows what happens when two groups of users hold different ideas.
For years, desktop computers treated scrolling as moving a window over a page. Roll the mouse wheel toward you, and the window moves down the page, so the text on screen moves up. Then touchscreens arrived. On a phone, the finger touches the page itself, so the page moves with the finger.
The 2011 switch on the Mac
In 2011, Apple changed the Mac to behave like its phones. OS X Lion made the touchscreen direction the default and called it natural scrolling. Apple describes the setting as moving the contents of a window the same way the fingers move. On a Mac before Lion, moving two fingers down the trackpad moved the page up; after Lion, the same movement moved the page down.
Many users disliked the change. A 2016 Low End Mac column by Dan Knight objected that the old direction had been the standard across the industry for years. It also argued that treating a scroll wheel like a finger on a screen is not intuitive. Both groups are right about their own device. With a finger on glass, the page-moves-with-the-finger mapping is natural, because the finger is on the thing that moves. With a wheel or a trackpad, nothing under the finger moves, so either direction is a habit.
The misreading: "natural" means built in
The word "natural" suggests a mapping that everyone is born with. Most are not. The scrolling dispute and the Taiwan stove results point the same way. Many mappings that feel natural are learned from the products and the culture a person grew up with. Even the stove knob has no built-in link to heat, because turning something does not by itself suggest more warmth. A designer should read "natural" as "matches what these users already expect", and then check who these users are.
072 min
Where natural mapping stops working
Even a mapping that matches what users expect can fail, in two situations that the stove does not show.
More realistic is not always more usable
Video game controllers are a common place to see natural mapping tested. The Nintendo Wii let players steer in Mario Kart Wii by turning the controller like a steering wheel, often inside a plastic wheel. In 2012, Doug Bowman, Ryan McMahan and Eric Ragan reported a study that compared this with the older thumb joystick, using the game's time-trial mode. With the joystick, players drove faster and more accurately than with the wheel. The authors suspected the wheel was not realistic enough. It was held in the air with no fixed base, did not push back to the centre like a real wheel, and responded with a small delay. Small hand muscles may also be more precise than the arm movements a real wheel needs.
The same article describes designers borrowing only part of a real movement. In one Wii Sports Resort game, players ride a bicycle by moving their arms in the pattern of pedalling legs, because the controller cannot follow the legs. Other designs go past realism on purpose, such as a virtual arm that reaches further than the player's real arm stretches. A 2014 study by Mitchell McEwan and colleagues sorted three racing-game controllers by how closely their physical inputs matched the game's actions. It then tested them with 64 players to measure how intuitive each one was. A 2015 study by Masitah Ghazali, Alan Dix and Kiel Gilleade looked at people who overshoot a setting. They reversed the same movement on the same control, whatever the mapping.
When the user cannot see the layout
Natural mapping depends on seeing the layout. A blind person using a screen reader hears controls one after another, so the fact that a button sits to the right of a picture tells them nothing. WCAG 2.2 Success Criterion 1.3.3, Sensory Characteristics, covers this: instructions must not rely only on shape, size, visual location or orientation. Its guidance is to describe controls by name. A good layout can reduce visible labels, but every control still needs an accessible name that says what it does.
082 min
Natural mapping vs. affordance, signifiers and conventions
Natural mapping often appears next to other ideas from Norman's work, and each answers a different question. A stove knob can do well on one and badly on another.
| Concept | The question it answers | Stove example |
|---|---|---|
| DesignAffordance | ||
| DesignSignifiers | ||
| Not in the library yetNatural mapping | ||
| DesignJakob's Law |
The most important difference is between natural mapping and convention. A convention works because people learned it from other products. A natural mapping works because the layout shows the answer, even to someone who has never used this kind of product. Many mappings sit between the two, as the scrolling case showed.
Natural mapping is also not skeuomorphic design, which copies how a real object looks. Te Papa's gallery menu copied the order of the prints, not the look of the gallery. Its design guide says a realistic look was not the aim.
Mapping and signifiers both help people across the gulf of execution, the gap between what a person wants to do and knowing which action will do it. A good mapping also helps a person build an accurate conceptual model of the product, so they can predict what an unfamiliar control will do.
092 min
Applying natural mapping
Back at the stove, the cook needed one thing: to see which knob heats which burner without reading. These moves give users that on any product.
Start from what is being controlled.
Sketch the layout of the things that change, such as burners, windows, speakers or areas of a screen, before you place any control. Then copy that layout.
Use Norman's order of preference.
Put the control on the object if you can, next to it if you cannot, and in a matching pattern as the last option.
Match direction as well as position.
Use up for more, and movement in the same direction as the effect, for quantities that add up. For colour, pitch or options with no order, use labels.
Test performance, not opinion.
Time people on the real control and count wrong first moves. People can rank a layout highly and still use it badly.
Test with the users you will sell to.
Expectations about order and direction differ between populations and between devices.
Keep accessible names.
The layout can remove visible labels, but a screen reader still needs a name for every control.
How to spot natural mapping going wrong
- In session recordings, people reach for the wrong control first and then correct themselves.
- Users add their own labels to the product: tape, stickers, notes. A label added by a user marks a layout that failed.
- Support tickets ask which control does what.
- A setting is changed and then changed straight back, the overshoot-and-reverse pattern.
Pick the control people use most often, check it against these signs this week, and fix its layout before you add another label.
?4 questions
Questions people ask
Is natural mapping the same as intuitive design?
Does natural mapping apply to keyboard shortcuts?
How do right-to-left languages change natural mapping?
Why do most stoves still put the knobs in a row?
§16 sources
Sources for natural mapping
Rauterberg, M. Interaction Design lecture notes, slide "The Concept of Natural Mappings". Eindhoven University of Technology. A lecture slide that quotes Norman's 1988 definition and labels principle.
Borchers, J. (2024). Designing Interactive Systems I, Lecture 3: Mappings, Constraints, Seven Stages of Action. RWTH Aachen University.
Wikipedia. Stimulus–response compatibility (on Fitts and Seeger, 1953).
Hoffmann, E. R., & Chan, A. H. S. (2011). Alternative approaches to the design of four-burner stoves. Ergonomics, 54(9), 777–791.
Show all 16 sourcesShow fewer sources
Shinar, D., & Acton, M. B. (1978). Control-display relationships on the four-burner range: population stereotypes versus standards. Human Factors, 20(1), 13–17. Abstract record, Ben-Gurion University research portal.
Hsu, S.-H., & Peng, Y. (1993). Control/display relationship of the four-burner stove: a reexamination. Human Factors, 35(4), 745–749.
Bowman, D. A., McMahan, R. P., & Ragan, E. D. (2012). Questioning naturalism in 3D user interfaces. Communications of the ACM, 55(9).
Apple. Change Trackpad settings on Mac. Mac User Guide.
Knight, D. (2016). Apple's Natural Scrolling: There's Nothing Natural About It. Low End Mac.
W3C. Understanding Success Criterion 1.3.3: Sensory Characteristics. WCAG 2.2.
Museum of New Zealand Te Papa Tongarewa. Natural mapping, a page of the museum's design principles for interactives.
Shortform. UX Mapping: Using Memory and Human Experience (a summary of Norman's The Design of Everyday Things).
McEwan, M., Blackler, A., Johnson, D., & Wyeth, P. (2014). Natural mapping and intuitive interaction in videogames. CHI PLAY 2014.
Ghazali, M., Dix, A., & Gilleade, K. (2015). The relationship of physicality and its underlying mapping.
Wikipedia. Natural mapping (interface design). (interface_design)
Dorin, A. (2008). FIT3084 Lecture 4: Introduction to HCI. Monash University. Lecture notes summarising Norman's The Design of Everyday Things.