Measurement units UX: change display, not data

Name actual units when the bundle is not enough

Zepp separates distance, weight, temperature and another measurement category, showing concrete units beside the settings. LazyFit presents height and weight as distinct selections. This approach lets a person recognize the result before making a change rather than translating a broad system name into every affected unit.

Use a concise summary on the parent settings row. Distance: kilometers is more informative than a chevron leading to an unlabeled Metric choice. Where a combined unit such as feet and inches is supported, show an example format so the input model is understandable.

Do not assume that language, location and measurement preference are identical. An initial regional default can reduce setup work, but retain an explicit override and preserve it. Changing the interface language should not quietly rewrite carefully chosen units unless the product clearly explains and requests that broader change.

Use system presets without pretending every category behaves alike

Monju offers a Metric bundle containing kg, cm and ml and an Imperial bundle containing lb, ft/in and oz. CREME separates measurement system from temperature and notes that some cooking measures are used across both choices when suitable. These examples reveal the limits of a single switch.

A preset can establish several defaults at once, but explain which categories it affects. If later overrides are allowed, represent the resulting mixed state honestly instead of continuing to display Metric as though every value followed the preset. Avoid making a minor temperature change reset unrelated preferences.

Treat recipe measures carefully as an interface model. A count, a spoon measure, a mass and a volume are not interchangeable labels for one stored number. Whether a product can translate between them depends on the underlying information. Do not promise a universal conversion merely because two options appear in the same preference panel.

Separate allowed input measures from display preference

BitePal combines a Metric or Imperial selection with ingredient-unit options such as grams, slices, cups and pieces. CalCounter+ separates Weight, Height, Length and Distance and previews a height format. One surface controls which measurement choices are available; the other describes a preferred representation for a category.

Do not use the same switch language for both jobs. Disabling a unit from future ingredient entry should not imply that historical entries in that unit will be deleted. Likewise, choosing a preferred display unit does not necessarily mean other entry units are forbidden.

Explain what happens when an older item uses a now-hidden measure. Preserve the original quantity and provide a readable fallback rather than replacing its label with a preferred but incompatible unit. Counts such as pieces can remain useful even when the person usually views mass in grams. The information model should lead the control, not the other way around.

Explain when the new representation takes effect

AI Health Pulse states that data will be displayed in the new unit and includes a Save action. Alive shows an Imperial or Metric weight choice with Save in a profile context. A separate confirmation provides room for a pending preference, but the interface needs to say what remains unchanged while it is pending.

If changes apply immediately, update the settings summary and relevant preview consistently. If they require Save, distinguish the candidate from the committed preference and keep Back behavior predictable. Leaving the page should not apply some categories and discard others without explanation.

Consider an active entry form elsewhere in the app. A person may have typed a number under the old unit before opening settings. Preserve that draft’s unit or deliberately convert its displayed value with clear context. Never leave the same digits in place, change only the suffix and treat the result as though the person entered it that way.

Allow useful category overrides without losing the overview

Livity lists units separately for distance, temperature, energy, weight and water. MacroFactor shows weight in kilograms while height uses feet and inches, alongside a separate clock preference. A mixed configuration is not necessarily a mistake; it can reflect how the person actually understands different quantities.

Keep each summary readable and provide a predictable way to change one category. Do not force a user through a large regional setup flow to switch a single preference. If the product offers Reset to regional defaults, describe the scope before replacing several independent choices.

Carry unit labels through charts, goals, history and shared output. A preference page can be perfectly clear while an old chart axis silently keeps another unit. Review summaries and detail views together. If an external export intentionally uses a fixed representation, disclose it rather than letting the recipient assume it follows the current on-screen setting.

Keep quantity, source unit and display policy explicit

Store enough information to interpret the original quantity reliably. A normalized internal amount can support consistent display, but retain source context where needed for editing, auditing or unsupported conversions. A bare number with a globally mutable unit label is not a dependable history model.

Use a defined conversion and rounding policy at presentation boundaries. Repeatedly switching units should not progressively alter the saved source through round-trip rounding. Keep display precision appropriate to the task and avoid implying more precision than the original input contains. A formatted number is not a new measurement.

Treat preference propagation as an observable state change. Update relevant views together, including cached summaries and open drafts, or explain any delayed application. Keep incompatible quantity types separate and do not manufacture a conversion where required context is missing. The user interface should expose that limit instead of returning a plausible-looking but unsupported number.

Use a round trip and an unfinished draft

Record a value, change the display unit, switch back and compare the saved source. Repeat several times to catch cumulative rounding. Include an empty field, a decimal input and a combined-unit format. Verify that selection, validation messages and accessible labels describe the actual unit used for entry.

Leave a draft open, change a preference and return. Check whether the same digits still mean the same quantity. Test a historical item using a measure hidden from new entry, and inspect chart labels, goals, exported files and shared summaries. These are different surfaces of the same representation contract.

Evaluate reinterpretation errors and inconsistent labels separately from preference saves. A successful Save event does not establish that historical data remains understandable. Ask the tester to explain an existing value before and after the change. The desired outcome is continuity of meaning, not merely a fast toggle.

Unit preferences need more than a system switch

Use these checks on a working build, not only on the polished reference state.

• ✓Every setting exposes the actual affected unit or an understandable example.
• ✓System presets and category-specific overrides have explicit scopes.
• ✓Allowed entry measures are distinct from preferred display units.
• ✓Saved quantities are not reinterpreted when preferences change.
• ✓Open drafts retain their unit meaning through a settings change.
• ✓Round-trip changes do not accumulate source-value rounding.
• ✓History, charts, goals and output show the correct representation.

Connect preferences to the data people already trust

Use the settings-screen examples to communicate scope, pending changes and the difference between preferences and account actions.

The calorie-tracker design article supplies a broader product context for measurement entry without turning a display preference into a calculation guarantee.

Compare sharing and export screens when deciding whether outgoing data follows the user’s display preference or a documented fixed format.