Introduction

A procurement team checks the price text on a vending screen and finds it readable. Beside the price sits a pale selection mark. The manual describes the mark, yet the customer still needs to perceive it on the live screen. Text readability and mark visibility answer different questions. Approval should identify which visual parts carry the task information and which colours immediately surround those parts.

This guide focuses on required non-text visual information: an icon that identifies a function, a check that indicates selection, a boundary needed to locate an input, or a graphical feature needed to understand an operator chart. These are proposed review fixtures, not features observed on the products below. The buying objective is to obtain evidence for the actual interface rather than a general statement that its design has “good contrast.”

W3C WCAG 2.2 Success Criterion 1.4.11, Non-text Contrast, is the primary reference. It is a Level AA web-content criterion. Applying its concepts to a software procurement scope does not certify a physical machine or settle local legal obligations. Define the relevant software surfaces and request a competent evaluation of the delivered configuration.

Quick Answer

The reference threshold is at least 3:1. W3C says computed values must not be rounded up: 2.999:1 does not meet 3:1. Record the actual colours and result, not just a rounded green badge from a tool. Default, selected and focused versions can place the same mark against different adjacent colours.

Do not measure every decorative line indiscriminately. A control does not always require a visible border around its hit area if another visible feature identifies it. Inactive controls and appearances determined by the user agent and unmodified by the author have specific exceptions. Meaningful graphics also have an essential-presentation exception. Classify the case before calling it a failure.

Comparison Table

These rows guide a buyer-requested demonstration. None establishes that a particular listed product includes these controls or graphics. Confirm the surfaces in the ordered configuration first.

Who Should Buy This

Use this scope when the purchase includes a customer menu with symbols, custom selection indicators, authored focus treatments or an operator view containing essential graphical information. It is useful when a supplier provides screenshots but no record of which non-text elements were measured. Buyers need a traceable result for the required indicator, not a measurement of a nearby title that happens to pass.

A brand palette can create different problems across surfaces. A white icon on a coloured badge has one adjacent background; an external outline around a button may sit on the page background. An operator chart may place marks against both a plot area and a highlighted region. Grouping all three under one approved brand colour does not establish their individual contrast.

If the purchase contains no relevant authored interface, narrow the request accordingly. Hardware lighting, glare and cabinet visibility can require separate reviews. This article does not treat a web-content contrast ratio as a complete assessment of a machine under outdoor illumination or a customer’s viewing distance.

How We Evaluate Smart Vending Machines

The shortlist uses three public WEIMI product listings reviewed on 10 October 2026. We have not independently tested their displays or measured installed icon colours. We compare the documented retail formats and the non-text contrast evidence a buyer should ask for. A product photograph does not establish compliance.

For a proposed acceptance review, create an object record: screen and build, control or information, required visual part, operating state, foreground colour, adjacent colour, measured ratio and interpretation. Attach a screenshot showing where the object occurs. If a graphical part is not required because equivalent information is available, document that alternative and how the user reaches it.

W3C Technique G207 covers icons required for understanding, including solid icons on solid or gradient backgrounds. Its procedure begins with the foreground and adjacent background, testing the least contrasting part of a gradient or pattern. Where part is below 3:1, the procedure asks whether the icon remains recognisable if that area is treated as invisible. This is more specific than averaging several sampled values.

Techniques are examples of ways to meet the criterion, rather than mandatory implementations. The supplier may use another supported approach. What matters for procurement is a clear result tied to the actual required visual information and the delivered build, with exceptions and unverified areas openly identified.

Key Buying Factors

Measure the part carrying information

An icon may sit in a circle that contrasts strongly with the page while the icon itself barely contrasts with the circle. If the symbol is what communicates the function, measure the symbol against its immediate background. Conversely, when a control’s fill is the feature that identifies its presence, its contrast with the surrounding area can be relevant. A good ratio for the wrong pair is weak evidence.

Keep state checks separate from before-and-after comparisons

SC 1.4.11 does not require the colour of a component in one state to contrast 3:1 with its colour in another state when the two are not adjacent. The required indicator must remain distinguishable from its neighbours in each applicable state. Measuring an unfocused blue against a focused green can therefore miss a faint mark on the actual background. Hue-only meaning is a separate Use of Color question.

Explain whether a boundary is necessary

Do not impose a rule that every button border must contrast 3:1. W3C explains that a control with visible content sufficient to identify it does not necessarily need a boundary around the hit area. An input whose boundary is the only visual way to identify its presence is a different case. Keep the identification rationale in the object record. A contrasting focus indication still needs its own review.

Review fine marks beyond a nominal pass

W3C notes that anti-aliasing can make very thin lines appear fainter than their defined colours. Evaluate using colours obtained from the user agent or underlying authored styles rather than treating softened edge pixels as the normative colour pair. Then inspect the rendered shape. A nominal ratio does not make a hairline a strong practical indicator; thicker shapes or additional contrast may be a sensible design choice.

Use exceptions accurately

An inactive control is visible but unavailable for interaction; a supplier should demonstrate that state before claiming its exemption. An unmodified user-agent appearance is different from an authored replacement. An essential graphic presentation also needs a rationale. Branding alone is not a blanket explanation for an optional low-contrast treatment chosen by the author. Record the basis rather than labelling every difficult graphic exempt.

Preserve evidence across themes and backgrounds

A necessary arrow on a solid tile may pass while the same arrow on an advertising gradient loses its recognisable shape. Request the relevant configured backgrounds, including content changes the operator can make. For a gradient, investigate the least contrasting region and the remaining information. Do not substitute the best-looking screenshot for the configuration intended for delivery.

Best Smart Vending Machines

The following real products are buying candidates for different retail formats. “Best” means a possible fit after configuration and evidence review, not an independent performance ranking. None has a verified non-text contrast score in this article.

Feature Comparison

There is no defensible winner based solely on non-text contrast because the listings do not provide the necessary measured interface data. Ask the supplier to identify the exact build and deliver an evidence set. A proposed redesign is useful design material, while a verified acceptance result requires evaluation of the agreed configuration.

Cost & ROI Analysis

This is a hypothetical review allowance, not a machine price, supplier quotation or measured saving. Assume a reviewer inventories 16 required graphical indicators and reviews their applicable states in 14 hours at $75 per hour: $1,050. Assume design changes cost $850 and release verification costs $350. The initial allowance is $2,250. Hardware, freight, tax, stock and payment charges are excluded.

The formula is allowance divided by assumed annual benefit, multiplied by 12. It ignores discounting and future maintenance. No increase in vending revenue or reduction in support requests is established here. Replace these inputs with recorded costs and a supplier quote before making an investment claim.

Ask the quotation to separate object inventory, measurement, design revision and regression verification. A quote that changes one icon colour may leave several states and backgrounds outside scope. Define what happens when operators change a banner, a palette or a dashboard view after handover. The cost of maintaining the evidence depends on those actual controls and release responsibilities.

Best Choice by Scenario

For packaged drinks sold through an access-based fridge, consider the single-door AI format once the access and payment configuration is confirmed. The useful acceptance request is an inventory of required visual information along that route. Avoid demanding an imagined menu when the ordered configuration has another interface.

For a touchscreen snack and drink selection process, consider WM22 after the delivery mechanism is confirmed for the product range. Request a configured selection-indicator demonstration and the measured adjacent-colour pairs. The screen size is a format fact, not proof of the visibility of its smallest essential mark.

For two selling areas with a shared main display arrangement, consider the dual-cabinet station after mapping and installation review. Include any required graphical identification of the selling area. Choose the candidate that fits the retail operation and supplies evidence for its actual indicators; do not choose on a generic claim that all graphics pass.

Applications

A proposed icon-only support control uses a telephone shape inside a badge. The badge-to-page ratio does not answer whether the telephone shape is perceivable inside the badge. The acceptance record should identify the symbol as the required object, retain its immediate colour pair and show the configured state. Whether the icon’s name and action are accessible requires separate checks.

A proposed stock graph uses a line to show a quantity over time. Determine whether the line and any necessary value guides carry information the user needs. If an equivalent labelled value table is available in a conforming manner, document that route and review whether the graph remains required for understanding. Do not assume that a hidden export automatically replaces the active chart’s information.

A proposed gradient promotion tile contains a necessary arrow. Use the least contrasting adjacent region as the first check. If a region is below the threshold, assess whether the remaining shape still communicates the arrow when that region is treated as invisible. Record the visual reasoning. Averaging a dark region and a light region can conceal the very part that makes the symbol recognisable.

FAQ

Final Recommendation

Approve a record that names the required visual object, its state and its adjacent colours. A single palette certificate or a screenshot with readable words is not enough to assess a necessary check mark, arrow or control indicator. Retain the measurement and the reasoning that explains scope, exceptions and recognisability.

Review the actual delivered software, then repeat the relevant fixtures when authored themes or backgrounds change. Keep web-content contrast evaluation, site visibility conditions and whole-machine obligations distinct in the acceptance scope. The practical outcome is traceable evidence for important visual information, without pretending that a public product listing supplied an independent test.

CTA

Send WEIMI the intended product range, machine format and quoted software surfaces. Include the icons, indicators or diagrams required for the operation and ask for a build-specific evidence proposal. Request separate pricing for design changes and verification so the acceptance deliverables are clear.