To choose between app icon options, place them side by side on the same background and view them at 64, 32, then 16 pixels. Check silhouette, legibility, contrast, differentiation, and fit with the product. Keep qualitative notes instead of a score that claims to measure user response.
Set up a fair comparison
- Put the options in a grid and give them neutral names such as A, B, and C.
- Show them at the same dimensions and on the same background. Repeat the check on a light and a dark background.
- Reduce each image to 64, 32, and 16 pixels. These are check sizes for this exercise, not store compliance thresholds.
- Hide file names during the first pass so you can focus on the shapes.
A set of questions, without artificial scores
| Criterion | Question | Observation to keep |
|---|---|---|
| Legibility | Does the main shape remain visible at 16 px? | Detail preserved, merged, or lost. |
| Silhouette | Can you distinguish the options when details are hidden? | Distinct outline or overly familiar shape. |
| Contrast | Does the symbol separate from light and dark backgrounds? | Readable area or area that blends in. |
| Differentiation | Do the options represent genuinely different directions? | Difference in metaphor, shape, or finish only. |
| Product fit | Does the idea connect to the purpose and audience in the brief? | Clear, uncertain, or missing connection. |
Options A to C are illustrations from Iconise’s public gallery. They support the reading exercise; no prompt, ranking, or test result is attributed to them.
An illustrative decision
Imagine that, in your own reduced view of the three illustrations, A keeps a clearer outline than B. Record that observation, then check whether A also suggests your app’s purpose more clearly. If B explains the product better, silhouette alone does not decide. This is a way to reason through a choice; it does not report a test of these SVGs or a measured preference.
Confirm the choice in context
A reduced preview cannot replace checking the icon in the interface and on the backgrounds where it will actually appear. Masks and requirements differ between an operating system, a store listing, and a browser. Keep two options if the criteria are still split, then ask for feedback with the same context shown to each person; do not present a small exercise as a representative user test.
- Keep one primary option and explain how it answers the brief.
- Keep an alternative only if it contributes a distinct strength.
- Check the iOS, Android, or web outputs you need before preparing the files.
The sizes and formats guide covers those use cases and links to official sources.