# Wardrobe App Benchmark 2026 — data dictionary

Canonical source: https://beautyai.app/wardrobe-app-benchmark-2026

- Dataset version: `2026.05`
- Snapshot date: `2026-05-02`
- Documentation update: `2026-08-08`
- Sample size: 11 apps
- License: [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)
- Preferred credit: `Source: Beauty AI — Wardrobe App Benchmark 2026 (v2026.05)`

## Files

- `wardrobe-app-benchmark-2026-v2026.05.csv` is the immutable May 2026 snapshot.
- `wardrobe-app-benchmark-2026.csv` is the stable latest-version URL. For this release its rows are byte-for-byte equivalent after line-ending normalization to the versioned snapshot.

Historical snapshots are added as new versioned files. Existing versioned files are not silently overwritten.

## Sample composition

The curated sample contains Beauty AI, Fits, Stylebook, Whering, Acloset, Indyx, Cladwell, Alta Daily, OpenWardrobe, SimpleCloset, and GetWardrobe.

An app was eligible when, at the snapshot date, it had an accessible official app-store or product listing and publicly positioned itself as a digital wardrobe, closet organizer, outfit planner, or AI styling product. The sample deliberately covers several product models and is not a random or exhaustive census of the market.

## Reproduction workflow

1. Open the `source_url` supplied for every row.
2. Record the public pricing model, supported platforms, published language count or label, feature claims, and store signal visible on the snapshot date.
3. Normalize pricing and platform names using the rules below.
4. Code capability depth with the published feature-group thresholds below.
5. Preserve uncertain or unavailable public information as descriptive text instead of estimating it.
6. Save a new versioned snapshot whenever row values are materially rechecked or changed.

Only official app-store listings and public first-party product pages are primary sources. Ratings, user totals, pricing, languages, and features are treated as dated signals.

## Data dictionary and coding rules

| Column | Type | Reproducible rule |
| --- | --- | --- |
| `app` | text | Public product name on the official listing. |
| `pricing_model` | category | `Free + in-app purchases` when installation is free and paid purchases are listed; `Free + premium` when a free install and paid premium upgrade are public but no subscription-led model is stated; `Free tier + subscriptions` when an ongoing subscription tier is explicit; `Subscription-led` when subscription is the principal model; `Paid upfront` when purchase is required before installation. |
| `platforms` | list | Unique public platforms shown by official listings. iPhone and iPad remain separate display labels; Mac, Web, Desktop, and Android are recorded only when explicitly available. |
| `languages` | count or text | Published language count when an official listing exposes one; otherwise the exact public label such as `English` or `Multiple`. No language count is inferred. |
| `ai_help` | ordinal | `Strong` = at least three verified capability groups among recommendations, outfit feedback, generation, and photo analysis; `Moderate` = two groups; `Limited` = zero or one group. |
| `digital_wardrobe` | ordinal | `Strong` = clothing inventory, item metadata, and outfit linking are all publicly documented; `Moderate` = inventory plus one of those groups; `Limited` = a lighter or unclear closet layer. |
| `outfit_planner` | ordinal | `Strong` = saved outfits plus scheduling/calendar and situation-based planning; `Moderate` = two of those groups; `Limited` = zero or one group. |
| `closet_stats` | text | Exact type of publicly documented wardrobe analytics. `Basic tracking` is used only when activity is tracked without deeper cost, value, spend, or wear-rate reporting. |
| `visual_search` | boolean | `Yes` only when official public material documents photo-led item discovery or matching; general image upload or AI styling alone does not qualify. |
| `store_signal` | dated text | Public rating/count or first-party adoption claim captured as context, never converted into a quality score. |
| `source_url` | URL | Official listing or first-party product page used for the row. |

## Derived figures

All figures use the 11-row snapshot as the denominator:

- Platform coverage = number of rows whose normalized `platforms` list contains the platform. Result: iPhone/iPad 11, Android 3, Web/Mac/Desktop 2.
- Strong AI coverage = rows where `ai_help = Strong`. Result: 6.
- Cost/value statistics coverage = rows whose `closet_stats` explicitly contains cost per wear, spending, wardrobe value, or resale value. Result: 6.
- Visual-search coverage = rows where `visual_search = Yes`. Result: 1.

No weighting or composite winner score is used.

## Limitations

- The sample is curated, non-random, and intentionally limited to 11 visible category products.
- Public listings can omit capabilities available after onboarding or inside paid tiers.
- Prices, ratings, language counts, platform support, and product claims can change after the snapshot date and can vary by country.
- Ordinal depth labels measure documented capability breadth, not output accuracy, interface quality, or long-term user satisfaction.
- Beauty AI is both the dataset creator and one of the products in the sample. The published rules, row-level source URLs, dated snapshot, and absence of a composite winner score are intended to make that conflict inspectable.

## Change log

- `2026.05` (`2026-05-02`): initial 11-app data snapshot.
- `2026.08-docs` (`2026-08-08`): added the versioned snapshot, data dictionary, formulas, limitations, citation formats, and embeddable summaries. No row values were changed.

## Citation

APA:

> Beauty AI. (2026). *Wardrobe App Benchmark 2026* (Version 2026.05) [Data set]. https://beautyai.app/wardrobe-app-benchmark-2026

Markdown:

```md
[Beauty AI — Wardrobe App Benchmark 2026 (v2026.05)](https://beautyai.app/wardrobe-app-benchmark-2026)
```