CLADWELL CAPSULE WARDROBE REPORT 2026
Data release 2026-10-19.v1. Published October 19, 2026.
Snapshot 2026-10-20T00:57:56.096957+00:00 (October 19 in New York).
Source and full methods: https://cladwell.com/capsule-wardrobe-report#methodology

CITATION
Flynn, Erin. Cladwell Capsule Wardrobe Report 2026. Cladwell, October 19, 2026.
Data release 2026-10-19.v1. https://cladwell.com/capsule-wardrobe-report
Please credit Cladwell and link to the report when quoting findings or sharing charts.
Keep the population, date and counting notes with any reused chart or statistic.

POPULATION AND TIME
7,146 focused capsules from 3,984 active, nonstaff accounts, last-used July 21 to
before October 20, 2026 UTC. Excludes whole closets, official templates and
challenges. Clothing-and-shoe size analyses use 6,726 nonempty capsules from
3,587 accounts. Current memberships, no deletion timestamp, not in storage.
Logging window: July 21 through October 18, 2026, inclusive, UTC.
Account counts do not mean verified unique people. The app users represented
are not a representative sample of everyone who uses a capsule wardrobe.

FILES AND GRAINS
summary.csv: one summary metric per row. Size percentiles count clothing/shoes.
sizes.csv: one size band per row, including zero. Owners may overlap across bands.
owner-sizes.csv: one owner capsule-count band per row plus ALL. Each owner first
  contributes a median size; median_owner_median is the median of those medians.
composition.csv: one size band per row. Item memberships, not unique garments.
shared-pieces.csv: one aggregate row. Uses capsules with any current item,
  INCLUDING accessories. Its nonempty population differs from size analyses.
outfit-repeats.csv: one logged-date band per row, including zero-log capsules.
  Owners may overlap across bands. Percentages use each row's capsule denominator.
*.png / *.svg: five chart exports with populations and caveats in the image.
report-social.png: 1200x630 report preview graphic.
manifest.json: source and generated-file checksums, not a separate data snapshot.
All CSV rows have data_version and snapshot_at columns. No personal rows,
identifiers, customer photos or private account records are included.

COLUMN DEFINITIONS
eligible_capsules / eligible_owners: selected capsules / distinct account owners.
nonempty_capsules / nonempty_owners: selected capsules / account owners with
  at least one clothing or shoe membership. Accessories do not count here.
size_p25 / size_median / size_p75 / size_p90: clothing-and-shoe size percentiles.
size_band: current number of clothing and shoe pieces, using unequal bands;
  01-15 includes 1 through 15, 61+ has no upper bound. Not a density histogram.
capsules / owners: count of capsules / distinct owners within this row.
mean_clothes_shoes / median_clothes_shoes: per-capsule piece-count summaries.
mean_accessories: mean accessory memberships per capsule.
capsule_count: number of nonempty clothing-and-shoe capsules per account; ALL
  is the overall owner-weighted result, not an additional disjoint population.
median_within_owner_size_spread: median of per-owner max size minus min size.
tops / layers / bottoms / dresses / shoes / accessories: pooled current item
  memberships by catalog category. A piece in two capsules counts twice.
median_pieces: median clothing-and-shoe size in the row.
no_tops_layers / no_bottoms_dresses / no_shoes: capsules missing the stated
  category combination. These counts can overlap and should not be summed.
owners_with_nonempty_capsule / owners_with_2plus_nonempty_capsules: distinct
  accounts with one / at least two capsules containing any current item,
  including accessories; specific to the shared-membership table.
owners_reusing_any_item: those multi-capsule accounts sharing at least one item.
median_owner_item_reuse_share: median fraction (0 to 1, NOT percent) of an
  owner's unique items appearing in more than one selected capsule.
logged_day_band: number of different UTC dates with a matching-owner log.
capsules_with_repeated_outfit: capsules with the same saved outfit ID on
  multiple UTC dates. Identical clothes under different IDs are not a repeat.
raw_logs: calendar log rows before date-level deduplication.
distinct_outfit_days: distinct outfit/date pairs, counted within each capsule.
distinct_outfits_within_capsule: outfit IDs counted distinctly per capsule,
  then pooled across capsules; this is not a global unique-outfit count.
median_days / median_distinct_outfits: per-capsule logged-date / outfit medians.
repeat_percent: 100 * capsules_with_repeated_outfit / capsules in the row.

LIMITS
Current memberships and mutable last-used timestamps are not historical capsule
sizes. Calendar entries are not verified wears, and can be entered retrospectively.
No logs does not mean no wear. More logs create more opportunity to observe a
repeat. Nonstaff test accounts and copied templates cannot be fully identified.
Results do not establish an ideal capsule size, causality, savings, sustainability
impact, lifetime downloads or population-wide behavior. Source values retain
precision for reproducibility; extra decimal places do not imply measurement accuracy.
Analysis prepared by Cladwell from aggregate app records; report author: Erin Flynn.
Questions or corrections: hello@cladwell.com