The work,
with the evidence.

What we audited

All public pagination was exhausted across the Videos, Shorts and Live tabs: 3,830 unique video IDs — 3,473 standard videos, 319 Shorts and 38 live streams. This matches the channel’s public video count at capture. Private, deleted and unlisted uploads are outside this public snapshot.

Titles were audited for all 3,830 entries. Exact publication dates are available for 337 and full public descriptions for 336, including one older sponsorship video checked directly in the browser. YouTube rate-limited further automated description requests (HTTP 429). The first 300 standard uploads have uninterrupted date coverage, extending to October 2024, which encloses the entire annual reporting window.

How the four figures work

1. 85+ brands in 12 months: 85 distinct normalized brands identified across 151 standard uploads published September 1, 2025 through August 31, 2026. Of those, 146 have identified brands and five remain unclassified. The plus sign marks a conservative identified count.
2. 3,261 brand-coverage videos: count of standard videos with at least one recognized brand in the title or a manually reviewed description. Videos are counted once even when several brands appear. This includes product reviews, setup tutorials, comparisons, accessory compatibility and event/show-floor coverage. It is not a count of sponsored videos.
3. 15+ years of product coverage: anchored to Installation Of Rear Door Sills Lexus Ct 200 H, published April 8, 2011, verified on the public video page. This demonstrates product coverage by 2011; it does not establish the start of paid partnerships.
4. 40.3% repeat brand coverage: 279 brands with at least two retained videos ÷ 692 identified brands × 100, rounded to one decimal. The scope is the audited standard-video history. This is repeat coverage, not a return-sponsor rate.

Normalization and deduplication

Case-insensitive whole-name matching uses the supplied alias dictionary, followed by documented per-video corrections. Examples include xTool / X-Tool / X Tool, Haotian / Hoatian, Sharge / Shargeek, Poly / Plantronics, and VRS Design / Verus. Product families are mapped where distinctive (for example iPhone to Apple). Distinct consumer sub-brands such as Anker, soundcore, eufy and eufyMake remain separate; these are not distinct corporate-client counts. The previously supplied partner list does not mark individual videos as sponsored.

Brand names are matched in titles rather than description boilerplate, so recurring camera-gear lists and affiliate footers do not inflate totals. Twenty documented description-based additions resolve recent videos with generic titles. Comparisons and named device compatibility count as brand coverage; a title mention alone is not proof of hands-on testing or a commercial relationship. Unknown and ambiguous products are left unclassified.

The calculation excludes every Shorts-tab and Live-tab entry, one standard upload tagged #short, and five duplicate normalized titles. That leaves 3,467 standard uploads, of which 206 have no identified brand. Duplicate titles retain the older entry. This conservative title rule may merge distinct edits with the same title. Renamed reposts and unmarked clips cannot be conclusively detected without content fingerprinting; no claim is made that these are distinct campaign contracts.

What the disclosures establish

One inspected video explicitly identifies a sponsorship: Sponsored Review: Poly Voyager Free 60+ UC True Wireless Earbuds for Work and Play, October 18, 2023. Its title, description and #ad / #sponsored labels agree. This is a confirmed minimum, not a lifetime total or the earliest possible sponsorship.

Among the 336 retrieved descriptions, 332 say the video is not sponsored and 329 include a product-provided disclosure. Those categories overlap. Three videos carry YouTube’s “Includes paid promotion” marker while also saying they are not sponsored: Anycubic P1 Max, Anycubic Kobra S1 Combo and Oleap Archer. Their commercial terms are not inferred. Samsung First Look 2026 uses #SamsungPartner; payment terms are likewise unspecified.

Affiliate links and products supplied for review are recorded separately from explicit sponsorship. This partial description sample cannot establish a representative repeat-sponsor rate, a lifetime number of sponsors, or distinct campaign contracts. No missing years are extrapolated.

Inspect or reproduce the audit

Public catalog snapshot (JSON) ↗ · Every classification and exclusion (JSON) ↗ · Brand counts and all evidence IDs (JSON) ↗ · Summary figures (JSON) ↗

Brand aliases ↗ · Reviewed corrections ↗ · Recalculation script ↗

To recalculate offline, keep the downloaded files together and run python3 analyze.py. No credentials or additional packages are required. Counts are a dated snapshot and do not change with the homepage’s recent-upload feed.

Brand evidence

Counts below exclude Shorts, streams and duplicate titles. “12 months” means September 2025–August 2026. Two source links are shown where available; the downloadable file includes every matching video ID.