History

History — A Founder Who Built It, Broke It, and Spent Five Years Rebuilding It

Yatsen IPO'd in November 2020 as China's hottest digitally-native beauty story — Perfect Diary's "disruptive DTC" model, eight brands, gross sales up 363% in 18 months [1] [2]. Within two quarters the founder was conceding on a public earnings call that the team had "moved too fast" [3]. Five years later the company has lost ¥3.0 billion to operating losses, written down its prestige skincare M&A twice, executed a de facto 5-for-1 reverse split to stay listed on the NYSE, and — only in fiscal 2025 — finally turned a non-GAAP profit of ¥8.4 million on ¥4.3 billion of revenue [4]. The story this tab tells is not whether the business is good (the answer is "increasingly"), but whether the founder who built it and then broke it has now earned the right to be trusted to run it for compounders. The verdict — narrowly, conditionally, on probation — is yes.

All financial figures are in renminbi (¥) — Yatsen's reporting currency. Share-price and US$ buyback figures are stated in US dollars. Per-ADS amounts after March 2024 reflect a 1-for-5 ADS-ratio change.

The arc at a glance

FY2025 Revenue (¥M)

4,298

YoY Growth

26.7%

FY2025 Gross Margin

78.2%

FY2025 Non-GAAP Net Income (¥M)

The headline is jagged: revenue grew from ¥5.2B in 2020 to ¥5.8B in 2021, then collapsed 36.5% to ¥3.7B in 2022 [5], and floor-skipped sideways through 2023 (¥3.4B) and 2024 (¥3.4B) before snapping back to ¥4.3B in 2025 [6]. The composition tells you why — color cosmetics (the Perfect Diary cash cow) was cut in half between 2021 and 2022 and never came back, while clinical/premium skincare (Galénic, DR.WU, Eve Lom) climbed from ¥855M to ¥2.28B over the same five years and is now the majority of revenue.

The five years are best read as five chapters. The IPO peak in late 2020; the crack that opened in Q2 2021 within weeks of going public; the collapse in 2022; the broken promise in 2023; the bottom in 2024; the reset in 2025; and the wobble in Q1 2026 that is the open question today.

Chapter 1: The IPO peak — a digitally native DTC darling (Nov 2020)

The original promise, set down in the November 2020 prospectus, was a single coherent thesis: Yatsen had cracked a "disruptive DTC business model" that would let a founder-led platform build "a family of iconic beauty brands" by combining proprietary content, data, and a young-consumer-first product engine [7]. The numbers behind that promise looked extraordinary: gross sales grew from ¥757M in 2018 to ¥3.5B in 2019, a 363.7% increase "or roughly 30 times the growth rate of the broader China beauty industry" [1]. The company priced at $10.50 per ADS on the NYSE on November 19, 2020 and ran into the mid-$20s within months — a market cap above $13 billion at the peak. Color cosmetics was 94% of gross sales in 2020 and Perfect Diary was the hero brand carrying the model [8].

Two things are worth noting about that opening chapter. First, founder Jinfeng Huang built the business himself — he founded Yatsen in 2016, has held the chairman-and-CEO role since inception, and remains in both seats today [9]. There is no inherited quality story here. Whatever the business is, he built it. Second, the IPO prospectus described the company as "China's leading beauty group" leveraging a "digitally native direct-to-customer business model" [8] — language that, four years later, has been quietly cut from the opening of the annual report.

Chapter 2: The crack — moves too fast, then walks them back (mid-2021)

The story bent in Q2 FY2021 — the second public earnings call. Listening to that call now, in light of what came after, the pivot is naked. Revenue was up 53.5% YoY; on the surface a triumph. But inside the script Huang acknowledged that the team had "launched in May" a set of "strategic growth initiatives" to "sharpen our growth model" — translation: management was already, three months after the IPO, restructuring around a recognized problem [10]. When pressed, he conceded directly: "previously, we did move too fast to reallocate our talent into the skincare BU" [3]. On the same call he introduced the framing that would carry the entire next five years — "the path of high-quality, sustainable growth" [10] — and quietly retired the "DTC" language that had carried the IPO.

By Q3 FY2021 (November 2021), Q3 revenue had decelerated to +6% YoY (from +53% the prior quarter). On that call management did three things that signal a corporate story shifting from offense to defense:

Replaced the "premiumize Perfect Diary" headline with the new phrase "a near path to profitability in the medium to long run" [11] — a multi-year hedge.
Announced the company's first share repurchase program — up to US$100 million over 24 months, authorized November 17, 2021 [12].
Asked investors to wait: "investors need to be patient about what we are doing right now" [13].

The stock closed 2021 at $10.75, down 88% from its February peak (both numbers split-adjusted). The "patience" request would eventually take five years to pay off — and as of mid-2026 has not yet been fully repaid in the share price.

Chapter 3: The collapse and the promise that broke (2022–2023)

The 2022 collapse was structural, not COVID. Revenue fell -36.5% YoY (¥5.84B → ¥3.71B) [5], with the color-cosmetics segment cratering from ¥4.87B to ¥2.42B in one year. The zero-COVID policy was a contributor but not the cause — Chinese color cosmetics had simply rotated out of Yatsen's hand as Douyin live-streaming changed the unit economics of building a mass color brand [14]. On the Q4 FY2021 call (March 2022), with the worst quarter still ahead, the founder set down what would become the most consequential — and most broken — public promise of the era:

"…we are expecting to break even on a yearly basis next year. So that is our current goal." [15]

He reaffirmed the same FY2023 break-even target on the Q1 FY2022 call in May 2022, hedged this time on COVID [16]. 2023 came and went. The full-year operating loss for FY2023 was ¥559M; the full-year non-GAAP net loss was approximately ¥296M. The promise was missed by roughly a billion renminbi.

Meanwhile management built the new narrative scaffolding under the wreckage. The "five-year strategic transformation plan" was launched at the start of 2022 and embedded as the official frame in that year's 20-F [17]. The first full-year ESG report was issued in May 2022, the MSCI rating was upgraded to "A," and 47% of the workforce was cut — headcount fell from 3,497 at end-2021 [18] to 1,837 at end-2022 [19]. The board upsized the share repurchase program from $100M to $150M in August 2022 and a NYSE minimum-bid deficiency notice received April 11, 2022 was cured by August 1 of the same year [20].

Then 2023 produced its own concessions:

The first Eve Lom goodwill impairment of ¥354 million was booked in Q4 FY2023 and disclosed on the March 2024 earnings call. Management's own words explained the underlying admission cleanly:

"…amount by which the carrying value of the Eve Lom reporting unit exceeded its fair value based on quantitative goodwill impairment test due to weaker operating results than expected at the time of acquisition." [21]

Eve Lom had been acquired for roughly ¥1.28 billion in March 2021 — including ¥965M of cash consideration and ¥569M of identifiable intangibles [22]. The impairment was the first explicit on-the-record concession that the prestige-skincare M&A strategy had been overpaid for. Same year, Yatsen "strategically phased out" Abby's Choice, its home-grown 2020 mass-skincare brand [23]. New chief scientific officer Jing Cheng — a 17-year Estée Lauder veteran — joined in early 2023 to lead a re-platforming of the company around clinical R&D [24], and the company moved the buyback authorization from $150M to $200M with the term extended through November 19, 2025 [25]. The buyback was no longer an opportunistic capital-return signal; it had become an existential mechanism to keep the share price off the NYSE delisting threshold while the operating business healed.

Chapter 4: The bottom — a reverse split, a second impairment (2024)

By 2023 year-end the share price was at $3.66 (split-adjusted). On November 2, 2023 the NYSE delivered a second non-compliance letter (the prior April 2022 letter had been cured). Management's response in March 2024 was a 1-for-5 ADS ratio change — converting one ADS from four Class A ordinary shares to twenty, which functioned as a 5-for-1 reverse split and immediately pushed the bid price above $1 [26]. The NYSE confirmed compliance on April 10, 2024.

Inside the operating business, 2024 was a holding pattern with two ugly disclosures:

A second goodwill impairment of ¥403.1 million — ¥397.8M against the Eve Lom reporting unit (again) plus a ¥5.3M write-down on a color-cosmetics unit (likely Pink Bear) [27].
Headcount fell to 1,350 — the low of the cycle — from 3,497 at the end of 2021 [28].

Despite the impairment, this was also when the "story" began visibly converging with the cash flows. The Yatsen Global Innovation R&D Center — described by management as "China's first global R&D hub for a national beauty brand" — opened in Shanghai in May 2024 [29]. Q4 FY2024 delivered the first non-GAAP profitable quarter of the rebuild [30]. The team had stopped retreating.

Chapter 5: The reset — skincare crosses the line (2025)

If the story has a vindication chapter, this is it. 2025 was the year management's five-year plan moved from PowerPoint to P&L.

FY2025 Revenue (¥M)

4,298

Gross Margin

78.2%

Skincare Share

53.0%

Revenue grew +26.7% to ¥4.30 billion — the first growth year since 2021 [31]. Skincare revenue grew +63% YoY to ¥2.28 billion and crossed 53.0% of total revenue — clearing the "skincare to exceed at least 50%" target Huang had set on the Q4 FY2022 call [32]. Gross margin reached 78.2%, up from 68.0% in 2022 [33]. Most importantly, the full year delivered ¥8.4 million of non-GAAP net income — a margin of 0.2% but, in absolute terms, the first profitable full year on a non-GAAP basis since the IPO [4].

And the opening line of the FY2025 annual report — the most carefully-edited sentence in any company's disclosure — was rewritten. It now reads:

"Yatsen is a leading China-based beauty group with the vision of becoming a world-class pioneer in beauty innovation." [34]

The "digitally native DTC" framing of 2021 is gone. The strategy section was re-anchored around three explicit pillars, the first of which is now "Driving R&D-led product innovation" — with the disclosure that R&D has consistently been ≥3.0% of revenue since 2022 [35]. Perfect Diary, which had been the entire company in 2020, now occupies a paragraph under a new doctrine called "makeup skintification" [36] — color cosmetics is no longer the headline brand, it is one application of the skincare platform.

Chapter 6: The wobble — and the open question (Q1 FY2026)

The Q4 FY2025 call introduced one new word for the first time. On a March 2026 earnings call, six years into the AI cycle:

"…application of AI in areas such as molecular structure prediction." [37]

The lateness of that introduction is itself a tell. Yatsen had described itself as "AI" and "big-data" enabled in the IPO prospectus four years earlier, then dropped that language entirely during the rebuild, and is now bringing it back exactly as the R&D platform reaches scale. A skeptic would read this as marketing the recovery for the next AI-themed multiple. A constructive reader would read it as a CSO-led R&D function finally having something AI-adjacent worth talking about.

The Q1 FY2026 call (May 2026) delivered the first regression of the new chapter. Revenue grew 22.5% YoY (in-line with guidance) but operating expenses ran ahead of revenue — 89.9% of net revenues, up from 83.2% a year ago [38] — and the company reversed back into a ¥57.3M non-GAAP net loss (5.6% margin), undoing in one quarter the ¥7.1M non-GAAP profit that had been the celebrated milestone of Q1 FY2025 [39]. Management blamed "elevated industry-wide [traffic] acquisition costs on the [Douyin] platform" [40] — the same channel mechanic that broke color cosmetics in 2021–2022, now back biting skincare.

Two weeks later (May 21, 2026) the company closed its first dilutive financing of the historical period — a private placement of convertible notes and warrants [41]. This was the same week the new (and much smaller) $30M repurchase program — authorized in May 2025 [42] — was already burning through capital intended to support the share price. The combination — burning buyback capital while raising new dilutive capital — is hard to reconcile cleanly. The most generous reading is that the founder wanted to add a strategic capital partner to support the rebuild; the less generous reading is that the company needed liquidity and didn't want to say so.

Promises made and promises kept

Across five years of public commentary, the founder made a finite set of public, valuation-relevant commitments. The track record:

No Results

Of ten material commitments, seven were kept (some convincingly beaten), two were clearly broken (the FY2023 break-even promise and the mid-2024 guidance that needed an in-quarter revision), and one is on probation after a single quarter's reversal. Two observations matter more than the score:

Both broken promises were quantitative and public. Management did not try to spin them away — the FY2023 break-even target was simply not mentioned again after Q2 FY2022, and the mid-July 2024 guidance revision was disclosed cleanly on the next call [43]. That is structurally honest behavior. Compare with management teams that simply quietly drop the target and never acknowledge it — Yatsen does not do that.
The buyback was actually completed. Yatsen authorized a $200M program and deployed $202.2M against 40.2 million ADSs through February 2026 [44]. Of total IPO net proceeds, US$499.3M went to share repurchases by year-end 2025 [45] — a figure most public-tech management teams would have left in cash and quietly written press releases about. That is a meaningful credibility signal, especially given the share price never recovered. It also means the buyback was, in aggregate, a destruction of capital at IPO prices — but a real return of capital to shareholders at the prices at which it was actually executed.

Narrative drift — what management stopped saying

The phrases that vanished from the script are as informative as the phrases that arrived. Three are worth pinning:

"DTC" / "digitally native" carried the IPO and was the single most prominent phrase of the prospectus [7]. It vanishes from the script after the Q2 FY2021 call and is no longer in the opening line of the FY2025 annual report. The thesis that justified the IPO multiple is gone.
"Premiumize Perfect Diary" was the central tactic of Q2 FY2021 [46]; by Q3 FY2021 it had been replaced by the buyback announcement. The premiumization tactic never produced a Perfect Diary that could carry the company on its own.
"Break even on a yearly basis next year" [15] was repeated in May 2022 and then dropped. It is the single most important phrase management quietly stopped saying.

Credibility — the verdict

Management credibility score (1–10)

A 5 out of 10 — with the arrow pointing up.

The case for a lower score is straightforward and was real for three years. Founder built the business, built it on a single channel and a single brand, missed the most important quantitative promise he ever made by a billion renminbi, paid roughly ¥1.28 billion for Eve Lom in 2021 and then wrote ¥752 million of that goodwill off in two tranches (2023 and 2024), executed a 5-for-1 reverse split to stay listed, and presided over a 96% drawdown in the share price between February 2021 and March 2024. At the end of FY2023 a fair score for this team would have been a 3.

The case for the arrow pointing up is the FY2025 reset. Every quarterly guide from Q3 FY2023 forward has been met or beaten, sometimes by a large margin (Q2 FY2025 came in at +36.8% against a +2-12% guide). The skincare-majority promise from the Q4 FY2022 call was kept. The non-GAAP profitability target was hit. The buyback was actually executed in full — at value-destructive IPO-era prices, yes, but executed. The Eve Lom impairments were taken honestly when the operating results required them, not delayed by another year of optimism. The CSO from Estée Lauder was hired before the recovery and there is genuine R&D infrastructure behind the "innovation" narrative now (CNAS-accredited lab, Cosmax JV, Hôpital Saint-Louis partnership). The founder did break the business and did then rebuild it — not by waiting for the macro to turn but by amputating brands, halving the workforce, and re-platforming around skincare.

What holds the score at 5 and not 6 are three things:

The Q1 FY2026 reversal into a non-GAAP loss followed by a dilutive convertible-note raise, in the same week the buyback was still drawing down capital, is a credibility-fragile event.
The AI vocabulary added to the Q4 FY2025 narrative is six years late — it reads more like multiple-management than R&D substance.
The chairman and CEO roles remain combined in the founder; an independent chair would meaningfully change the calculus of trusting this team with the next decade of capital allocation, given the post-IPO M&A record.

What the story is now

The story today is no longer "China's leading DTC color cosmetics brand." It is a clinical/premium skincare platform — Galénic + DR.WU + Eve Lom — with Perfect Diary as a profitable but secondary application of that platform under the "makeup skintification" framing. The five-year transformation is over and management is now arguing it has begun the next chapter: an R&D-driven, profit-disciplined, multi-brand operator in the prestige and clinical tier of the Chinese (and increasingly international) beauty market.

A reader should believe the structural transformation is real. The numbers — gross margin 68% → 78%, skincare share 16% → 53%, headcount 3,497 → 1,350 low in 2024 then rebuilt to 1,623 in 2025 [47], R&D infrastructure stood up — are too consistent to be a story. A reader should discount the AI vocabulary and the "world-class pioneer" rhetoric as positioning for the next equity raise rather than as substance. And a reader should watch Q2 FY2026 closely — if the non-GAAP profitability comes back, the story has earned the right to be told as a successful rebuild; if it does not, the convertible note raise will be re-read in retrospect as the moment the rebuild stalled.

The single most important fact about Yatsen in mid-2026 is that the founder who built it and broke it is still running it, and is starting — for the first time since 2021 — to look like he is doing it well. Whether that lasts is the question this tab cannot answer; the answer is in the next four quarters.

References

Generated Jun 19, 2026 10:16 UTC