Prepared by Yujie Cheng · Completed September 1, 2026
Analysis date: August 31, 2026. Built on 11 original research blocks and 347 deduplicated public sources, supplemented by three targeted public-record checks.
Original Word report (16 pages)
1. Core Judgment
ODDITY's customer acquisition engine broke in early 2026. The company's response reveals something more interesting: management was already building its next vehicle before the engine failed.
2. Key Data
Financial snapshot
| Metric | Q2 2025 | Q1 2026 | Change |
|---|---|---|---|
| Revenue | $241.1M | $197.9M | -17.9% vs. Q2 2025; -26% YoY (vs. Q1 2025: $268.1M) |
| Operating income | $57.1M | $(25.5)M | Swing from profit to loss |
| Gross margin | 72.3% | 69.7% | -2.6pp |
| IL MAKIAGE CPA index | 1.5 (H1 2025) | 2.8 (H1 through May 2026) | +83% YoY; 2.8x the H1 2022 baseline of 1.0 |
| First orders | baseline | ~-50% | Q1 2026 YoY, CEO-disclosed approximation |
| Repeat orders | baseline | ~-15% | Q1 2026 YoY, CEO-disclosed approximation |
| Cash, equivalents & investments | $815M (Jun 30, 2025) | $667.4M (Mar 31, 2026) | -$147.6M over nine months |
Note on comparisons: quarterly revenue is strongly seasonal (Q4 2025: $152.7M), so year-over-year comparisons are more meaningful than sequential ones. Q2 2025 is shown as the strongest recent quarter by operating income; it is not a sequential comparator.
Audience scale signals
| Brand | Trustpilot reviews* | Rating | Monthly visits (Similarweb est., Jul 2026) |
|---|---|---|---|
| IL MAKIAGE | ~97,000 | 4.3/5 | 2.4M |
| SpoiledChild | ~11,000 | 4.4/5 | 5.7M |
| METHODIQ | ~2,000 | 4.4/5 | Not shown on free tier (bounce rate: 71.8%) |
Snapshot at collection (Aug 2026); counts and ratings are dynamic.
Where the headcount is going
| Hiring direction | Open roles | Share |
|---|---|---|
| ODDITY Labs (Boston - peptides, drug discovery, bioengineering) | 5 | 45.5% |
| Parent company (engineering, finance/accounting, new ventures) | 5 | 45.5% |
| IL MAKIAGE (social media internship) | 1 | 9.1% |
| SpoiledChild | 0 | 0% |
| METHODIQ | 0 | 0% |
| Sales / BD | 0 | 0% |
Insider trading signals
| Signal | Data |
|---|---|
| CEO Form 4 filings (Aug 31, 2025-Aug 31, 2026) | 0 - no transaction-triggered filing found |
| CFO gross shares disposed | 192,735 shares: 130,567 sold in broker-assisted cashless option exercises; 62,168 sold for RSU tax withholding |
| Open-market purchases by any insider | 0 Form 4 transactions coded P |
| Largest single transaction | 100,000 shares sold at a weighted-average $15.09 ($1.509M gross), Jun 30, 2026, in a broker-assisted cashless option exercise |
3. Competitive Positioning
What ODDITY chose
| Choice | Evidence |
|---|---|
| 100% DTC, no offline retail | CEO stated no offline retail plans (Q4 2025 call); "we are 100% D2C" (Q1 2026 call) |
| Algorithmic acquisition over community | Meta Ad Library showed ~1,110 active US ads across the three brands' official pages on Aug 31, 2026 (IL MAKIAGE ~290, SpoiledChild ~400, METHODIQ ~420); near-zero organic brand advocacy found on Reddit/X |
| Technology narrative | Homepage, investor materials, and CEO remarks consistently framed around AI shade-matching, personalization, and the data platform |
| Subscription / auto-replenishment lock-in | Try Before You Buy (TBYB) pre-authorization -> auto-conversion to subscription; identical pattern across all three brands |
| Extension into prescription care | METHODIQ dispenses doxycycline and spironolactone; ODDITY Labs works on small molecules and peptides |
[Signal] Channel optionality, not a disclosed channel shift
On July 1, 2026 - 29 days after the Q1 2026 earnings call - ODDITY Tech filed a USPTO trademark application for IM PRO MAKEUP NY, covering IC 003 (cosmetics), IC 021 (cosmetic tools), and IC 035 (retail and wholesale store services). The IC 035 description states that all listed services may also be provided online.
The filing does not establish a physical-retail launch or a B2B/professional-channel launch. The "PRO" wording, combined with wholesale coverage, is consistent with - but does not prove - optionality toward a professional or wholesale channel. "NY" is part of the mark; no geographic strategy can be inferred from the filing. This does not contradict current D2C operations. It is a reserved right worth monitoring.
What ODDITY gave up
| Trade-off | Cost, now visible |
|---|---|
| Offline retail / omnichannel | 100% DTC concentrates exposure to ad-platform algorithm changes. The CEO first acknowledged this in the final Q&A round of the Q1 2026 call |
| Community and organic brand culture | No power-user base, no independent advocate community. When the paid funnel broke, no organic traffic cushioned the fall |
| Transparent pricing and simple checkout | The #1 complaint is identical across all three brands: unexpected charges, subscription lock-in, cancellation friction. This is by design - but the design may now be feeding back into acquisition signal quality (see Section 5) |
| US Google search ads | 0 of 6 tracked keywords showed any ODDITY brand in a US SERP session (Aug 2026). The only advertiser visible in Google's Transparency Center is an Israeli entity (IL MAKIAGE Beauty School). US acquisition appears Meta-dependent [inference - see Section 7] |
Peer comparison
| Dimension | ODDITY | Estee Lauder | Glossier | PROVEN | Curology |
|---|---|---|---|---|---|
| Channel | 100% DTC | Omnichannel | DTC -> retail | DTC + subscription | DTC -> Walmart |
| Acquisition | Paid Meta | Brand equity + retail + digital | Community + content + retail | Quiz + SEO | Quiz + paid + retail |
| Tech narrative | Strong (AI shade-match, data platform, Labs) | Weak | Weak | Strong (Skin Genome, AI) | Medium (algorithmic formulation) |
| Rx capability | METHODIQ | None | None | None | Core capability |
| Representative prices* | $15-$179 one-time | $52-$235 | $16-$116 | $34.99-$167.97 | $14 one-time; up to $59.90 per 30-60 days (Rx/custom) |
| Single-point-of-failure risk | Very high (one ad partner) | Low | Medium | High (small scale) | Medium |
Displayed one-time list prices from representative SKUs at collection, not exhaustive catalog ranges. ODDITY brands and Curology also offer subscription/auto-replenishment billing.
4. Statements vs. Actions: Five Contradictions
A. "We are not changing our model" - while changing it
What the CEO said:
"I want to be very clear, despite the dislocation in our new user acquisition we are currently facing, we are not changing our model, our strategy, or our long-term focus on growth."
- Q4 2025 earnings call
Note the structure of the sentence itself: the acknowledgment of an acquisition "dislocation" and the denial of any model change share a single breath.
What happened next:
-
40% of acquisition revenue moved from TBYB to standard purchase (Q1 2026)
-
Several "tech products / funnel offerings" were shut down during Q1 for troubleshooting
-
Channel diversification was acknowledged as under consideration for the first time (final Q&A round, Q1 2026 call)
TBYB is ODDITY's signature acquisition mechanic - "try first, return free if unsatisfied." Moving 40% of acquisition revenue off it is a model change by any reasonable definition. The defensive framing ("I want to be very clear") reads as an attempt to hold market confidence while the underlying approach shifts, and sell-side analysts noticed the gap.
B. "A technical problem, not a structural one" - but only ODDITY was hit
What the CEO said:
"[W]e don't believe that the problem sits on our end."
- Q1 2026 earnings call (excerpt)
What an analyst challenged (Anna Lizzul, Bank of America, final Q&A round): other beauty companies were not reporting an equivalent algorithm shock - why did the problem look concentrated in ODDITY?
What the data shows:
-
IL MAKIAGE's CPA index rose 83% YoY (1.5 in H1 2025 -> 2.8 in H1 through May 2026; the H1 2022 baseline was 1.0). Within the sources reviewed for this analysis, no other listed beauty DTC company disclosed a comparable shock in the same period. This screen was not exhaustive - it covers the earnings communications in the underlying dataset, and "not found" is not "confirmed absent."
-
ODDITY itself concedes that 100% DTC makes it more exposed to algorithm changes than omnichannel peers.
-
TBYB's pre-authorization mechanics plausibly generate elevated return and dispute rates, precisely the signals ad platforms use to score conversion quality.
If this were a platform-level fault, it should have hit the whole category. A shock concentrated in one advertiser suggests that advertiser's signals carry some feature the algorithm penalizes. The strongest candidate: TBYB's return rates and subscription complaints degrading the account's conversion-quality score.
C. Revenue is in beauty; hiring is in biotech
Essentially all FY2025 revenue came from beauty and personal care: IL MAKIAGE at roughly $560M and SpoiledChild at roughly $250M, against total net revenue of $809.8M. METHODIQ is guided to roughly $25M for 2026.
Meanwhile, 45.5% of open roles (5 of 11) sit in ODDITY Labs' biotech R&D - and the two brands generating the revenue have one opening between them, a social media internship.
Read against the $667M cash position, this looks less like misallocation than deliberate sequencing: current cash flow funding the foundation of a different company.
The trademark record makes the R&D direction more concrete. In the 18 months through August 31, 2026, ODDITY filed 20 new trademark applications. Nine are ODDL-series proprietary-ingredient names (including ODDL1893 and ODDL1105) covering goods in IC 001/003/005: chemical additives for manufacturing cosmetics, cosmetics, and medicated preparations. Five more use pharmaceutical-style names - LAZORAQ, SALORIL, LEXORIL, CLINDALAQ, and TRAZELAQ - for medicated acne-treatment preparations. CLARIXA covers dietary and nutritional supplements.
The signal is commercial rather than technical: ODDITY is building a trademark portfolio around proprietary ingredient names and medicated-product identities. The hiring data says the company is building biotech capacity; the trademark data shows it reserving names for potential outputs. Trademark protection covers the brand name, not the underlying molecule. Patent protection, if any, is not visible in this dataset. The filings also do not establish that the underlying compounds have reached commercial production.
D. Five-star and one-star reviews describe the same company
Positive reviews praise shade-match accuracy, product performance, and responsive support. Negative reviews describe unexpected charges, subscription lock-in, and cancellation friction.
These are consistent, not contradictory. The product works; the purchase and retention flow is the problem. Users like the foundation - they dislike discovering they were auto-subscribed. The moat under strain is not product quality but the cleanliness of the acquisition funnel.
E. METHODIQ's bounce rate contradicts its narrative
METHODIQ's bounce rate is 71.8%, against 48.9% for IL MAKIAGE and 44.3% for SpoiledChild (Similarweb estimates, Jul 2026). The company positions METHODIQ as a platform "disrupting the medical care system."
For a product that asks users to complete a health questionnaire, upload photos, and accept a prescription, about 72% of visitors bouncing indicates a break between ad promise and landing experience: wrong traffic, insufficient trust for a medical-grade interaction, or both.
5. The Structural Conflict: Retention Mechanics vs. Acquisition Economics
The central hypothesis of this analysis [inference]: the mechanisms ODDITY built to keep customers appear to be degrading its ability to acquire new ones.
How the retention engine works:
-
TBYB places a full-price pre-authorization -> user trials the product -> non-returns convert to purchases
-
Purchases default into subscription / auto-replenishment
-
Cancellation involves friction (a consistent theme across negative reviews)
-
Result: repeat sales made up roughly 70% of FY2025 revenue
The by-products this engine generates:
-
Full-price pre-authorizations that some banks flag as suspicious -> elevated dispute exposure
-
Unexpected charges -> complaints, chargebacks, negative reviews
-
Structurally high return rates - TBYB is designed on the assumption that many users will return items
-
A consistent negative-signal footprint across Trustpilot, Reddit, and app stores
How an ad platform reads those by-products:
-
Returns -> this advertiser's conversions don't hold
-
Chargebacks and disputes -> this advertiser's conversions are problematic
-
Negative social signals -> lower auction quality score
The chain:
Subscription lock-in -> high retention -> ~70% repeat revenue
Downstream: high returns + disputes + negative signals
Downstream: ad platform discounts conversion quality
Downstream: CPA index +83% YoY, first orders ~-50%
CEO: "the problem does not sit on our end"
But only ODDITY shows this profile in the reviewed sources - because only ODDITY's retention model produces this signal signature.
If this hypothesis holds, it explains four otherwise disconnected facts:
-
Why the CEO insists the problem is not structural - conceding it would mean the core of the business model (subscription lock-in) needs rework
-
Why 40% of acquisition revenue was already moved off TBYB - internal data may have surfaced the TBYB-CPA correlation before it became public
-
Why METHODIQ was under construction before the CPA crisis - management may have seen the ceiling on the existing model earlier than the market did
-
Why 5 of 11 open roles sit in ODDITY Labs' biotech R&D
Implications:
-
Near term: CPA may not "normalize" back to baseline, because the driver is a by-product of the model, not a platform bug
-
Medium term: continued TBYB reduction is the likely direction, at direct cost to first-order conversion
-
Long term: ODDITY's real bet is not beauty DTC; it is whether ODDITY Labs and METHODIQ can build a second revenue curve before the cash runs down
The data flywheel behind METHODIQ
Block 10 app-release logs show each METHODIQ update expanding medical data capture: weekly check-ins with skin images and treatment tracking, Care Team messaging (v2.2.0-v2.3.0), enhanced side-profile image capture (v2.5.0).
[Inference] The front end is telehealth; the back end is a longitudinal dermatology dataset feeding ODDITY Labs. This is also the most economical explanation for hiring peptide-discovery and bioengineering scientists while METHODIQ itself has zero open roles. METHODIQ's value is not its projected $25M revenue; it is the dataset being built underneath it.
Consistent with this: on August 31, 2026, METHODIQ - the smallest brand by revenue, with zero open roles - was running more active US Meta ad creatives (~420) than IL MAKIAGE (~290) or SpoiledChild (~400). Creative counts are not spend, but the allocation signal points the same direction: the hiring goes to Labs, the acquisition push goes to METHODIQ. Both point at the new vehicle.
Infrastructure signals
Certificate-transparency records show repeated certificate activity for IL MAKIAGE hostnames associated with three development labels: develop, develop2, and develop3.
[Inference] The web-quiz subdomain family was first observed in CT logs in July 2025 and appeared across development, pre-production, and production environments, a pattern consistent with development and staged deployment of a quiz-based web feature. It could relate to PowerMatch or to a separate questionnaire; CT data cannot distinguish between them. The webchat subdomain family was first observed in August 2025 in the same deployment pattern. The hostname is consistent with a chat layer but does not reveal whether it is AI-driven or human-supported. Certificate activity alone cannot confirm the specific product function behind either hostname.
qr.spoiledchild.com appeared repeatedly in CT logs, with 36 certificate records since February 2023. [Inference] The hostname is consistent with QR-linked web infrastructure, but certificate records do not establish its use case, traffic, placement on packaging, or distribution channel.
6. Implications for Competitors
Curology. METHODIQ is entering your category, prescription tele-dermatology, with a 71.8% bounce rate and zero open roles, meaning ODDITY is not yet investing in the trust infrastructure this category requires. The window: reinforce the dermatologist-founded narrative in paid social while ODDITY's "AI-driven" story is weakened by the CPA crisis, and bid on METHODIQ comparison keywords before the landing-page trust gap gets fixed.
Glossier. ODDITY's crisis is live market evidence that 100% DTC with no organic community is fragile under platform algorithm changes. The move into Sephora and owned retail is being validated in public, at no cost. No action required.
PROVEN. The Google Ads transparency check found no brand bidding on the six tracked US personalization keywords, including "personalized skincare quiz." With Meta CPAs rising for the category, this is an unusually cheap SEM/SEO opening for a smaller player that cannot win a paid-social bidding war against ODDITY's three-brand operation (~1,110 active US ad creatives as of Aug 31, 2026).
7. Methodology and Data Quality
This analysis is built entirely on public sources. The original research base consists of 11 blocks, 347 deduplicated source URLs, and a 2,365-line underlying dataset. Three supplemental public-record checks - USPTO trademarks, certificate-transparency logs, and SEC ownership filings - were added separately and are documented in Appendix A7. Every non-inferred figure carries a source URL in the original dataset or supplemental records and can be verified line by line.
A separate line-item verification pass (39 checks against the raw dataset and original filings/transcripts) was completed on August 31, 2026; corrections were applied wherever source support was insufficient. Meta Ad Library counts are additionally archived in a dated evidence pack - a counts CSV, a method note, and three page-level screenshots, all captured August 31, 2026. Load-bearing sources are listed in the Appendix below; see Data Availability at the end of the Appendix for what can be provided on request.
Claims marked [inference], with basis and falsification conditions:
1. TBYB return/dispute rates degraded Meta ad-account quality scores
Basis: only ODDITY reported a shock of this magnitude in reviewed sources; TBYB's design structurally produces returns; ODDITY itself is cutting TBYB exposure.
Not directly observable: ODDITY does not disclose return or dispute rates.
Falsified if: peers disclose comparable CPA shocks for the same period, or ODDITY's CPA normalizes without further TBYB reduction.
2. METHODIQ's data collection serves ODDITY Labs R&D
Basis: same corporate entity; successive app releases expand skin-image capture; Labs is hiring in drug discovery and bioengineering.
Not directly observable: no company disclosure explicitly links the two.
Falsified if: data-use disclosures show clinical data siloed away from Labs.
3. US acquisition is heavily Meta-dependent
Basis: 6 of 6 tracked keywords showed no ODDITY brand ads in a single US SERP session; Google's Transparency Center shows only an Israeli advertiser; Meta Ad Library showed ~1,110 active US ads across the three brands' official pages (Aug 31, 2026 snapshot).
Caveats: "no ads observed in one session" does not equal "zero Google spend." Ad Library figures are approximate, page-level creative counts - not spend - and the snapshot postdates the Q1 2026 funnel shutdowns, so it cannot describe pre-crisis behavior. Meta-dependence is plausible; any specific share of spend is unverified.
4. CEO Form 4 absence is consistent with no Form 4-triggering transaction after reporting began
Basis: EDGAR returned 12 ODDITY Form 4 filings through August 31, 2026; none was filed by CEO Oran Holtzman. Holtzman filed a Form 3 on March 18, 2026, confirming that he became subject to Section 16 reporting when the foreign-private-issuer reporting requirement took effect.
Limit: the absence of a filed Form 4 does not establish the absence of every trade or every potentially reportable transaction. The filing pattern is consistent with - but does not prove - a long-term holder who made no Form 4-triggering transaction after March 18, 2026.
Falsified if: a CEO Form 4 covering a transaction within the March 18-August 31, 2026 review window is later identified.
Additional scope note: the peer screen in Section 4B covers the earnings communications collected in the underlying dataset. It was not an exhaustive review of all listed DTC companies' disclosures.
Appendix - Load-Bearing Sources
Sourcing policy: the report body carries no URLs. This appendix lists the 71 source URLs that directly support the report's figures, quotes, and named claims, grouped by report section. The original underlying dataset - 11 research blocks and 347 deduplicated source URLs - is available on request; the supplemental public-record checks are documented in A7.
Access dates: original web pages and documents were collected August 29-30, 2026 (UTC) unless noted. Supplemental records were collected through August 31, 2026. Volatile sources - ad libraries, review counts and ratings, traffic estimates, job listings, trademark status, and certificate-transparency results - change continuously; each figure in the report is tied to its stated capture date. Meta Ad Library counts are additionally archived as a dated evidence pack (CSV, method note, page-level screenshots, captured August 31, 2026).
A1. Financial results and management statements
Supports: Key Data financial snapshot; Sections 4A-4C; all CEO quotations.
1. ODDITY Tech Q1 2026 results press release (GlobeNewswire, Jun 2, 2026) - Q1 2026 revenue, operating loss, gross margin, CPA index (2.8, +83% YoY), 40% acquisition-revenue shift off TBYB, cash position $667.4M, METHODIQ 2026 guidance $25M
2. Q1 2026 earnings call transcript (The Motley Fool) - first orders ~-50% / repeat orders ~-15%; "we don't believe that the problem sits on our end"; temporary funnel shutdowns; channel-diversification acknowledgment and Bank of America analyst exchange in final Q&A
3. Q1 2026 earnings call transcript, secondary copy (StockAnalysis)
https://stockanalysis.com/stocks/odd/transcripts/568524-q1-2026/
4. Q4 2025 earnings call transcript (Seeking Alpha) - "we are not changing our model, our strategy, or our long-term focus on growth"; no-offline-retail statement; FY2025 repeat sales ~70%; IL MAKIAGE ~$560M and SpoiledChild ~$250M FY2025 revenue
https://seekingalpha.com/article/4874828-oddity-tech-ltd-odd-q4-2025-earnings-call-transcript
5. Q4 2025 earnings call transcript, secondary copy (StockAnalysis)
https://stockanalysis.com/stocks/odd/transcripts/402471-q4-2025/
6. ODDITY Launches METHODIQ, a Telehealth Platform Reimagining Medical Care (ODDITY Investor Relations, Nov 18, 2025)
https://investors.oddity.com/node/7606/pdf
Note: Q2-Q4 2025 quarterly figures (revenue $241.1M / $147.9M / $152.7M, gross margins, operating income, cash of $815M / $793M / $776M, FY2025 net revenue $809.8M) are compiled from the corresponding quarterly releases and call transcripts in the underlying dataset (Block 7).
A2. Product, pricing, and conversion mechanics
Supports: Section 3 (choices); Section 5 (retention engine).
7. IL MAKIAGE Try Before You Buy - mechanics, 14-day window, pre-authorization
https://www.ilmakiage.com/tutorial/5d112c0e0e91b4001bd254f0/try
8. IL MAKIAGE Terms of Use
https://www.ilmakiage.com/termsofuse/
9. SpoiledChild Auto-Refill FAQ - subscription default, pause/cancel terms
https://www.spoiledchild.com/faq/category/auto-refill-subscriptions
10. METHODIQ Doxycycline - prescription oral antibiotic, $59
https://www.methodiq.com/doxycycline
11. METHODIQ Spironolactone - oral tablet for hormonal acne, $59
https://www.methodiq.com/spironolactone
A3. Audience, traffic, and review signals
Supports: Key Data audience table; Section 4D; Section 4E bounce rates. Review counts, ratings, and traffic estimates are snapshots (Aug 2026; Similarweb figures are July 2026 estimates).
12. Trustpilot - IL MAKIAGE (~97,000 reviews, 4.3/5 at snapshot)
https://nz.trustpilot.com/review/ilmakiage.com
13. Trustpilot - SpoiledChild (~11,000 reviews, 4.4/5 at snapshot)
https://nz.trustpilot.com/review/spoiledchild.com?page=2
14. Trustpilot - METHODIQ (~2,000 reviews, 4.4/5 at snapshot)
https://nz.trustpilot.com/review/methodiq.com?page=4
https://www.trustpilot.com/review/methodiq.com?page=8
15. Similarweb - ilmakiage.com (visits, bounce rate)
https://www.similarweb.com/website/ilmakiage.com/
16. Similarweb - spoiledchild.com (visits, bounce rate)
https://www.similarweb.com/website/spoiledchild.com/
17. Similarweb - methodiq.com (bounce rate 71.8%; visit volume not shown on free tier)
https://www.similarweb.com/website/methodiq.com/competitors/
18. Apple App Store - METHODIQ Health - release notes v2.2.0-v2.5.0 (data-capture expansion) and user reviews
https://apps.apple.com/us/app/methodiq-health/id6739997374
A4. Hiring signals
Supports: Key Data headcount table; Sections 4C and 5. Snapshot 2026-08-30; listings change continuously.
19. LinkedIn Jobs - ODDITY entities filter (11 open roles at snapshot; role-level breakdown in dataset)
20. ODDITY Labs careers page - displayed three mutually inconsistent states on one page at collection; recorded as a conflict entry only, not used for counts
https://www.odditylabs.ai/careers
A5. Advertising footprint
Supports: Section 3 (algorithmic acquisition); Section 6; inference #3 in Section 7.
21. Meta Ad Library - IL MAKIAGE official Page, US, All ads, Active: ~290 results (captured 2026-08-31)
22. Meta Ad Library - SpoiledChild official Page, same scope: ~400 results (captured 2026-08-31)
23. Meta Ad Library - METHODIQ official Page, same scope: ~420 results (captured 2026-08-31)
Counts are Meta-displayed approximations at Page level; they are creative counts, not spend. Dated evidence pack (CSV, method note, three screenshots) archived and available on request.
24. Meta Ad Library - individual creative records cited in the dataset
https://www.facebook.com/ads/library/?id=25916613107935334
https://www.facebook.com/ads/library/?id=2114300509442319
https://www.facebook.com/ads/library/?id=1287997406517432
https://www.facebook.com/ads/library/?id=941912721717440
25. Google Ads Transparency Center - advertiser AR03841466509578731521 ("IL MAKIAGE Beauty School," Israel), the only ODDITY-related advertiser found; representative creative records (seven further creative-level URLs in dataset)
26. Google US SERP session - six tracked keywords, logged-out, hl=en / gl=us / pws=0, 2026-08-30 UTC; no ODDITY brand ads observed in this session (single-session observation; not evidence of zero Google spend)
https://www.google.com/search?hl=en&gl=us&pws=0&q=online%20foundation%20shade%20match
https://www.google.com/search?hl=en&gl=us&pws=0&q=foundation%20for%20acne%20prone%20skin
https://www.google.com/search?hl=en&gl=us&pws=0&q=personalized%20skincare%20quiz
https://www.google.com/search?hl=en&gl=us&pws=0&q=IL%20Makiage%20alternative
https://www.google.com/search?hl=en&gl=us&pws=0&q=best%20foundation%20for%20mature%20skin
https://www.google.com/search?hl=en&gl=us&pws=0&q=online%20acne%20treatment%20with%20dermatologist
A6. Peer comparison
Supports: Section 3 peer table; Section 6. Representative displayed prices at collection, not exhaustive catalogs.
27. Estee Lauder - representative product pages
28. Glossier - site and representative product pages
https://www.glossier.com/pages/about
https://www.glossier.com/products/balm-dotcom
https://www.glossier.com/products/boy-brow
https://www.glossier.com/products/cloud-paint
https://www.glossier.com/products/futuredew
https://www.glossier.com/products/glossier-you
29. PROVEN - positioning, subscription model, catalog
https://www.provenskincare.com/
https://www.provenskincare.com/why-proven/
https://www.provenskincare.com/personalized-subscription/
https://www.provenskincare.com/acne-treatment/
https://release.provenskincare.com/collections/all-skincare-products/
30. Curology - positioning, pricing, formulation narrative
https://support.curology.com/en_us/pricing-and-products-BJOLQnd2O
https://curology.com/blog/custom-formula-skin-concerns-acne-treatment/
31. LinkedIn company pages - peer scale reference
https://www.linkedin.com/company/glossier
https://www.linkedin.com/company/provenskincare
https://www.linkedin.com/company/curology
A7. Trademark, certificate-transparency, and insider-transaction signals
Supports: Section 3 channel-optionality signal; Section 4C trademark corroboration; Key Data insider-trading table; Section 5 infrastructure signals; inference #4 in Section 7.
32. USPTO Trademark Search - owner/applicant queries - "ODDITY Tech Ltd" (44 results), "IL MAKIAGE" (26), "SpoiledChild" (5), and "ODDITY Corp" (0); 72 unique serial numbers after deduplication. USPTO owner-query state is session-based and is not preserved in the URL. Individual trademark records are accessible through the TSDR direct links in entry 33 and the underlying dataset.
33. USPTO TSDR - representative individual trademark records - 20 new applications filed March 1, 2025 through August 31, 2026; each record's direct TSDR URL is retained in the supplemental dataset
IM PRO MAKEUP NY: Serial 99916291 (filed July 1, 2026; Pending at the report cutoff)
CLARIXA: Serial 99464300 (filed October 27, 2025; registered August 18, 2026)
LAZORAQ: Serial 99495003 (filed November 13, 2025; Pending at the August 31 cutoff; registered September 1, 2026)
34. crt.sh Certificate Transparency logs - subdomain enumeration for oddity.tech, ilmakiage.com, and spoiledchild.com (captured August 31, 2026: 2,341 raw rows; 1,493 distinct certificates; 145 unique subdomains after deduplication)
https://crt.sh/?q=%25.oddity.tech
https://crt.sh/?q=%25.ilmakiage.com
https://crt.sh/?q=%25.spoiledchild.com
Availability note: all three crt.sh queries returned HTTP 502 during verification on September 2, 2026. Query results were captured August 31, 2026; the raw data is preserved in the supplemental dataset.
35. SEC EDGAR - ODDITY Tech Ltd ownership filings, CIK 0001907085 - Form 4 review period: August 31, 2025 through August 31, 2026; CEO Form 3 filed March 18, 2026
https://www.sec.gov/Archives/edgar/data/1907085/000110465926031530/0001104659-26-031530-index.htm
36. SEC - Holding Foreign Insiders Accountable Act: Section 16(a) Reporting Requirements - foreign-private-issuer officers and directors became subject to Section 16(a) reporting effective March 18, 2026; published jurisdiction exemptions do not include Israel
Data availability
This appendix lists the 71 load-bearing source URLs (36 entries above). The following materials are available on request:
-
Full original research dataset - 11 research blocks, 2,365 lines, 347 deduplicated source URLs; every non-inferred figure in the original report ties to a row-level source in this dataset
-
Supplemental public-record checks - USPTO trademark records, certificate-transparency exports, and SEC ownership filings used for the additions in this revision
-
Meta Ad Library evidence pack - counts CSV, method note, and three page-level screenshots documenting the ~290 / ~400 / ~420 active-ad counts, captured August 31, 2026
-
Block-level research files - per-topic collection records underlying the dataset
Prepared by Yujie Cheng - August 2026