Better, Faster than Your RA
3 minutes after the call
US channel strategy for K-beauty entry
Amazon-first velocity, then Ulta before Sephora. Margins reach parity once CAC is loaded.
Key Takeaways
- Private-label pressure is concentrated in mass, not prestige. Prestige margins have held for 3+ years.
- Amazon is the discovery channel now. 60–70% of new brand volume starts there before any retail listing.
- Retail media spend is table stakes. 8–12% of gross sales to hold shelf position.
- The DTC margin advantage is overstated. Load CAC and contribution lands at Amazon parity below ~$50M.
- Korean brands win on speed. 6-month concept-to-shelf vs 18 months for US incumbents.
Market Structure
- US skincare is roughly $21B at retail, and the growth is lopsided toward prestige.
- Prestige is compounding at 9% YoY while mass sits at 2%.
- The top-4 retailers control about 55% of prestige distribution between them.
- Amazon’s share of total beauty went from 18% to 27% in three years.
- The shift is driven by brand discovery, not price shopping.
Channel Economics
- Amazon 3P nets around 25% after fees and the ad spend they effectively require.
- Take rates are non-negotiable below $10M GMV.
- Terms only start to open up at 1P vendor status.
Channel Strategy
- How should we prioritize channels for a new entrant?
- Amazon-first for 12 months to build velocity data, then leverage that into Ulta before Sephora.
- How much leverage do brands have on take rates?
- Effectively none below $10M GMV. Negotiation only opens at 1P vendor status.
- What kills brands in year one?
- Inventory, not demand. Stockouts during a viral moment, not weak sell-through.
Figures Mentioned
| Metric | Value | |
|---|---|---|
| US skincare retail market | $21B | |
| Amazon share of beauty, 3-year change | 18% → 27% | |
| Retail media to hold shelf position | 8–12% of gross | |
| Amazon 3P net margin | ~25% | |
| Concept-to-shelf, KR vs US | 6 mo vs 18 mo |
Deck-Ready Quotes
- “Ulta says yes to data, Sephora says yes to brands Ulta proved.”
- “You take the standard terms and make the math work on product cost.”
- “We lose to ‘we’ll do it ourselves’ more than we lose to vendors.”
On Amazon 3P you’re at maybe 25 points net once you pay fees and the ad spend they basically require now.
And how do DTC margins actually compare at mid scale?
Honestly people overstate the DTC margin thing. Once you load in customer acquisition you’re basically at Amazon parity until you hit real scale.
Concept to shelf for the Korean brands is six months. The US incumbents plan in eighteen.
Ulta says yes to data, Sephora says yes to brands Ulta proved.
Your questionnaire
4 questions prepared1What drives churn in enterprise accounts?
Integration debt, not price. Renewals fail when IT owns the contract and the product sits outside daily workflows.14:22
2Who is the real competitor in a deal?
In-house builds. “We lose to ‘we’ll do it ourselves’ more than we lose to vendors.”22:05
3What would make the product sticky?
Owning the weekly ops review. Once the meeting runs on your numbers, nobody rips you out.31:40
4How are renewals priced at the top end?Not answered
The expert moved on before answering this one.
＋Which segment churns first in a downturn?Asked mid-call
Mid-market, roughly two quarters ahead of enterprise. Enterprise renewals lag the cycle because of annual budgets.38:57
You bring the questions.
We bring the answers.
Every question you prepared, answered and cited. Plus the ones you improvised mid-call.
- ✓Answers land under the exact questions you wrote
- ✓Questions you added during the interview are captured too
- ✓Anything the expert skipped is flagged, never quietly dropped
A transcript you never have to fix
Names, numbers, and jargon corrected.
The same in English, Korean, Japanese, and Chinese.
- ✓Misheard company names, metrics, and terms corrected
- ✓Fillers and false starts stripped, so it reads like a document
- ✓Speaker labels that stay right for the whole call
“uh so gross margin’s like forty for DTC, but on amazon three pee you’re at maybe twenty five after fees and, you know, the ad spend they basically require now…”
Gross margin is around 40% for DTC, but on Amazon 3P you’re at maybe 25% after fees and the ad spend they effectively require now.
On Amazon 3P you’re at maybe 25 points net once you pay fees and the ad spend they basically require now.
And how do DTC margins actually compare at mid scale?
Honestly people overstate the DTC margin thing. Once you load in customer acquisition you’re basically at Amazon parity until you hit real scale.
Every claim, backed by the exact words
Every bullet links to the exact sentence behind it.
Click, land, verify.
- ✓Exact-span highlights, down to the phrase
- ✓One click from note to source
- ✓No more “where did this number come from?”
Stop retyping numbers into Excel
Every figure from the call, already in one table.
Cited to the second.
- ✓Market sizes, margins, and growth rates pulled automatically
- ✓Each figure cited to the moment it was said
- ✓Copy the table straight into your model
“Ulta says yes to data, Sephora says yes to brands Ulta proved.”
“We lose to ‘we’ll do it ourselves’ more than we lose to vendors.”
“You take the standard terms and make the math work on product cost.”
Deck-ready quotes, already picked
The striking lines from every call, verbatim and attributed, ready to paste straight onto the slide.
- ✓The memorable lines picked out of every call
- ✓Verbatim and attributed, safe to cite
- ✓Copy straight onto the slide
Bot, desktop, mobile, or upload
Wherever the conversation happens, every path ends in the same note.
- ✓Meeting bot for Zoom, Meet, and Teams
- ✓Desktop app captures system audio, no bot in the room
- ✓Mobile app for in-person and on the go
- ✓Upload audio files or existing transcripts
Ask Claude.
It knows your interviews.
Connect through MCP. Your interviews become context Claude can search and cite.
- ✓Ask across every project and interview in plain language
- ✓Answers cite the interviews they came from
- ✓Works in the Claude apps your team already uses
In their words
“I really like the Q&A script, especially how everything is properly cited and sourced. The note quality is better than what ChatGPT provides.”
AssociateTMT boutique firm
“We stopped handing note-taking to our RAs. Honestly, I barely take notes in my own interviews anymore. I just run the conversation.”
Engagement ManagerGlobal strategy firm
“We run calls in Japanese and Chinese, and the notes come back just as clean as the English ones. Language coverage used to be its own workstream.”
ConsultantAPAC practice, strategy firm
“Sharing takes one click. The partner reads the takeaways, the team works from the full note, and nobody waits for my recap email.”
PrincipalGrowth-stage fund
Safety first, by design
Built for firms where confidentiality is the product.
SOC 2-aligned infrastructure
Built on architecture with strict controls for data protection, access, and operational security.
No model training on your data
Your transcripts, notes, and communication data are never used to train AI models. Ever.
Full encryption
All transcripts and notes are encrypted both in transit and at rest.
Regular penetration testing
Independent security audits validate the strength of the infrastructure.
Strict project-level data isolation
Each project's data is isolated to prevent cross-project leakage.
Secure external sharing controls
Share notes outside your team without losing control of access.