# Supernormal Financial Model

AI-powered automatic meeting notes for Google Meet (expanding to Zoom and Teams), delivered via Chrome extension with no user input required.

- Canonical: https://finamodel.com/startups/supernormal
- Excel download: https://finamodel.com/startup-models/supernormal.xlsx
- Category: Enterprise/Security
- Model type: SaaS ARR / Valuation
- Funding round: Seed
- Funding: $10M
- Founded: 2023
- Geography: Global; ICP anchored on Google Workspace companies. Named customers include Netflix, Wayfair, Salesforce, Clover, TNW. [DECK, slides 11, 16]
- Customer: B2B

## About the company

Supernormal creates AI-generated meeting notes through a low-friction Chrome-extension workflow for Google Meet, with Zoom and Teams expansion planned. The product requires no active note taking, turning recorded meetings into a shareable artifact for the wider organisation.

The business uses a product-led, seat-based motion. Shared notes can bring new users into a workspace, while paid seats and team plans monetise broader adoption; the deck identifies AI inference cost per recorded hour as a key operating driver.

The model tracks activation, shared-note sign-ups, paid conversion, seats, expansion, and churn. Recorded hours and inference cost build delivery COGS, while product-led acquisition, sales assistance, gross margin, headcount, and operating expenses determine cash runway.

## What's included

- 5-year monthly revenue build with stage-appropriate growth assumptions
- Full P&L, headcount plan, and operating-expense schedule
- Cash-flow statement, runway, and burn-rate tracking
- Valuation via exit multiple with a DCF cross-check
- Returns analysis with MOIC and IRR
- Unit economics including CAC, LTV, payback, and cohort retention

## Product & value proposition

- Chrome extension that auto-captures Google Meet audio, produces a full recording, transcript, and AI-crafted structured notes (Summary, Decisions, Action Items, Blockers) in <2 seconds after meeting ends.
- Notes are organized into extensible "Sections" - each section is a prompt run through a fine-tuned base model.
- Action-item auto-routing: detected action items can trigger automated follow-up emails / task assignments.
- Team/org layer: personal recordings → shared team spaces → org-wide repository (searchable recordings + transcripts).
- Quality benchmark: 8.2/10 notes rated indistinguishable from human-written.
- 60,000 hours of annotated meetings used to build proprietary training data (bootstrapped).
- Expanding to Zoom and Teams (beta).

## Market

- TAM: $22B
- SAM: $12B
- Early Target Market (Google Meet users): $250M
- Problem quantification: Lack of meeting notes costs businesses $10,000 per employee annually; duplicate meetings up 25% without notes; avg person could skip >20% of meetings with an alternative way to stay informed.

## Revenue model

- Seat-based SaaS subscription.
- Bottom-up / PLG motion: individual adoption within org → team expansion → org-wide deal.
- Channels: Chrome Web Store (Google Meet integration), performance marketing, vertical-by-vertical GTM (Sales+CS, Design, Ops, HR, Finance).
- K-factor virality: AI meeting notes are shared by default, creating inherent viral loop.

## Traction & metrics

- $600K in productivity gains delivered over last 3 months; implied value-based metric, not revenue.
- 125 paying teams in a month
- Avg paying organization starts with 5 seats in first week, doubling every 10 days
- 48% sign-up to aha-moment conversion
- 72% DAU/MAU
- 3.8 avg recordings per user per day
- Hours saved chart (slide 10 image): exponential growth from ~Feb 2022 to Oct 2022; y-axis peaks at ~16,000 hours/month by October 2022.
- Fundraising target: $4M ARR with 200K+ DAU
- Notable customers: Netflix, Wayfair, Salesforce, Clover, TNW, Pillar, Exness, Veolia Water, Headset.

## Unit economics

- Cost per recorded hour (AI inference):
  - GPT-3 Da Vinci 2: ~$8.50 at Q1, declining to ~$4.80 by Q4 (year 2)
  - GPT-4: ~$11.00 at Q2, declining to ~$6.80 by Q4 (year 2)
  - Supernormal V1: ~$4.20 at Q1, declining to ~$2.90 by Q3
  - Supernormal V2: ~$1.70 at Q3, stabilizing ~$1.30 by Q1 (year 2)
  - Supernormal V3: ~$0.90 at Q1 (year 2), stable ~$0.80 by Q4
  - On Device: approaching ~$0.20 by Q2+ (year 2)
- Seat expansion: 5 seats → doubling every 10 days implies ~40 seats within a month per org.

## Competition / moat

- No explicit competitor slide. "Why Now" cites: "No one on the market has a completely automatic solution."
- Moats: 60K hours proprietary training data; custom real-time processing protocol (<2 sec latency); fine-tuned models (V1/V2/V3) with dramatically lower inference cost than GPT-3/GPT-4; on-device roadmap for near-zero cost.
- Named tech benchmarks: GPT-3 Da Vinci 2, GPT-4 used as cost-per-hour comparators.

## Team & funding ask / use of funds

- Colin Treseler, Co-founder: ran ML teams at Facebook/Instagram, Klarna.
- Fabian Perez, Co-founder: Director of Design & Engineering at GitHub, Splice.
- Jim Kleban, Head of ML (ML PhD): ML at Stripe, Facebook, Microsoft.
- Julio Ody, Full Stack Engineer: open-source author/contributor 10+ years.
- Committed in round: EQT Ventures (pre-seed) @ $1M+; Angels @ $200K (Jehad Affoneh / Toast CDO, Nir Eyal / Hooked author, Mick Johnson / FB News Feed, Alex Schleifer / fmr CDO Airbnb, Josh Brewer / fmr Principal Designer Twitter, David Helgason / Unity founder, et al.).
- Round target size: Not explicitly stated.
- Fundraising target milestone: $4M ARR, 200K+ DAU.

## Recommended financial model

- **Archetype + why:** Bottom-up SaaS ARR model with PLG seat-expansion mechanics. The revenue engine is seat-based subscriptions sold via viral individual adoption expanding to teams and orgs. The dominant growth driver is the K-factor (shared notes → new sign-ups → paying teams), and the key cost driver is AI inference cost per recorded hour (declining rapidly). A standard SaaS ARR waterfall (new ARR, expansion ARR, churn ARR, net revenue retention) fits perfectly.

- **Forecast horizon & granularity:** Monthly for Year 1–2 (aligns with seed-stage operational cadence); quarterly for Years 3–5. 3-year base case minimum; 5-year if valuation is needed.

- **Key drivers & assumptions:**

| Driver | Value |
| -- | -- |
| Starting paying teams | 125 |
| Avg seats per new org (week 1) | 5 |
| Seat doubling period (within org) | 10 days → ~3x per month |
| Max seats per org (ceiling) | 50 |
| Monthly new org additions | Extrapolated from hours-saved curve growth rate |
| Seat price (monthly) | $15–$25 / seat / month |
| Sign-up → paying conversion | 48% sign-up to aha moment; paying conversion rate unknown |
| DAU/MAU | 72% |
| Avg recordings per user per day | 3.8 |
| Avg meeting duration (hours) | 0.75 hrs |
| AI cost per recorded hour (current) | ~$0.80–$0.90 (V3) trending to ~$0.20 (on-device) |
| Gross margin | 60–70% |
| Monthly churn (team-level) | 2–4% |
| Net Revenue Retention | 120–140% |
| Headcount additions (Y1) | ~8–12 FTEs |
| S&M spend | 20–30% of revenue |

- **Scenarios (Base / Bull / Bear - which variables flex):**
  - **Bull:** Seat doubling holds at 10-day pace, higher org ceiling (100 seats), price at $25/seat, on-device cost realized by Q3 Year 2.
  - **Base:** Seat expansion slows post-month-1 (doubling at 20 days), org ceiling 50 seats, $18/seat, V3 cost stays at ~$0.85/hr.
  - **Bear:** High churn (5%), low conversion, expansion plateaus at 10 seats/org, price pressure to $12/seat.

- **Required sheets / outputs:**
  1. Assumptions - all drivers listed above with toggle for Base/Bull/Bear
  2. User & Seat Model - new sign-ups → paying orgs → seat expansion → total paying seats (monthly)
  3. Revenue - MRR/ARR waterfall (new, expansion, churn, net), broken out by tier if pricing tiers are added
  4. COGS & Gross Margin - AI inference cost per recorded hour × total hours; infrastructure/hosting
  5. OpEx - headcount (R&D, S&M, G&A), other opex
  6. P&L - monthly income statement through EBITDA
  7. Cash & Runway - cash from raise, burn rate, months of runway
  8. KPI Dashboard - DAU, MAU, DAU/MAU, recordings/user/day, paying teams, NRR, CAC payback (once data available)

## Frequently asked questions

### Is the Supernormal financial model free?

Yes. The Supernormal model is a free Excel download with live formulas.

### Can I change the assumptions?

Yes. The workbook is editable and its live formulas recalculate when assumptions change.