B Babelio Playbook · 18
Investor Update M1
Founder comms · Lines not dots

Investor update #1

A pre-drafted first-month update built on Mark Suster's "lines not dots" structure. Honest about where we are: traction is a working prototype, not users or revenue. Swap the yellow [update] values each month and forward as-is — no more than five numbers change.

Cadence
M1monthly
Stage
Prototype
ICP
Immersion learners
ARPU
$12/mo
How to use this Edit only the yellow [update] spans. They are the ≤5 numbers that change month to month (KPIs, cash, hires). The prose, the asks, and the structure are stable — that consistency is what makes investors read every update. Then hit Copy update below and paste into your email client.

The update

Subject: Babelio update #1 — [May 2026]

Hi all,

TL;DR

We shipped a working desktop prototype that captures any app's audio and renders a live translated subtitle/dub — the OS-level layer no Zoom/YouTube/Teams silo offers. This month was about validation, not scale: we ran [12] Mom-Test interviews with immersion-method language learners (our single ICP) to test the $12/mo subtitle-led wedge. No users or revenue yet — and we won't claim product-market fit until we measure it.

KPIs vs plan

Metric M1 plan Actual Δ
ICP interviews (Mom Test) 15–20 [12] −25%
Total users 300 [0] −100%
Paying users 40 [0] −100%
MRR $480 [$0] −100%
Dub COGS / active-hr ≤$0.50 [$0.48] on plan

Plan = monetization base-case M1. We are deliberately not driving signups before WTP is validated — the only KPI that matters in M1 is interview quality and the Cartesia dub-cost de-risk.

Wins

  • Prototype works end-to-end — per-app audio capture → live translated subtitle/dub on Mac, sub-700ms in our tests. The native-desktop capture primitive (the part no browser extension can reach) is real, not slideware.
  • Cartesia dub cost confirmed at [$0.48]/active-hr — under our $0.50 ceiling, so the subtitle-led $12/mo model is margin-positive on retail APIs. The earlier "must self-host TTS or die" assumption is dead.
  • Strong early signal from immersion communities — interviewees in Refold/Migaku circles already pay $60–180/yr for fan-subs and Anki tooling, and confirm the live/native-client gap we target. Our eval harness + legal posture (EU AI Act Art.50, voice-clone gating) are already above-bar for the vertical.

Lowlights (honest)

  • WTP ceiling may sit below our cost floor. Validated willingness-to-pay for this ICP is $6–12/mo (capped by free extensions), which is near the COGS floor for heavy dubbing. We're betting subtitle-led usage keeps margin healthy — but that bet is unproven until the Van Westendorp + metered concierge test lands next month. It's the #1 thing that can break the model.
  • Founder-market fit + the key hire are open. The entire native-desktop moat rests on one rare hire — a Rust + CoreAudio/WASAPI audio engineer. Until that person is committed, our 4-month MVP timeline is contingent, and any VC will (correctly) flag team risk. This is our top ask below.

Asks

Intro 1 — distribution

An intro to a founder/admin of a large immersion-learning community (Refold, Migaku, MoeWay, or a 100k+ language-learning Discord/subreddit) — we want to time a launch to a recurring immersion-challenge kickoff.

Intro 2 — talent

An intro to a senior Rust / low-level audio engineer who has shipped CoreAudio (macOS taps) or WASAPI (Windows audio sessions) work — ideally someone you've backed at an audio, media, or systems startup.

Intro 3 — capital

An intro to a pre-seed angel or fund that backs consumer-AI / prosumer-desktop tools ($150k–$500k checks) — we expect to open a small pre-seed once WTP and the audio hire are confirmed.

Hire — first hire (the make-or-break role)

Founding Audio / OS-Internals Engineer (Rust + CoreAudio/WASAPI). This is the single hire the moat depends on. If you know someone who lights up at per-process audio capture, low-latency pipelines, and signed-installer trust on Mac & Windows — please forward. [Contractor lead in progress; not yet committed.]

Hires made

[None this month — founder solo; audio hire is the top ask above.]

Cash position

[$110K] in the bank, burning [$6K]/mo (lean: founder + infra/inference) → [~18] months runway. The subtitle-led COGS path is breakeven-capable around ~460 paying users, so we are not forced to raise — a pre-seed would buy speed and the audio hire, not survival.

Next month focus

  • Close the WTP question. Finish 15–20 Mom-Test interviews, run the Van Westendorp price meter + a metered $5/hr concierge test on 30–50 buyers. Apply the <30% kill-criterion to metered-dub WTP.
  • Lock the audio hire. Convert the contractor lead into a committed founding engineer (or sign a clear contract path) so the 4-month MVP timeline becomes real.
  • Instrument the telemetry flywheel day-1. Wire up per-app capture + correction signals now, so the only real moat candidate starts compounding at launch, not "someday."

As always — reply with the intro or the name and I'll take it from there. Thank you.

[Founder name]
Founder, Babelio

Copy behavior: the Copy update button copies the email body in whichever language is currently shown — toggle EN/RU first, then copy. Yellow [update] markers are copied verbatim as bracketed placeholders so you can't accidentally send a stale number; replace them in your draft before sending.

Why this structure

Suster's rule: investors back the line (your trajectory across updates), not a single dot (one month's numbers). A monthly cadence with stable structure turns each send into a data point on a trend they trust. The fastest way to lose that trust is to hide a lowlight — so we lead with the two that matter and pair each with a concrete next step.

×
Don't inflate the dot

Reporting "300 users" when you have a prototype destroys the line. We report 0 users and a working prototype — that's the honest dot, and the trajectory is what we're selling.

×
Don't bury the ask

Three named-type intros + one specific hire. Vague "let us know if you can help" gets ignored; "intro to a CoreAudio engineer" gets forwarded.

×
Don't change the structure

Same six sections every month. Only the ≤5 yellow numbers move. Predictable shape = readable line.

Babelio · Playbook artifact 18 of 24 · Founder comms Lines not dots · forward as-is