Once a year, we publish the boring numbers. Here are 2025's.
Headline
| Paid accounts (Dec 31) | 7,140 (+153%) |
|---|---|
| Free accounts (Dec 31) | 21,500 (+133%) |
| Revenue (CAD) | $498,000 |
| Profit margin | 11% (~$55,000) |
| Headcount | 6 (+2) |
| Servers running production | 11 (+5) |
| Mail delivered | ~340 million messages |
| Average inbound queue time | 2.3 seconds |
| Uptime | 99.992% |
| Customer support tickets | 4,180 (avg first response: 6h) |
| Complaints filed against us with the spam control bodies | 3 (all dismissed) |
| Law enforcement requests received | 2 (one quashed, one complied with under MLAT) |
What surprised us
The free tier converts at 7%. When we first set up the free plan, we expected most people to stay on it forever. They mostly don't. Of the people who sign up free and stick around for 60 days, about 7% pick a paid plan within their first year. We didn't expect that, and it changes the economics of free-tier abuse for us — it now pays for itself.
Custom domains drive Plus signups more than aliases do. When we shipped Plus in early 2024, we thought the killer feature was unlimited aliases. Turned out it was custom domains. Of people who upgrade from Free to Plus, 64% do so within a week of adding their own domain.
Bare metal saved us more than we expected. We left AWS in November 2024. We projected ~$3,800/month in savings; we actually got $4,650/month. Latency to typical receivers is also down meaningfully — Yahoo's been the biggest improvement, probably because they ratelimit cloud IPs more aggressively than residential-ish ones.
The "I want to send mail from my own domain" use case is bigger than we thought. About a quarter of our Pro customers run a small business — bakery, therapist's practice, freelance design, art studio. None of them want Microsoft 365 or Google Workspace; both are expensive and feel corporate. They want what we have: a clean per-mailbox setup with no enterprise SSO theater.
What didn't go well
The January queue saturation incident. A single customer's account was attacked with a flood of spam, and our shared inbound queue ground to a 45-minute delay before we mitigated. Affected everyone, not just the customer being attacked. We've since added per-mailbox rate limiting on inbound, which would have caught it. Postmortem is on the blog.
The webmail rewrite shipped six months late. Sara's been working on the V2 webmail since Q4 2024. It was supposed to ship in Q2 2025. It shipped in early Q4. The slip wasn't catastrophic but it was bigger than the team-of-three plan supported, which is part of why we hired Tomás.
We declined a buyout offer. A private-equity-backed roll-up made us an offer in September that would have changed the company permanently. We declined. None of us would still be working here in two years if we'd taken it. It was a real number, but small enough relative to what each of us makes here that turning it down wasn't actually agonizing.
What we're doing in 2026
- Mobile apps. Native iOS and Android, written in Swift and Kotlin respectively. No cross-platform framework, no React Native. We want them to feel like first-class system apps. Targeting Q2 launch.
- Group accounts. Better admin tooling for families and small teams sharing a Pro plan. Currently it's basically "the admin manages everything from one Settings page" — we want shared calendars, shared contacts, per-member usage stats.
- OpenPGP everywhere. Native WKD, automatic key publishing, optional client-side encryption in the webmail. Going to take all year.
- Hire one more. A second support engineer, ideally in a European timezone so we have better coverage for our growing UK customer base.
- Hold steady on price. Plus stays at $4 CAD, Pro stays at $9. We're profitable; we don't need a raise.
Thank you to everyone paying us $4 a month. Genuinely. It's wild to us that you'd choose this small operation over the giants, and we don't take it lightly.
— Maya