b3126044e8
- Simplify flatten_vault API to return usize instead of MigrationResult struct - Add KEEP_FOLDERS: attachments/ and _themes/ alongside type/, config/, theme/ - Use HashSet for collision tracking in unique_filename - Update wikilinks from path-based [[folder/slug]] to title-based [[slug]] - Clean up empty directories after flattening - Flatten demo-vault-v2: move all notes from type-based subfolders to root - Update smoke tests for flat vault structure - Remove migrate_to_flat_vault from repair_vault (one-time migration only) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
5.2 KiB
2025
Theme
2025 is the scale year — the year the question shifted from "can this work?" to "how big can this get?" Coming off the momentum of 2024, every metric pointed upward: subscribers, revenue, podcast downloads, team capacity. The goal wasn't just to grow more, but to grow better — building systems and a team that could sustain 85k subscribers, EUR 22k MRR, and a product (Laputa) without everything depending on me.
But scaling surfaced a tension I hadn't anticipated: the more successful Refactoring becomes, the more it pulls me away from the craft that made it successful in the first place — writing. Managing sponsors, coordinating a team, shipping a product, speaking at conferences — these are all valuable, but they're not writing. The creative struggle of 2025 has been protecting the space to think deeply and write well, while everything else demands attention. Some months I won that battle. Others, I didn't.
Highlights
- Newsletter grew from 50k to 82k+ subscribers (on track for 2025-reach-85k-subscribers by year end)
- measure-sponsorship-mrr grew from EUR 11.4k to EUR 20.1k by end of Q3, tracking toward 2025-reach-22k-mrr
- Shipped three versions of Laputa: 25q1-laputa-v1, 25q2-laputa-v2, 25q4-laputa-v3 — from internal prototype to closed beta candidate (2025-ship-laputa)
- Podcast grew to 18k downloads/month across Seasons 3 and 4 (25q2-podcast-season-3, 25q3-podcast-season-4)
- Published 25q3-ebook "The Engineering Leader's Playbook" — 1,400+ copies sold in the first month
- Launched 25q3-community-launch with 850 founding members
- Rode Stelvio from Bormio in 1h52m — 2025-ride-stelvio achieved
- Keynote at 25q3-leaddev-london to 1,200 attendees — biggest speaking engagement to date
- 25q2-team-retreat in Tuscany aligned the team for H2 and catalyzed the community idea
- Executed 25q1-rate-increase — sponsor rates up 25%, retention at 90%
By the numbers
- Subscribers: 50.2k to ~82k (projected 85k by Dec) — +64% YoY
- Sponsorship MRR: EUR 11.4k to EUR 20.1k (Q3 close) — targeting EUR 22k by year end
- Podcast downloads: 8.2k/month to 18k/month — +120% YoY
- Premium subscribers: 320 to 680+ — +112% YoY
- Ebook sales: 1,400+ copies (launched Q3)
- Community members: 850 (launched Q3)
- Books read: on track for 2025-read-20-books (16 through Q3)
- Cycling km: ~5,800 km through Q3, projected 7,500+ for year — measure-cycling-km-per-month
- Stelvio: 1h52m from Bormio — personal milestone
- Conference talks: 2 (LeadDev London keynote, plus a workshop)
- Total revenue: projected ~EUR 260k (+68% vs 2024)
- Resting HR: measure-resting-hr averaged 48 bpm, down from 52 bpm in 2024
Reflections
The most important thing I did in 2025 was hire well and delegate aggressively. person-matteo-cellini now runs the entire sponsor relationship lifecycle — from prospecting to renewal — and does it better than I did. person-paco-furiani owns operations, invoicing, and now coordinates the community logistics. person-sara-ricci handles podcast production and guest coordination. For the first time, I can take a week off and nothing breaks. That's the real milestone, even if it doesn't show up in the metrics.
The ebook and community launches in Q3 represented a philosophical shift. For three years, Refactoring was a one-to-many broadcast channel: I write, you read. The community made it many-to-many. Engineering leaders helping each other, sharing war stories, giving feedback on each other's challenges. I'm a facilitator now, not just a broadcaster. It's a better model, but it requires a different kind of energy — more listening, less performing.
Laputa has been the most surprising journey of the year. What started as a weekend hack in Q4 2024 has become a serious product with three major versions shipped. Using it daily to manage my 9,000+ file vault has been the best kind of dogfooding — every friction point I hit becomes a feature. The AI chat panel in v3 was a breakthrough moment: asking your own vault questions and getting contextual, linked answers feels like the future of personal knowledge management. Whether this becomes a real product with real users or stays a personal tool, building it has made me a better engineer and a better thinker.
If I'm being honest, Q3 nearly broke me. Three major launches, a keynote, and peak cycling training in the same 13 weeks was too much. I hit a wall in late August — low energy, poor sleep, short temper. person-paco-furiani noticed before I did. I took 10 days fully off in early September and came back feeling human again. The lesson is one I keep having to relearn: ambition without recovery is just a burnout timeline. Going into Q4 and planning for 2026, I'm trying to build more white space into the calendar. Fewer launches per quarter. More weeks with nothing scheduled. The business can handle it — the question is whether I can resist the urge to fill every gap with a new project. History suggests I can't, but I'm working on it.