"Nobody's Coming" Is the Worst Part
If you've ever shipped a side project, you know this feeling.
Launch day. Open Google Analytics. Real-time users: 0.
Next day, still 0. Day 3, finally 1 — it was me.
When I launched my developer tool suite, the first week's engaged MAU was 5. I didn't investigate how many were me.
Three weeks later, MAU hit 175. I spent zero on ads.
This is the full record — what worked, what flopped, and all the real numbers.
The tool suite discussed in this article is PureMark — 7 developer tools (JSON, Base64, Diff, JWT, Timestamp, and more) that work with zero clicks. Just copy something to your clipboard and open the page.
What I Built: 7 Tools, One Umbrella
Small tools developers use daily — JSON formatting, Base64 decode, URL encode, diff comparison, JWT decode, timestamp conversion — bundled into one branded suite.
| Tool | What It Does |
|---|---|
| Annotate | Add arrows & text to screenshots (PWA) |
| JSON Formatter | Auto-formats on paste |
| Base64 Encoder/Decoder | One-action encode/decode |
| URL Encoder/Decoder | Percent-encoding conversion |
| Diff Checker | Text comparison with custom Myers diff |
| Timestamp Converter | Unix timestamp ↔ human-readable dates |
| JWT Decoder | Safely inspect tokens locally |
Every tool shares one design principle: zero click. Copy something, open the page, and the result is already there. No Ctrl+V needed.
Plus a Chrome extension (PureMark Detect) that auto-detects clipboard format and opens the right tool in one click.
All the Numbers
No point hiding them, so here's everything:
MAU Growth
W10 (2/28–3/2): 5 ← Measurement started. Mostly me.
W11 (3/3–3/9): 80 ← Published 3 Zenn articles at once
W12 (3/10–3/14): 135 ← SEO accumulation kicking in
W13 (3/14–3/18): 175 ← Zero articles published. Organic only.
By Channel
| Channel | W13 Numbers | Notes |
|---|---|---|
| Zenn (JP dev blog) | 687 total PV / 13 likes | 12 articles. +4 likes in a no-publish week |
| Dev.to | 336 total PV | 11 English articles |
| Google Search | 22 impressions / 0 clicks | Still impression-only stage |
| Chrome Extension | 23 installs | 1 active user (painful) |
| X (Twitter) | 1,242 impressions | Basically crickets |
Ad spend: $0. Running costs: Cloudflare free tier + domain registration.
What Worked #1: Designing Articles as SEO Machines
90% of growth came from articles. But not random ones.
One Article = One Problem, Not One Tool
I never wrote "I built a JSON Formatter." Instead:
- "Do You Know Where Your JWT Goes When You Paste It Into a Developer Tool?" — Lead with privacy concerns, introduce the tool as the solution
- "130 Lines of Myers Diff From Scratch — And the Off-by-One Bug That Took 2 Days" — Lead with a real debugging story
- "How I Turned My Development Logs Into Self-Accumulating Assets with Claude Code × Obsidian" — Lead with a workflow pain point
Every article starts with a problem the reader recognizes. The tool appears naturally as part of the solution.
Failure Stories Get Likes. Tutorials Don't.
After 12 articles, the pattern is unmistakable:
| Articles That Got Likes | Common Thread |
|---|---|
| Skitch Alternative PWA (3 likes) | Flutter detour → PWA pivot. 3 hours on arrow rendering |
| Claude Code × Obsidian (5 likes) | 2 AM frustration → bind mount solution |
| Myers Diff 130 lines (1 like) | Off-by-one bug that cost 2 days |
Every article without a failure story got zero likes. n=12, no exceptions.
Try the zero-click experience yourself: visit JSON Formatter with some JSON in your clipboard — the result appears instantly.
What Worked #2: Making Tools "Refer" Each Other
This is why I built a suite instead of separate sites.
When you format JSON and there's a JWT token in your clipboard, the tool shows a button: "Looks like a JWT. Decode it?" Click it, and you're on the JWT Decoder with the token already parsed.
W13's cross-tool transition rate was 5.75% — 6 out of every 100 users moved from one tool to another.
This has compound effects: longer sessions improve Google's quality signals, and users discover tools they didn't know existed.
What Flopped — Spectacularly
Show HN → Zero
Posted to Hacker News. Result: 0 upvotes, 0 comments.
I'd analyzed it beforehand — developer tools are a saturated category on HN, and without genuine technical novelty (novel algorithm, unique architecture), it won't land. Knowing that in advance kept the emotional damage manageable.
Lesson: Pre-calibrate your expectations. "Light toss" mentality saved me from spiraling.
Share URL Feature → 4 Weeks at 0% Usage
Every tool has "recipe URLs" — encode input data in the URL hash so sharing a link reproduces the exact state. Technically elegant implementation.
Times used: Zero. Four consecutive weeks.
Button placement? No demand? I changed the UI in W13 — next week will tell.
X (Twitter) → Invisible
Tweeted. Made demo videos. Result: 1,242 impressions total, 3 likes.
Plenty of "use Twitter for indie marketing" advice out there. Reality: if your account has no followers, nothing reaches anyone. Not even good content.
Design Decisions That Mattered
"Everything Runs Locally" Pays Off Slowly
Every PureMark tool processes data in the browser. No server calls. Clipboard reading, JSON formatting, JWT decoding — all client-side.
Initially just a differentiator. But when I wrote the privacy-focused article, it got 2 likes in 4 days — fast for my scale. "Do you know where your JWT goes?" resonated. There's a real audience that cares about this.
KPI Dashboard Before Tools
I built the measurement infrastructure first. Cloudflare Workers + KV, pulling GA4/GSC/Zenn/Dev.to/X data every 6 hours into one dashboard.
Best decision I made. Without numbers, you can't tell what's working. The difference between "article week" and "no-article week" is visible in data, which makes the next action obvious.
3-Week Takeaways
Worked
- Zenn article SEO accumulation — MAU grew +30% in W13 with zero new articles. Content is self-sustaining.
- Story-driven articles — Only articles with failure stories earned likes.
- Cross-tool navigation UI — 5.75% transition rate validates the suite strategy.
- Dev.to English versions — GSC impressions from English queries growing.
Didn't Work
- Show HN — 0 upvotes. Saturated category needs technical novelty.
- X (Twitter) — Doesn't work without an existing audience.
- Share URLs — 4 weeks at 0% usage. Feature ≠ adoption.
Still Unknown
- When will GSC clicks start? — 22 impressions but 0 clicks. Need higher search rankings.
- Chrome extension active rate — 23 installs, 1 active. Need a reason to keep using it.
- Is this growth rate sustainable? — Gate 1 target is MAU 500 by mid-April. Need +11%/week.
Before / After
Before (3 weeks ago):
Built 7 tools. Was proud of them. MAU: 5. Show HN: 0. "Build it and they will come" was a lie.
After (now):
MAU 175. ~30% weekly growth. Users increasing even in weeks with no new content. What changed wasn't the product — it was spending equal time on "how to reach people" as on writing code.
In indie development, writing code is half the job. The other half is getting discovered, and there are reproducible methods for that.
All tools run in your browser. No sign-up, no data sent to servers. Try them → puremark.app
Top comments (0)