Build in Public: A 7-Stage Plan to Launch an Extension
Originally published on ExtensionLaunch. Most Chrome extensions launch to an empty listing and a trickle of strangers. Building in public is the cheapest way to change that: share the problem, the progress and the numbe
Originally published on ExtensionLaunch.
Most Chrome extensions launch to an empty listing and a trickle of strangers. Building in public is the cheapest way to change that: share the problem, the progress and the numbers while you build, and launch day stops being the day you start from zero. This is the 7-stage plan, with the store metrics worth posting and the one number almost everyone misreads.
Why extensions are the easiest product to build in public
Extensions fix small, visible annoyances in apps people already use, so they demo in seconds. A 20-second screen recording of a button appearing in Gmail beats three paragraphs of feature copy.
They also come with natural story beats: the first prototype loaded unpacked, the first permission decision, a rejection and the fix, the first 100 installs. Each one is a post. Geoffrey Litt, who writes about shipping browser extensions as side projects, shared a build with roughly 100 beta testers and found DMs were the best channel for early feedback â building in public is how you find those first 100 people.
The 7 stages and what to post
| Stage | Post | Example |
|---|---|---|
| Before any code | The problem and who has it | Moving Analytics tables into Sheets costs me 10 minutes a week; I'm building the one-click fix. |
| First prototype | A short screen recording | Day 4: one click, table copied with formatting. Rough, but it works. |
| While building | Decisions and trade-offs | Cut the broad "read all sites" permission; it now runs only on click. |
| Store submission | The process, setbacks included | Rejected for a vague description. Rewrote, resubmitted, approved in two days. |
| Testing | What testers found | 12 testers, 3 hit the same dark-mode bug. Fixed in v0.3. |
| Launch | One link and one ask | It's live. Weekly Analytics users, I'd love your feedback. |
| After launch | Numbers with context | Week 3: 420 installs, 180 weekly users â one newsletter mention drove most of them. |
Posts about decisions get more replies than progress updates, because readers can disagree. Permission choices, the monetization model, the framework pick (WXT or Plasmo) and which browsers to support first are all safe, productive topics.
Post numbers that survive a fact-check
The Chrome Web Store developer dashboard gives you three shareable reports, as Google's store metrics documentation describes:
- Installs and uninstalls per day, by country, language and OS.
- Weekly users, split by country, language, OS and item version.
- Impressions: how often people view your listing in the store.
Read before you post: the weekly users figure "only captures installations; it doesn't monitor whether users are active or not." Call them installed users, not active users â honest labels beat big numbers.
The Impressions report also breaks listing views down by utm_source, utm_medium and utm_campaign. Tag your store link in posts (?utm_source=x&utm_medium=social&utm_campaign=build-in-public) and you can later prove which platform actually sent people to your listing.
Numbers worth posting with context: first-week installs and where they came from, first-week uninstalls and the top reasons, views-to-installs conversion before and after changing screenshots, and revenue ranges if you're comfortable.
What never to post
- User data. No names, emails, browsing data or support messages in screenshots. Crop, blur, or recreate with test data.
- Secrets. Check every code screenshot for API keys, tokens and private URLs.
- Other people's words. Ask before quoting a tester or customer, even anonymously.
- Implementation details you aren't ready to defend.
One rule is not optional: never offer anything in exchange for installs, ratings or reviews, and don't ask friends to rate the listing â the Chrome Web Store treats both as manipulation. Asking your audience for honest feedback is fine; rewarding a five-star rating is not.
From followers to installs
An audience only matters if there's a next step. Build the path before you need it:
- A landing page with a waitlist form, live before launch.
- A tester group from your most engaged followers, using a private or unlisted release.
- One launch day: announce it a week ahead, share one link on the day.
- A follow-up post a week later â what happened, what broke, what you're fixing. It often beats the launch post.
Consistency beats volume: a weekly update for six months is worth more than ten posts in launch week. A 15-minute routine works: Monday, note one decision, one number and one lesson; midweek, record a 20-second clip of what you shipped; Friday, post the weekly update and reply to every comment.
Key takeaways
- Every post should help someone who has never heard of you: a demo, a decision, a number with context, or a lesson.
- Extensions demo in seconds â a short recording beats a paragraph.
- Post aggregate store numbers, but explain them: weekly users counts installations, not active people.
- Never share user data, secrets or private feedback, and never trade installs or reviews for rewards.
- Give followers a path: waitlist, tester group, one launch day, one follow-up.
The full build-in-public guide for Chrome extensions adds platform-by-platform advice (X, Bluesky, LinkedIn, Indie Hackers), a deeper look at what to keep private, and the launch-day steps that follow. If you want to see what shipped this week, browse the new browser extension launches on ExtensionLaunch.
Which one post earned you the most installs, and what did you share in it?
Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes â full credit and traffic to the original publisher.