Skip to content
Shaflex
Back to the blog

Strategy

Building in Public: A Playbook for Founders

What to share, where to share it and how often, with real examples from founders who've done it for years, and the risks that have grown since AI made products easier to copy.

9 min read

Building in public works when you share the decisions, lessons and numbers behind what you're building, on a steady weekly rotation, adapted to each network, and you answer the people who reply. It's not daily screenshots of your dashboard, and in 2026 it's not sharing everything either.

This is the playbook: five kinds of post to rotate through, how each one lands on X, LinkedIn, Bluesky, Mastodon and Threads, a weekly routine that fits in about 15 minutes a day, and what's become too risky to share.

What it looks like when it works

A few people who've built in public for years, and what they actually share:

  • Justin Welsh started by writing for a very narrow audience, early-stage SMB SaaS founders in healthcare. In 2025 he published a breakdown of $10 million in revenue by line of business: products, consulting, sponsorships, subscriptions and community, with the story of how each one started.
  • Buffer has published every employee's salary since 2013 and keeps a live page of its revenue and customer numbers. It has also written honestly about the costs: transparency "can create rigidity" and "does inherently create extra work", and an offer made with the wrong salary band once cost it some trust in its own system.
  • Simon Willison runs a link blog where every link gets his own notes and quotes, and the people he links to are always named. Thousands of posts in, it's a public record of what he's learning.
  • Molly White publishes on her own site first and posts adapted versions to each network: split into threads where the limit is tight, linked back where it isn't, with replies answered on each platform.

None of these are daily "day 47 of building" posts. Each post is useful on its own, and the building is the context.

The risk that's grown

Arvid Kahl built his startup FeedbackPanda in public, shared its revenue as it grew, and sold it. In April 2026 he wrote that sharing revenue and architecture is now "outright dangerous", because AI coding tools make it cheap for someone to clone a product they've watched you build step by step. His test for any post:

"Interesting to participate in. Not easy to clone."

In practice:

  • Share freely: decisions and the reasoning behind them, mistakes and what you learned, how you work, what customers taught you, your opinions on your field.
  • Think first: exact revenue by channel, your full acquisition playbook, architecture details that are the hard part of your product.
  • Don't: share anything about a customer they haven't agreed to, security details, or numbers you'd be embarrassed to correct later.

Five kinds of post

Rotate through these. Covering three a week is plenty.

1. Progress

What you shipped and why it matters to someone. "We added a scheduling feature" is forgettable. "We shipped timezone-aware scheduling after three users in Tokyo told us their posts went out at 4am" tells a story and says who it's for.

2. Lessons and mistakes

What went wrong and what you'd do differently. The pricing change that backfired, the bug that ate a Wednesday, the feature nobody used. These tend to get the best replies because people recognise the problem.

3. Decisions

Why you chose one thing over another: a framework, a price, a first integration, a market. Decision posts start conversations because readers have their own opinions, and they show how you think without giving away how you build.

4. Numbers, with context

Users, revenue, churn, conversion, if you're comfortable sharing them. A number on its own is a brag; a number with a cause is a lesson. If you'd rather not share absolute figures, percentages and trends still work: "trial-to-paid went from 4% to 7% after we cut onboarding to three steps." Our guide to the metrics that matter helps pick which ones.

5. Behind the scenes

The whiteboard, a design before and after, the support message that changed your roadmap (shared with permission). These make you a person rather than a product.

Where to post it

Each network has its own norms. The personal brand playbook by network covers how each one ranks posts; here's how building-in-public posts fit.

X. Still a big founder and developer audience. Posts are 280 characters, or up to 25,000 with Premium. A short post with one specific result, or a thread for a step-by-step story, works well. X's ranker weighs a predicted reply about 7.5 times a like, so end on something people can answer, and spread posts out: each extra post from you in someone's feed scores lower.

LinkedIn. Reaches B2B buyers and people hiring. Posts can run to 3,000 characters. A story from your work with the lesson spelled out does well, and LinkedIn says it favours posts grounded in knowledge and advice. Skip "comment YES" bait; LinkedIn is filtering it.

Bluesky. Strong developer and indie communities. Posts are 300 graphemes. Technical decisions and honest updates fit; marketing language doesn't. Find the custom feeds and starter packs for your field, because the default feed is chronological and that's how strangers find you.

Mastodon. Suits open-source and technical products. Posts are 500 characters by default, and each server sets its own limit. Write like a person, add alt text to every image, use a few CamelCase hashtags, and read your server's rules: some require you to keep interacting if you schedule posts. Our guides to scheduling Mastodon posts and content warnings cover the details.

Threads. A broader, less technical audience. Posts are 500 characters, with an optional text attachment for longer pieces, and one topic tag. Meta says replies drive almost half of Threads' views, so this is where answering comments matters most.

A week of it

DayPostExample
MondayProgress"This week we're rebuilding onboarding. The current version loses 60% of people at step 4. Here's the new flow."
TuesdayDecision"We're launching on Bluesky before Mastodon. Here's why, and what would change our minds."
WednesdayLesson"We spent three hours on a timezone bug. The cause: Brazil dropped daylight saving time in 2019 and our test data hadn't."
ThursdayBehind the scenesThe whiteboard sketch of the new onboarding, with one line on what changed
FridayNumbers"Onboarding v2 after five days: step 4 drop-off is down from 60% to 35%. Still too high. Next we're cutting step 3."

The examples are made up for this post. Each post goes to two or three networks, rewritten for each. Monday's might be a tight X version and a longer LinkedIn version; Thursday's sketch fits Threads and Bluesky.

The routine

Once a week, about 90 minutes: write the week's posts in one sitting, then schedule them. Keeping drafting and editing as separate passes makes both faster. If a week goes badly, a queue a week ahead keeps you posting; we go through that setup in the weekly routine in our scheduling guide.

Every day, about 15 minutes: answer replies on each network, and leave a few useful replies on other founders' posts. Buffer's data found that answering comments went with more engagement on every network it studied, most of all on Threads and LinkedIn. More on that in why replies grow your audience.

Every other week, 15 minutes: look at which posts got real replies and saves, not just likes, and write more like those.

How to tell if it's working

Follower counts move slowly and say little. Better signs:

  • Replies that share their own experience or push back on yours
  • Saves and bookmarks, because someone wanted to come back to it
  • Messages from strangers saying they have the same problem
  • Signups or calls that mention a post

It usually takes months of steady posting before these show up. Welsh, Kahl and Buffer were all at it for years before it looked like a strategy.

Common mistakes

  • Pasting the same text everywhere. Mentions don't resolve, lengths don't fit, and people who follow you on two networks notice. Adapt each copy; our cross-posting guide shows how.
  • Only sharing wins. A stream of milestones reads as a performance. The mistakes are what people remember.
  • Going quiet for weeks. Three posts a week, every week, beats a burst and a silence.
  • Ignoring replies. The conversation is the point, and on Threads and X it's also a big share of the reach.
  • Sounding like a brand. Write in your own voice. If you use AI for drafts, read our guide to writing posts with AI without losing your voice.

Where Shaflex fits

Shaflex is built for this routine. Write one draft and the Composer gives each network its own copy with its own limit, so the X, LinkedIn and Bluesky versions can differ. Split turns a long update into a thread on X, Threads, Bluesky and Mastodon. Set your weekly slots once and the queue publishes in your timezone. On Creator and up, replies from all five networks land in one inbox, so the daily 15 minutes happens in one tab.

Sources

Checked on October 20, 2026.

Questions readers ask

What does building in public mean?

Sharing your product's progress, decisions, problems and numbers openly while you build it, instead of waiting for a polished launch. Done well, it builds trust before people need what you sell and gets you feedback early. The trade-off is that your unfinished work, and sometimes your numbers, are visible to everyone, competitors included.

What should you post when building in public?

Rotate five kinds of post: what you shipped, what went wrong and what you learned, why you made a decision, a number with the context behind it, and a look behind the scenes. Covering three of the five a week gives you variety without inventing something new every day.

Is building in public still worth it in 2026?

For most founders, yes, but share differently. Arvid Kahl, who built and sold a startup in public, wrote in April 2026 that sharing revenue and architecture details is now riskier because AI coding tools make products cheap to clone. His test is whether a post is interesting to follow without being easy to copy. Share decisions and lessons freely; think before sharing the exact recipe.

Which network is best for building in public?

The one your future customers read. X still has a large founder and developer audience. LinkedIn reaches B2B buyers. Bluesky has strong developer and indie communities. Mastodon suits open-source and technical products. Threads reaches a broader, less technical crowd. Start with one or two and adapt each post rather than pasting the same text.

Should you share revenue numbers?

Only if you're comfortable with competitors and customers seeing them, and always with context. '$5K MRR' alone says little. '$5K MRR, up from $2.8K after we added a free trial' teaches something. Buffer has published every salary since 2013 and keeps its revenue public, and has also written about the extra work and rigidity that brings.

How much time does building in public take?

About 15 minutes a day once you have a routine, plus one longer writing session a week. Write the week's posts in one sitting, schedule them, and spend the daily time on replies. Replies are where most of the conversation, and on some networks most of the reach, comes from.

More posts

All posts

Strategy

A Personal Brand Playbook for Each Social Network

What X, Threads, LinkedIn, Bluesky and Mastodon say about how they rank posts, what each one penalises, and what that means for what you post and how often.

12 min read

Strategy

Why Replies Grow Your Audience Faster Than Posts

What X, Threads, LinkedIn, Bluesky and Mastodon say about replies, what Buffer's data shows about answering comments, and how to reply well without automating it.

7 min read

Strategy

How to Start a Personal Brand From Zero: A 90-Day Plan

A week-by-week plan for the first 90 days: a one-sentence thesis, one main network and one secondary, your first ten posts, and what to post on X, Threads, LinkedIn and Bluesky.

15 min read

Write the next post once.

Shaflex fits it to Threads, X, Bluesky, Mastodon and LinkedIn, schedules it on your rhythm and brings the replies back to you.

Free plan, no card needed.