Lesson 4 of 87 min read

Building in public without oversharing

Know what to share (numbers, decisions, mistakes, lessons) and what to keep private, and build a four-post weekly rhythm you can sustain.

In this lesson you will

  • Know which kinds of build-in-public posts are worth writing
  • Draw a clear line around what you never share
  • Plan a weekly rhythm of four post types that you can keep for months

"Build in public" sounds simple: share what you are doing while you do it. In practice developers get stuck at both ends. Some share nothing, because every update feels too small or too unfinished. Others share everything, including things they later regret: a customer's name in a screenshot, a server path in a log, a revenue number they were not ready to stand behind.

This lesson gives you a middle path. You will learn the post types that work, a hard list of what stays private, and a weekly rhythm built from four post types that you can keep up long after the first burst of enthusiasm fades. Releases (lesson 3) cover what you shipped. This lesson covers everything around the shipping: why, how, and what went wrong.

What is worth sharing

Building in public works because people like to follow a story with real stakes. They do not need every commit. They need the moments where something was decided, learned or measured. Four kinds of posts carry most of that.

Decisions. Why you chose one approach over another. Why you dropped a feature. Why you picked a boring database. Decisions are interesting because readers face the same choices, and because a good explanation shows how you think.

Mistakes. What broke, why, and what you changed so it does not happen again. A clear, blameless mistake post earns more trust than ten launch announcements, because it shows you take problems seriously.

Lessons. Something you learned that others can use without your product: a debugging technique, a library quirk, a pricing page lesson, a support pattern. These posts travel furthest, because they are useful to people who will never be your users.

Numbers. Real figures from your own project: signups this month, a load time before and after, the size of your test suite, how long a build takes. Numbers make a story concrete. Only share numbers you are comfortable having quoted back to you later, and always say what they measure.

A fifth type, progress, is the "here's what I worked on this week" post. It is useful as a weekly anchor but weak on its own. Always pair progress with one decision, mistake, lesson or number, or it reads like a status report.

What never gets shared

Write this list down and keep it next to your drafts. These are not judgment calls; they stay private every time.

  • Customer data. Names, emails, company names, usage data that identifies anyone, support conversations. Even "a big customer" can be guessed in a small niche. If a customer agrees in writing to be named, that is a different post, and still check what is visible in the screenshot.
  • Security details. Unpatched vulnerabilities, how your auth works internally, server names, internal URLs, key formats, which provider stores what. A postmortem can say "a misconfigured permission let one user see another's report" without a map of your infrastructure.
  • Anything under NDA or contract. Partner terms, pricing you were given privately, work you did for a client.
  • Other people's mistakes. Your postmortem is about your system. A vendor's outage or a contributor's bug is theirs to share.
  • Secrets in screenshots. Terminal windows, browser tabs, dev tools and environment panels leak tokens, account ids and internal hosts. Crop tight and check every corner before posting.

A useful test before posting: "If a competitor, a customer and an attacker all read this, is each of them fine with what they learn?" If the attacker learns something useful, cut it.

Numbers are a gray area. Revenue, user counts and churn are yours to share if you choose, but once posted they are public forever and will be compared to every future number. Decide your policy once (for example, "we share growth rates but not absolute revenue") and stick to it, so you are not making that call on a bad week.

Consistency over one big hit

Many people start building in public hoping for one post that takes off. Occasionally one does. Usually it does not, and the person concludes it does not work. The people who get results from building in public mostly have one thing in common: they kept posting for months, at a pace they could sustain.

That means the rhythm matters more than any single post. A sustainable rhythm has three properties:

  1. Fixed slots. The same days and times each week, so you never decide "should I post today?"
  2. Fixed types. Each slot has a post type, so you never stare at a blank box wondering what to say.
  3. Small enough. Small enough to keep in your worst week, not your best one.

For a solo developer we recommend four posts a week on your reach network, plus one longer piece a month on your depth network. That is a starting point. If four feels like too many, do two. Two a week for six months beats ten a week for three weeks.

A four-slot weekly rhythm

Here is a weekly plan built from the post types above. Each slot names the type and the question that fills it.

SlotTypeThe question to answer
MondayRelease or progressWhat changed for users since last week?
TuesdayDecisionWhat did I choose this week, and why?
ThursdayLessonWhat did I learn that someone else can use?
FridayMistake or numberWhat went wrong, or what did I measure?

You do not need to write these on the day. Most people write all four in one sitting, often Sunday evening or Monday morning, and schedule them. Thirty to forty-five minutes is a realistic budget once you have done it twice.

Post templates help a lot here. A template is a structure with blanks: you bring the facts, it brings the shape. Our free post templates include several written for builders, such as "Explain a hard decision you made on your project", "Show what you cut to ship on time", "A mistake with a lesson others can borrow", "Review your month in wins, misses, and lessons" and "If I could change one thing in my industry". Pick one template per slot and reuse it every week until it feels stale, then swap it.

Here is a filled example for the Tuesday decision slot, for a clearly hypothetical project:

Say you maintain a CLI for database migrations:

I made a call this week I went back and forth on: no GUI.
Everyone asked for one. But every hour on a GUI is an hour
not spent on rollbacks, which is why people use the tool.
So: terminal only for now, better rollbacks in two weeks.
Would you have made the same call?

It names the decision, the tradeoff, what happens next, and asks a real question. That is all a decision post needs.

Worked example: how we share our own build

We build our product in public too, and we had to draw the same lines. Here is how we apply this lesson to our own accounts.

What we share. What shipped (from our changelog), why we built it, and the decisions behind it. For a release like our money-back guarantee, "we have a guarantee now" is the least interesting post. The decision post is the one worth writing: what we wanted a new customer to feel when trying us, and what we gave up to offer it. For signing up from an agent, the decision worth explaining is the line we hold everywhere: an agent can draft, but nothing it drafts publishes without a person approving it.

What we never share. Customer names or data, anything about our infrastructure or vendors, security internals, and numbers we have not checked. We do not post screenshots of the dashboard with real customer content in them; our screenshots use our own accounts.

Our slots. We plan our week around X, LinkedIn, Bluesky and Discord. A release post comes from the changelog. A decision or lesson post comes from what we argued about that week. Discord gets the same news with more room for questions.

Here is a decision post in the shape we use, about letting coding agents sign up and draft posts:

We let coding agents sign up and draft posts for you.
The decision we spent longest on was the opposite one:
what an agent may never do. Our answer: it never publishes.
It drafts and proposes; a person approves. Every time.

It shares a real decision and the reasoning, reveals nothing about how it is built, and invites people to judge the choice themselves. It also avoids typing prices or internal details into a post; anyone who wants more follows the link.

Recap

  • The posts that work are decisions, mistakes, lessons and numbers. Pair weekly progress with one of them.
  • Customer data, security details, NDA material and secrets in screenshots never get shared.
  • Rhythm beats any single post: fixed slots, fixed types, small enough for your worst week.
  • Templates give each slot a shape; reuse them until they feel stale.

Exercise

Plan your four-slot week

  1. Write your "never share" list using the five categories in this lesson, adding anything specific to your project.
  2. Open the free post templates and pick four, one for each slot: release or progress, decision, lesson, mistake or number.
  3. Choose the days and times for your four slots. Pick times you can actually keep.
  4. Fill in all four templates with facts from your past week. Check each against your never-share list and the "competitor, customer, attacker" test.
  5. Done looks like: a simple table of four slots with day, time, post type and template, plus four ready drafts for next week.

Go further

Queue slots in FoPost

In FoPost you can save your weekly slots as a posting schedule, so each new post drops into the next free slot instead of you picking a time. Set your days and times in posting schedule, then start drafts from the template library and add them to the queue. See the calendar to move any post to another slot by dragging it.

Check yourself

  1. 1Why should a weekly progress post be paired with another post type?

    On its own it reads like a status report; a decision, lesson, mistake or number gives readers something to learn or react to.

  2. 2What is the "competitor, customer, attacker" test?

    Before posting, ask whether each of them would be fine with what they learn. If an attacker learns something useful, cut it.

  3. 3What makes a posting rhythm sustainable?

    Fixed slots, a fixed post type per slot, and a volume small enough to keep in your worst week.

Academy

Keep reading, free

Tell us a little about you and the rest of this lesson, and every lesson in every course, opens on this device. No account and no card.

What describes you best?
Where do you post today? Pick any that apply.
What do you want most right now?

We use your answers to suggest lessons and to improve the Academy. They are never shown publicly. Read our Privacy Policy.