What to post when you have no audience: 30 ideas that work at zero followers
The blank box is the whole problem. You know posting works, you have read that consistency matters, and then you open the app on Tuesday morning and have nothing to say that a stranger would care about. Here are thirty things you can say, grouped by where they come from, each one with a draft attached.
Most content advice assumes the hard part is distribution. At zero followers the hard part is supply. You are not choosing between good ideas, you are staring at nothing, and the standard answer to that is a list of prompts so vague they make it worse. "Share your journey." "Provide value." Nobody has ever opened an app and typed a post because somebody told them to provide value.
Posting in public is one of the four ways to get customers, and it is the slowest of them. It is also the only one that keeps paying after you stop working it, which is why it is worth starting on the day you have no audience rather than the day you have one. If you have not settled on it yet, the Core 4 explains what the four are and what each costs, and the channel quiz gives you a split for your actual situation in about two minutes.
Why the usual advice fails at zero
Advice posts and industry commentary are the two formats every founder reaches for first, and both are the wrong shape when nobody knows you. Advice borrows authority you have not earned. Commentary competes with people who have been at it for six years and have the reach to prove it. You lose both fights, quietly, three weeks in a row, and then conclude that posting does not work for you.
What you do have is the raw material. You are inside a problem that most of your readers have only glanced at. You have a support inbox, a churn number, a bug you shipped on Thursday, and a user who said something odd on the phone. None of that is available to the person with forty thousand followers writing about your market from the outside. It is the one asymmetry running in your favour, and almost nobody uses it.
- 30 posts is the smallest honest test, and the first ten are you working out what you sound like.
- 3 a week for ten weeks gets you there. At 45 minutes a post, replies included, that is about 22 hours in total.
- 22 hours is roughly fifteen minutes a day across a quarter. It is less than one bad week of building the wrong feature.
- Assume one post in five starts a conversation. Thirty posts gives you six. Assume one in five of those buys something and you have one customer, plus five people you can now message directly.
Those conversion assumptions are ours and they are illustrative, not measurements. Put your own in. The point of the sum is that the answer is small on customers and large on contacts, and the contacts are what the next quarter runs on.
Five rules that make any of the thirty work
Skip these and the ideas below turn into the same beige posts as everyone else's. None of them take extra time. Two of them save you time.
- 01
Write to one person, by name, in your head
Pick somebody who replied to you in the last fortnight and write the post to them. Not to an imagined market. The specificity that comes from having one real reader is the thing strangers stop scrolling for, and it is free. You already know what confused them, what they asked twice, and which word they used instead of yours.
- 02
One post, one claim
If you can pull two claims out of a draft, you have two posts and you should ship them on different days. A post that argues three things reads as a summary of your beliefs rather than an argument, and summaries do not get replies. The test: can somebody disagree with this in one sentence? If not, it is not a claim yet.
- 03
Show the artefact
The screenshot, the spreadsheet row, the email that came back, the crossed-out paragraph. An artefact is proof you did the thing, and proof is the only thing you have that a founder with a following does not have more of. Annotate it with one arrow. Two arrows and nobody reads either.
- 04
Publish where you already read
Whichever place you open without meaning to is the place where you understand the norms, recognise the names, and can tell a good post from a bad one. That instinct is worth more than any platform-choice argument. One place, properly, beats four places badly, and cross-posting the same words everywhere is how you learn nothing about any of them.
- 05
Count posts, not followers
The follower number is the one thing on the page you cannot move directly, so tracking it just makes you feel bad on a Tuesday. Count posts published and conversations started. Thirty posts and six conversations is a working quarter even if the follower count did almost nothing, because the six conversations are people you can now message.
The 30 post ideas
Six wells, five ideas each. Every idea has the reason it works and a draft written out, so you can see the shape rather than guess at it. The drafts are for an imaginary product and the numbers inside them are invented on purpose. Yours will not be, and that is the entire difference between a post that lands and a post that reads like a template.
Something you built this week
The build-in-public well. It never runs dry while you are still shipping, and it is the only category where being small is an advantage rather than a thing to hide.
- 01
The thing you shipped and the thing it replaced
Why it works: A launch is an announcement. A swap is a story, because the old version sets it up and the new one lands it.
DraftOur onboarding asked eleven questions. It now asks four. Two of the seven we cut were genuinely useful later and five were us being nosy on day one. Here is the old screen next to the new one.
- 02
The bug that reached a real person
Why it works: Admitting a specific fault costs nothing and buys the credibility that a hundred polished posts cannot.
DraftYesterday our reminder fired at 04:00 for everyone outside Europe. Fourteen people got woken up by a productivity app, which is a special kind of irony. The cause was a timezone stored on the server instead of the device. Fixed, and here is the two-line diff.
- 03
The thing you deleted
Why it works: Removal posts are rare, so they stand out, and they say more about your judgement than any feature announcement.
DraftDeleted the calendar streak. It punished people for one bad Tuesday, and everybody who broke a streak on day nine stopped opening the app entirely. Counting sessions instead of days fixes it, and nobody has broken one since.
- 04
The decision you are still arguing with yourself about
Why it works: An open question invites answers, and answers are the whole reason to publish when nobody is reading yet.
DraftShould a rest day break a hundred-day streak? If it does, the honest people quit. If it does not, the number means less. I have written both versions and I cannot pick. Which one would keep you going on day 41?
- 05
One screenshot with one arrow on it
Why it works: An annotated image stops a scroll better than a paragraph describing the same thing, and it takes four minutes.
DraftThis is the empty state nobody saw for six months, because you only reach it after archiving everything. Arrow points at the line we spent a whole afternoon on. Worth it.
Something a person actually said
The richest well and the least used. Every reply, support message and awkward sales call is raw material you have already paid for.
- 06
The exact sentence a user said, quoted verbatim
Why it works: Their words beat your summary of their words, and other people recognise themselves in a real sentence.
DraftA user last week: "I do not want another list. I want to be told the one thing." That sentence deleted two features off our roadmap and rewrote the top of the home page.
- 07
The question you were asked twice
Why it works: Asked twice means dozens wondered and did not ask, so answering in public turns one reply into a hundred.
DraftTwo people this week asked whether the app sends anything on their behalf. It does not, and here is exactly what it does instead. If two asked, more assumed the worse answer and left.
- 08
The objection you keep losing to
Why it works: Naming your weakest point publicly gets you the counter-argument for free, usually from someone who has solved it.
Draft"I already have a to-do app." I have lost this one four times and my answer is still not good enough. Here is what I say now, and I would genuinely like a better version of it.
- 09
A support thread rewritten as a public answer
Why it works: The thinking is already done and paid for, so the post costs you the ten minutes it takes to strip the names out.
DraftSomeone asked why we do not import their whole history on day one. The long answer involves rate limits and a very boring queue. The short answer is that the first day should be fast, not complete.
- 10
The customer you lost, and why
Why it works: Churn posts are uncomfortable, specific and rare, which is exactly the combination people forward to a friend.
DraftLost a paying user in month two. Reason, in their words: "It works, I just stopped opening it." That is not a pricing problem or a feature problem. It is a habit problem, and it is mine to fix.
A number you actually have
Numbers travel further than opinions and get quoted back at you months later. The rule is that every number arrives with its denominator, otherwise it is decoration.
- 11
One number with the denominator attached
Why it works: A number you can check is a number people repeat. A percentage with no base is noise and readers can smell it.
DraftI hand-wrote 41 first emails last month. Six replied, two booked a call, one paid. That is 14.6% replies on a list I built by reading job ads, and it took eleven hours in total.
- 12
The number that went the wrong way
Why it works: Bad numbers published on purpose are the cheapest trust you will ever buy, and almost nobody does it.
DraftWeek-two retention fell from 38% to 29% after the redesign. I have a theory involving the new home screen and no proof yet. Posting it now so I cannot quietly reframe it later.
- 13
The real cost of one thing you do
Why it works: Cost-per-unit posts are practical, they help the reader plan, and they are impossible to fake convincingly.
DraftCold email costs me about 80 euros a month in tooling plus roughly nine hours. At a 3% reply rate on 600 sends that is 18 conversations, so a bit under five euros and thirty minutes per conversation.
- 14
A before and after with a date on it
Why it works: Two states plus a date is the smallest complete story, and the date makes it checkable rather than a claim.
Draft12 May: signup took nine screens and 4m10s. 8 August: three screens, 51 seconds. Nothing clever happened in between. We just removed things one at a time and measured after each one.
- 15
The arithmetic behind your own goal
Why it works: Showing your sums invites people to check them, and a corrected sum is a conversation with someone who cares.
DraftA hundred customers in a hundred days at a 4% close rate means 2,500 conversations, which is 25 a day, which is not a thing one person does. So the goal is wrong or the rate is. I did the sum in the 100 in 100 calculator and then rewrote the target.
Something you know that your reader does not
The teaching well. You have been inside a problem for months, which means the boring middle of your process is genuinely new to almost everyone reading.
- 16
The checklist you actually run, boring steps included
Why it works: Real checklists have unglamorous steps in them, and those steps are the proof that this is not a made-up list.
DraftBefore every release: run the migration on a copy, open the app on the oldest phone in the drawer, send myself the welcome email, check the App Store screenshot still matches the screen. Step three has caught more problems than the tests.
- 17
The mistake you made twice
Why it works: Once is bad luck and reads as a humblebrag. Twice is a pattern, and a pattern is useful to somebody else.
DraftI have now twice shipped a pricing change without checking what existing subscribers see. Both times somebody found out from the website before they heard from me. Here is the four-line checklist that exists because of it.
- 18
A teardown of your own page
Why it works: Critiquing your own work gives you the teardown format without the risk of publicly embarrassing a stranger.
DraftMy home page had a headline that described a category, not a benefit. Three people read it and could not tell me what the app does. Here it is with the rewrite next to it and the reason for each cut.
- 19
How you do one specific thing, in numbered steps
Why it works: Steps embed a question that keeps people reading to the end, and the end is where your ask lives.
DraftHow I find 50 prospects in 40 minutes: filter job ads posted in the last 14 days, take the company name, find the person whose title contains the word from the ad, verify one address in ten by hand, stop at 50 even if it is going well.
- 20
A definition nobody has written down properly
Why it works: Clean definitions get screenshotted, quoted and cited, including by the machines people now ask instead of Google.
DraftWarm outreach is messaging someone who already knows you, one person at a time, with a specific ask. That last clause is the part everyone drops, and it is why "just catching up" messages go nowhere.
Something somebody else made
The borrowed-attention well. Reacting to work that already has readers is the fastest way to be seen by people who have never heard of you, provided you add something.
- 21
The reply that should have been a post
Why it works: You already wrote it, it already got a reaction in miniature, and almost nobody saw it in a comment thread.
DraftSomebody asked in a thread how to price a first product. I wrote four paragraphs in a reply that eleven people saw. Here it is properly, with the bit I got wrong the first time corrected.
- 22
Disagreeing with named advice, generously
Why it works: Conflict earns attention, and naming the source keeps it honest instead of vague sniping at an invisible enemy.
DraftHormozi says give distribution four hours every morning before anything else. I ship the product, answer the support email, and have about ninety minutes. The principle survives the compression. The schedule does not.
- 23
A book note with a page number, applied to today
Why it works: The book brings its own credibility, and the application to your Tuesday is the part only you can write.
DraftOgilvy: five times as many people read the headline as read the body. I spent eleven minutes on a blog post headline and four hours on the post. The ratio is backwards and I am fixing it this week.
- 24
A full answer to a public question, posted twice
Why it works: Answering properly in the forum earns you the right to post the same answer on your own account the next day.
DraftSomeone on a forum asked how many cold emails before you know if it works. Answer: about fifty of one version, because below that a reply rate is arithmetic noise. Here is the working.
- 25
Specific credit to a specific person
Why it works: Naming a real thing somebody did is memorable, unlike praise, and it starts a conversation with them too.
DraftThe onboarding fix came from a stranger who replied to a post with one line: "you are asking me things you could observe." I have not stopped thinking about it. Thank you, whoever you are.
Posts that ask for something
The reply well. Most posting advice is about reach, and reach does not help you at zero. These are written to produce a reply, which is the only currency that converts.
- 26
The very specific ask
Why it works: A vague ask gets sympathy. A precise one gets forwarded, because the reader can picture exactly who fits.
DraftLooking for five people who run recruitment at a company of 10 to 50 and still send candidate updates by hand. Twenty minutes, no pitch, and I will send you what I learn from all five.
- 27
The artefact in exchange for a reply
Why it works: A real spreadsheet or checklist is worth a reply, and a reply moves that person from stranger to contact.
DraftI built a spreadsheet that works out how many cold emails you need for a given customer target. Reply with the word "sums" and I will send it. No list, no sequence, I will just send the file.
- 28
The "tell me I am wrong" post
Why it works: People who will never like a post will happily correct one, and a correction is a conversation you can continue.
DraftI think paid ads are useless before you have a message that already works organically. I have spent 258 euros finding that out and got 21 signups and no customers. Convince me the test was the problem and not the timing.
- 29
The invitation to be in the thing
Why it works: Being included is a stronger offer than being sold to, and early users who feel like co-authors stay.
DraftWriting up the first ten people who used this, with their numbers and their names if they want. Two slots left. You need about thirty minutes and a willingness to say what did not work.
- 30
Tomorrow morning, in advance
Why it works: Saying the plan out loud creates a small public cost for skipping it, and the follow-up post writes itself.
Draft06:30 tomorrow: forty cold emails to operations leads at logistics companies, same message as today, no rewriting. Reporting the reply rate on Friday whatever it is.
Which well to draw from on a given Tuesday
You do not pick an idea from a list of thirty at half past six in the morning. You look at what happened since the last post and let that decide. Read the first column, take the row, write the post. The whole choice should take under a minute, because the choosing is where the session dies.
| If this happened | Draw from | Shape of the post | Minutes |
|---|---|---|---|
| You shipped, fixed or deleted something | A. Something you built | Screenshot plus three sentences | 15 to 25 |
| Somebody replied, complained or asked | B. Something a person said | One quote plus what you changed | 10 to 20 |
| You have a week of data in front of you | C. A number you have | Number, denominator, one conclusion | 20 to 30 |
| Quiet week, nothing happened | D. Something you know | Checklist, steps or a definition | 30 to 45 |
| You read or watched something good | E. Somebody else made it | Quote, then your disagreement | 15 to 25 |
| You need conversations, not reach | F. Posts that ask | One ask, on its own line | 8 to 15 |
Notice that the quiet week routes to the slowest column. That is deliberate. When nothing happened you have to make something out of what you already know, and those posts take longer, which is exactly why founders skip the quiet week and then have a two-week gap in the record. Put the teaching posts on the calendar for the weeks you can already see are going to be dull.
The replies are the channel
Twenty minutes answering comments will usually start more conversations than the post did. This is the part founders treat as optional and it is the part that converts, because a reply is warm outreach that arrived without you having to be brave first. Somebody who answered your post has already done the hard bit. All that is left is a specific question back.
Keep a list of every person who replies to anything, with the date and what they said. After ten weeks that list is thirty or forty names of people who know who you are, and it is the warm list you did not have in January. Posting is slow at producing customers and quick at producing contacts, and the contacts are the actual output.
This is also where the two failure modes live. The first is quitting at week nine, a fortnight before it turns, which is common enough that it has its own post-mortem. The second is treating publishing as a broadcast. Broadcasters get reach. People who treat their posts as an inbox get customers.
Fitting thirty posts into mornings you already do not have
Three posts a week does not mean three writing sessions. It means one session where you capture, and three where you publish. Keep a running note with one line per thing that happens: the bug, the sentence a user said, the number that moved, the thing you deleted. By Thursday there are six lines in it and picking one takes ten seconds.
The capture habit is what makes this survivable at ninety minutes a day. Without it you spend the first twenty minutes of every session trying to remember your own week, which is the most expensive way to run out of time. The daily routine covers how the block sits next to the other three ways of getting customers, and how to keep one morning a week reserved for this without the other three quietly eating it.
One last thing, and it is the least fun sentence here. Ten of these thirty ideas will not suit you. Some people cannot bring themselves to publish a churn number and some people write a terrible checklist. That is fine. Twenty usable ideas at three a week is still seven weeks of supply, and by week seven you will have invented four of your own, because the well that actually refills is the work itself.
Questions founders actually ask
- What should I post if I have no audience?
- Post the things only you could post. What you shipped this week and what it replaced. The exact sentence a user said to you. One number with its denominator attached. The decision you are still arguing with yourself about. Avoid general advice and industry commentary, because both borrow authority you have not built yet, and a stranger has no reason to take marketing advice from an account with forty followers. Specific beats broad at every audience size, and it beats broad hardest at zero.
- What is build in public content?
- Publishing the work as it happens rather than after it succeeds: what you shipped, what broke, what a number did this week, what you decided and why. It works for a founder with no audience because the material is free, it renews itself every week you keep building, and nobody else can copy it. The failure mode is posting progress with no stake in it. A changelog entry is not a post. A changelog entry plus the reason you chose that over the other thing is.
- How often should I post when nobody is reading?
- Three a week for ten weeks. That is thirty posts, which is the smallest number that tells you anything, and at 45 minutes each it costs about 22 hours in total. Daily posting at zero audience mostly buys you a faster route to quitting, because the quality drops before the reach arrives. Three a week is slow enough to keep each post specific and frequent enough that you get properly fluent before the ten weeks are up.
- Where should I post if I have no followers?
- Wherever you already read without meaning to. You know the norms there, you recognise the names, and you can tell a good post from a bad one, which is the judgement that actually decides whether posting works for you. One platform properly beats four badly. Communities where your customers already answer each other, such as a subreddit or a niche forum, usually pay back sooner than a personal feed, because the readership is there whether or not anybody follows you.
- How long before posting produces customers?
- Six to twelve weeks, or roughly thirty posts, before you can read the result honestly. Posting is the slowest of the four ways to get customers and the only one that keeps working after you stop. A post from March still brings people in September, while a cold email from March is in the bin. That is the whole argument for publishing, and the ten quiet weeks in front of it are the whole argument against.
- What do I post if my product is not launched yet?
- The problem, the research and the decisions. Post the conversations you are having with people who have the problem, the thing you decided not to build, the competitor you tried for a week and what annoyed you about it, and the number you are trying to move. Pre-launch is the easiest time to ask for replies, because you are asking for opinions rather than money, and the people who answer become the warm list you will need on launch day.
- Is it worth posting if only forty people see it?
- Yes, if you are counting conversations rather than views. Forty of the right people is a room, and one reply from that room is worth more than a thousand impressions from strangers who will never buy. The posts to prioritise at that size are the ones written to produce a reply: a specific ask, an artefact offered in exchange for a message, or a claim somebody will want to correct.
- Should I use AI to write my posts?
- Use it for the shape and the tidying, never for the substance. The parts that make a post work at zero audience are the artefact, the quote, the number and the mistake, and a model does not have any of those. It has not seen your support inbox or your churn. Drafting the structure with AI and filling the specifics by hand is the combination that holds up. If the finished post could have been written by somebody who has never used your product, delete it.