What did Pieter Levels commit to in 12 startups in 12 months?
To launch one startup a month for a year, and to report on it in public. Pieter Levels, the maker of Nomad List and Remote OK, announced the goal on 1 March 2014 in a post titled I'm Launching 12 Startups in 12 Months. Levels liked the idea of Jennifer Dewalt's 180 websites in 180 days but, unlike Dewalt, could already build things and wrote that the real challenge was to finish and launch them.
The post names two problems. Creative people struggle to finish things, and fear of failure stops them launching what they have built. The plan for each month was to pick an idea, develop it and launch it to press. Levels defined a startup as one that delivers a new product and grows it fast, and set a one-month test: a product that showed no sign of demand within the month counted as failed.
The public part was the progress report at the bottom of the post, which Levels said would be updated with links to each month's startup. The report lists projects including Play My Inbox, Tubelytics, Nomad List, Nomad Jobs, GifBook and Remote OK. The series page dates the first project post, Play My Inbox, to 9 March 2014, and the post on how Remote OK reached number one on Product Hunt to 2 April 2015.
What did Levels say building in public does?
It brings help, accountability and attention, at the cost of stress. Levels' clearest account is the 26 July 2017 post Building a startup in public, about Hoodmaps, a crowdsourced city map. Levels live-streamed the build on Twitch from the first line of code, posted progress on Twitter, then launched on Product Hunt, Hacker News and Reddit. The post calls it Levels' first time building 100% in public and lists what that did:
- Help. Whenever Levels got stuck, 100 people were watching and helping to solve the problem.
- Accountability. With people watching, there was no room to slack off.
- Marketing. Building in public was unusual enough to count as marketing in itself.
- Cost. It was stressful and meant risking failure in front of hundreds of people. Levels also notes some unfair advantages, such as a big Twitter following, but tells readers not to let that stop them.
An earlier post, What we can learn from Stormzy about transparency (16 April 2017), gives the reasoning: show the struggles, admit it's hard, apologise when you get it wrong, and share successes when they come. "People want to ride along with you on your journey," Levels wrote.
How did Levels turn revenue into headlines?
By putting the number in the headline and the breakdown underneath. These four posts were first posted on X and are archived on levels.io with their dates:
| Date | Headline number | What the post broke down |
|---|---|---|
| 5 Jan 2016 | 2015 revenue of $202,785 | One Linode server, one MacBook, no office, hardly any expenses and $0 funding |
| 11 Aug 2017 | $34,733 revenue in the last 30 days | Revenue per product (Nomad List $18,473, Remote OK $12,120, MAKE $4,140) and $3,950 a month of expenses, itemised |
| 26 Feb 2018 | $52,843 in monthly revenue | Revenue per product, including Hoodmaps at $0 a month, and an extrapolated $634,116 a year |
| 30 May 2019 | $1m annual revenue run rate, four years in | $83,333+ a month, 11,996 customers, no ad budget and no funding |
Three habits carry over to outbound, in our reading. The number leads. The breakdown shows where it came from. And the zeros stay in: the February 2018 post lists a product earning nothing next to the ones that pay. The August 2017 post also promised that all revenue would soon be shared live on Levels' sites.
Which outbound numbers should you post every week?
Three: people contacted, replies and meetings booked. A pre-revenue team has no revenue line to post, but it does have activity, and these three show both the effort and the response to it. Define each once and keep the definition, or week-on-week comparisons mean nothing. The definitions below are our suggestion.
| Number | Count this | Don't count |
|---|---|---|
| Sent | New people contacted for the first time this week, on LinkedIn, X and email | Follow-ups, or the same person on a second channel |
| Replies | People who replied, including a polite no | Auto-replies and out-of-office messages |
| Positive replies (optional) | Replies that ask a question, ask for details or agree to talk | Thanks-but-no-thanks |
| Meetings booked | Calls booked this week, whatever date they fall on | Calls you hope to book |
In our reading, Levels' three benefits carry straight over to outbound. Accountability: a Friday post with a zero in the sent column is hard to publish twice, so the sending happens. Help: readers who sell for a living will tell you why an opener isn't landing. Marketing: if your buyers are founders or sales leaders, the people reading your numbers are the people you want to reach. The cost carries over too, because a run of weak weeks is visible to everyone, prospects included.
Work out two rates: replies divided by sent, and meetings divided by sent. Count each reply in the week it arrives, so a Monday reply to last Thursday's message goes into this week. Post bad weeks too. Levels' revenue posts include a product at $0, and a weak week with a clear lesson is more useful to readers than a gap.
A template for posting outbound numbers on X and LinkedIn
Use a short format on X and a longer one on LinkedIn, on the same day every week. We suggest Friday, because most of the week's replies are in by then. Both templates below are examples; replace everything in braces with your own figures. X's developer documentation says a post can contain up to 280 characters and counts every link as 23, so the X version leaves room for a link.
For other numbers-update formats, see our X post templates and LinkedIn post templates. If Siela's Social Media agent publishes for you, set that account to show me every post first, so you check each figure before it goes out.
What should you keep out of a public numbers post?
Anything that identifies a prospect. The people you contacted didn't agree to be part of your story. These rules are ours:
- No names. Not of people, and not of their companies, unless they say yes in writing.
- No screenshots of replies. Blurring often leaves enough to recognise someone. Describe the pattern instead, for example that three people asked about pricing before the demo.
- No rounding up. If you sent 47, write 47. Readers who follow for eight weeks will notice when the maths stops adding up.
- No swapped metrics. If a week is bad, don't switch to profile views or impressions to fill the post. Keep the same three numbers.
- No claims for tools. If you use software to send, say so, but don't credit it with your results unless you can show the before and after.
The point of the post is the same as Levels' revenue headlines: a real number, posted on schedule, that people can follow over time.
Your first Friday numbers post: a checklist
Set it up on Monday and post on Friday. These steps are our suggestion for a founder or a team of up to five.
Write your definitions
Put the four definitions from the table above at the top of a spreadsheet, so next month's count matches this week's.
Log as you go
One row per person: date contacted, channel, replied, positive, meeting booked. Counting at the end of the week from memory is how numbers drift.
Fill in the template
On Friday morning, total the week, work out the two rates and write one line on what you'll change.
Post on both platforms
Use the short version on X and the longer one with the lesson on LinkedIn. Same numbers on both.
Keep going for eight weeks
One post proves nothing; eight show a trend. After a month, write a monthly roll-up with the total as the headline, Levels-style.
For more on growing an audience on X, read our guide for B2B founders on X.
Frequently asked questions
When did Pieter Levels start 12 startups in 12 months?
Pieter Levels announced the challenge on 1 March 2014 in a post on levels.io, with the goal of launching one startup every month for the next 12 months. The post kept a public progress report with links to each project, and the projects it lists include Nomad List and Remote OK.
What did Pieter Levels say about building in public?
In a 26 July 2017 post about building Hoodmaps live on Twitch and Twitter, Levels called it a first time building 100% in public. The benefits listed were 100 people helping when stuck, accountability and built-in marketing; the costs were stress and the risk of failing in front of hundreds of people.
What outbound numbers should a founder post each week?
Post three: new people contacted for the first time that week, people who replied, and meetings booked. Add positive replies if you track them. Keep the definitions fixed so weeks can be compared, and work out two rates, replies divided by sent and meetings divided by sent.
Should you post your outbound numbers when the week is bad?
Yes, in our view. A gap in a weekly series says more than a weak week does. Pieter Levels' February 2018 revenue post listed Hoodmaps at $0 a month next to the products that paid. Post the real numbers, add the one thing you are changing, and keep the same format the following Friday.
Is it OK to share screenshots of prospect replies in a build in public post?
We advise against it. The people you contacted did not agree to appear in your posts, and blurred screenshots often still identify them. Share the pattern instead, such as how many replies asked about pricing, and name a person or company only with their written permission.
Sources
- Pieter Levels: I'm Launching 12 Startups in 12 Months (published 1 Mar 2014; checked 20 Oct 2026)
- Pieter Levels: 12 Startups in 12 Months (series page) (checked 20 Oct 2026)
- Pieter Levels: Building a startup in public, from first line of code to frontpage of Reddit (published 26 Jul 2017)
- Pieter Levels: What we can learn from Stormzy about transparency (published 16 Apr 2017)
- Pieter Levels: My 2015 revenue was $202,785 (published 5 Jan 2016)
- Pieter Levels: Nomad List, Remote OK and MAKE generated $34,733 in revenue last month (published 11 Aug 2017)
- Pieter Levels: I just passed $52,843 in monthly revenue (published 26 Feb 2018)
- Pieter Levels: It took 4 years to reach $1m annual revenue with Nomad List and Remote OK (published 30 May 2019)
- X Developer Platform: Counting characters (checked 20 Oct 2026)