Who we feature
The people behind the numbers. Every founder we've sat down with, and the stories they told us straight.
A composite founder diary. Four contractors, one policy bought only to satisfy a customer's insurance rider, and a premium audit that asked for certificates of insurance rather than contracts. What I found reading the California Labor Code.
OSI filed $1,093,077 for 2024 and has self-published $667,000 for 2025, a year the IRS does not have. Its accounts and its tax return disagree, and the gap reconciles exactly.
A composite founder diary. Twenty one Forms I-9, sixteen flagged, and not one unauthorised worker among them. What I learned reading 8 U.S.C. 1324a and its penalty table: the good faith defence runs one way only, and the clock that sets the fine is the one I do not control.
OTF's audited revenue hit a record $53,115,458 in the year ended 30 September 2025. In the same year its cash halved to $25,283,707 and it closed with net assets of minus $123,660. Six grant years paid for one fiscal year.
A composite founder diary. At 65K MRR a banking partner asked whether client funds ever touched our balance. Nine days in the statutes turned up a federal felony that says your knowledge is irrelevant, an exemption written for somebody else, and a civil penalty of 10,556 dollars a day that the statute itself prints as 5,000.
The Tor Project's record 2025 revenue is $12,279,993, up 68.5%. The government money went the other way: down 47.5% since 2023, to 22.62% of revenue on Tor's own method, from 53.5% three years earlier.
A composite founder diary. $58,400 left on an instruction I never gave, and the statute that decides who eats it never once asks whether I was careless. Read end to end from UCC Article 4A at $64K MRR.
The headline is correct and nearly useless. $2,248,286 of the GNOME Foundation's record 2025 revenue arrived as cryptocurrency from a single donor, was sold at a $39,828 loss, and is 85% locked behind donor restrictions.
A composite founder diary. A contractor invoiced at $26 an hour for three years asked one question, and the number everyone quotes turned out to be the escapable one. Read end to end from the statute at $63K MRR.
A composite founder diary. Our SMS vendor handled every STOP perfectly, and a customer who typed a plain sentence instead was texted again nine days later. What the statute and the regulation actually require, read end to end at $62K MRR.
The DSF raised $296,584 in 2024 and spent $277,682. The structure is the story: 97.60% of income is donations, 78.56% of spending is one contracted programme, 84.30% is booked abroad, and there are no employees at all.
A composite founder diary. A preservation demand landed 22 days after my 90-day retention rule had already deleted the exact window it asked for. Rule 37(e) has three conjunctive conditions, and most of the panic failed at the first one.
Ruby Central's Form 990 reports $3,107,120 of revenue in 2024, the best year in its history, and a $160,579 surplus. Sixteen months later its board said the organisation was in real financial jeopardy. The number that reconciles the two is the reserve: 11.5 days at the end of 2023 and 44.8 days at the end of 2024.
A composite founder diary. I tried to write off $18,470 of stale customer credits and found that Delaware voids the expiry clause I had written myself, that the $50 threshold I relied on governs a letter rather than a liability, and that two states' notice windows never overlap.
Software in the Public Interest reported $685,055 of revenue and $925,737 of expenses in 2024, a third consecutive deficit. Its audited statements report a different revenue figure for the same year, and the Part X split shows the shortfall belongs to the projects it sponsors, not to SPI, whose own reserves rose 46.22% while Debian's balance fell 63.40%.
A composite founder diary. I sent a $21,600 collection letter and got back a sentence about a certificate I did not hold. California bars an unregistered company from suing on that business while deeming it to consent to being sued. The threshold that decided it was $177,000, not the $757,070 every guide quotes.
NumFOCUS reported $8,376,226 of revenue and $11,090,430 of expenses in 2024, its second consecutive deficit. The two returns carrying those figures are not on its own public records page, and the line that would tell you how much of the money it can actually spend is left blank on both.
A composite founder diary. A customer filed Chapter 11 and their counsel demanded back $8,550 of ordinary subscription payments. The statute prints a $5,000 floor; the operative floor is $8,575. I was twenty-five dollars under a line I did not know existed, and the date that decided it was not mine.
Two of Conservancy's Form 990s are filed with the IRS and published nowhere. Revenue fell to USD 2,487,144 in the year to February 2024 and recovered to USD 3,214,723 in the year to February 2025, while unrestricted reserves more than halved. Three filings, 63 arithmetic checks, and the one dollar that does not reconcile.
A composite founder diary. A customer's procurement form asked whether we were a high-risk AI system under the EU AI Act. The headline penalty is EUR 35 million or 7% of turnover. Article 99(6) quietly inverts that for small companies, and the exemption I chased turned out to carry its own registration duty.
Eleven years of ISRG's tax filings and seven annual reports, assembled into one table and audited three ways. Revenue collapsed 36% in 2023 and rebounded 85% to a record in 2024, the question has two correct answers 733,260 apart, and the filing that carries the answer has no structured record anywhere.
A composite founder diary. A 190 line fork of an AGPL library, shipped and forgotten for fourteen months, drew a demand for the complete source of a 148,000 line platform. The licence asked for something 779 times smaller, and the letter turned out to be the only route to permanent reinstatement.
Eleven years of the OpenStreetMap Foundation's filed accounts, assembled into one table and checked three ways. 2024 was its best year for income and its worst for losses, corporate membership now pays for half of it, and the most recent accounts are missing from the page that lists the accounts.
A composite founder diary. A dormant $49 a month account resolved to a sanctioned jurisdiction after 26 successful charges. The statutory maximum reads as $9.8 million, the real base is $26,000 or $637, and the 40.8x gap turns on one condition buried in OFAC's own definition of voluntary.
QGIS.ORG booked EUR 428,260.48 in 2024 against EUR 370,906.17 in 2023. All twenty-two income and expense columns across eleven financial reports foot exactly. The 2020 report is the one year whose bottom line does not, and its 2026 budget counts EUR 80,000 of reserves as income.
A composite founder diary. A demand letter valued four years of California subscription revenue at $284,610 and called it an unconditional gift. The statute said something narrower, and something worse: the 2024 amendments reached the whole existing book, not just new signups.
Blender Foundation booked EUR 3,106,361 in 2024, up 21.7%, and ran its first deficit in the published record. Twelve of the fourteen financial columns across five annual reports add up exactly. The two that do not are both in the newest one.
A composite founder diary. A Delaware annual report notice arrived for $102,215 against $52K MRR. The real figure was $850, and the notice was not an error: Title 8 section 503(b) makes the Authorized Shares Method the statutory default until you report your issued shares and gross assets at the time of filing.
The Godot Foundation's two financial reports, read together. Revenue up 62.6%, surplus down 71.3%, and a note that names four revenue categories while publishing a value for none of them.
A composite founder diary. My Stripe 1099-K reported $561,300.00 in a year my bank received $528,217.80. The form was right: 26 CFR 1.6050W-1(a)(6) defines the reportable amount as gross 'without regard to any adjustments for ... fees, refunded amounts or any other amounts.' Here is the $33,082.20 reconciliation, why the $20,000 threshold everyone quotes does not apply to direct card payments, and what typing Box 1a straight onto a return would have cost.
The Document Foundation raised EUR 1,976,825 in donations in 2025, up 42.1%, held spending flat and banked a EUR 718,654 surplus, its largest ever. Weeks later LibreOffice shipped its first in-product donation banner. Here is the full 11-year series from the foundation's own ledgers.
A composite founder diary (2026): at $49K MRR a letter of intent valued my company at four times ARR, and eleven weeks later a quality of earnings report put the price at $1,162,960 instead of $2,352,000. Why ARR overstated my trailing twelve month revenue by $70,700, why the two year recurrence test killed most of my add-back list, and the line I had never once thought about: the market salary I never paid myself, which cost more than the revenue restatement, the add-back fight and the deferred revenue combined.
Mentimeter reported SEK 598.0 million of net revenue in 2025 and its first profit since 2020. It also produced positive free cash flow in four of the five loss-making years before that. Here is the full audited series, assembled from the filings for the first time.
A composite founder diary (2026): at $48K MRR a competitor cloned eleven of the fourteen sentences on my landing page, comma splice included. A DMCA takedown notice got the page down in six days, and a counter notification put it back on the table on day nineteen. Why the CDN was the wrong recipient, why copying expression is actionable when copying features is not, and the arithmetic nobody publishes: ten to fourteen business days to file a lawsuit against a copyright registration that was averaging about four months.
Scoro's filed revenue ran EUR 10.43M (2022) to EUR 16.62M (2025) and the group booked its first ever profit in 2025, EUR 374,000. The Estonian and UK filings reconcile to the euro, the register leaves the profit cell blank, the UK copy is an unreadable scan, and both aggregators on page one are high by 20% to 30%.
Four early stories in this archive carried a publication date from before the domain existed. Here is what was wrong, what the dates are now, and the two free lookups that let you check it without taking our word for it.
Pipedrive OU's audited revenue ran EUR 160.42M (2022) to EUR 236.80M (2025) while net profit fell from EUR 46.53M to EUR 25.62M. The Estonian filings show the profit drop is mostly a first-ever dividend tax, that 2024's record profit year burned EUR 28.9M of cash, and that the famous 2022 growth spike was a transfer pricing change.
A composite founder diary (2026): at $47K MRR an enterprise customer's procurement team invoked section 8.2 and claimed a service credit for 47 minutes of downtime. Why their number and my number were different lengths and theirs was the one that counted, why the same 42 minute outage breaches in February and clears in July, why the $240 credit was a seventeenth of what the month cost, and the three clauses the large published SLAs carry that mine did not.
Toggl OU's audited revenue ran EUR 14.99M (2022) to EUR 18.03M (2025) while net profit fell from EUR 3.25M to EUR 1.39M. The Estonian filings show why that profit drop is an accounting artefact, why all of 2025's growth came from a line that is not the app, and why the three aggregators on page one are out by up to 17x.
A composite founder diary (2026): at $45K MRR a twelve page ADA demand letter arrived by certified mail alleging eleven barriers and giving fourteen days to respond. Why a demand letter is not a website accessibility lawsuit and what that changes, why the published filing counts of 3,117 and 4,928 describe a category I was not in, the overlay widget I nearly bought at 11pm, the nine defects behind the eleven allegations, and what the whole month actually cost.
Mullvad VPN AB's audited turnover ran SEK 40.7M (2020) to SEK 314.8M (2024), a 7.7x rise on a price that has not moved since 2009. The Swedish filings also show why page-one aggregators are out by a factor of eight, and what happened after the company removed recurring subscriptions in June 2022.
A composite founder diary (2026): at $43K MRR a sixteen-word email asking for "everything you have on me" turned out to be a valid data subject access request with a one month deadline that had already been running for nine days. What Article 15 actually requires, the eleven systems holding the data, why nine of the fourteen hours went on redaction rather than export, and the one scoping lever in Recital 63 that halved the work.
Ecosia's total income ran EUR 1.39M (2015) to EUR 43.24M (2025), with EUR 22.48M in H1 2026. Summing eleven years of its own monthly reports reveals a three-year plateau and the only decline in company history, plus an aggregator estimate off by more than half.
A composite founder diary (2026): at $39K MRR my password resets, receipts and failed-payment notices quietly started landing in spam. The cause was one launch email sent to 6,240 people in a single afternoon, which pushed me over the 5,000-messages-per-day line where Gmail and Outlook both apply strict bulk sender rules. Why 22 complaints was enough, why Google Postmaster Tools shows a solo operator nothing until the day they break something, and what the fix actually cost.
Games on Steam grossed roughly $5.5B (2017), ~$15.4B (2024), ~$17B to ~$20B (2025) and $11.1B in H1 2026 alone. Valve's own revenue was $4B+ from Steam (2025), ~$5B (2023) and ~$2.21B (2021). A source-typed ledger separating storefront volume from what Valve keeps, why the viral $50 million per employee figure is wrong by roughly four times, and what 79 people running Steam actually tells operators.
A composite founder diary (2026): at $41K MRR I found out I had crossed sales tax registration thresholds in five US states without noticing, because a monthly subscription bills each customer twelve times a year and the 200-transaction test counts invoices, not dollars. What economic nexus is, why subscriptions trip it at a third of the revenue you would expect, why uncollected tax comes out of your margin instead of your customers' pockets, how a voluntary disclosure agreement capped the damage, and why January 1, 2027 is on the calendar for anyone selling into California.
Epic Systems' revenue ran $1.2B (2012), $3.3B (2020), $4.6B (2022), $4.9B (2023) and $5.7B (2024), roughly 14 percent compound growth for twelve years on $70,000 of founding capital and no outside investment. A source-typed ledger, how an EHR vendor with 42 percent share reaches $5.7B, the one customer it has ever lost, and the three things that make the story less repeatable than it reads.
Grammarly's ARR runs $350M (2021) to more than $700M (May 2025), but a second series of much smaller figures circulates just as widely, and Google's own AI Overview quotes both without reconciling them. A sourced ledger, the accounting reason the numbers disagree, and the eight unfunded years the revenue figure hides.
JetBrains publishes growth rates but never a total. Its Czech filings show revenue climbing from 3.03bn CZK (2015) to 15.07bn CZK (2024), about $649M, while operating margin collapsed from 46.3% to 13.6%. A sourced ledger of the biggest software company that never took a cent of outside money.
Proton doesn't publish revenue, so 2024 to 2025 estimates run ~$97.5M (GetLatka) to $105M (ZoomInfo), up from ~$70M ARR in 2022. A sourced ledger of the Swiss privacy company that hit 100M accounts on roughly $2.5M ever raised, then gave itself to a non-profit foundation.
Kagi doesn't publish revenue, so estimates for 2026 run from ~$7M (Growjo) to $9.1M (ZoomInfo). A sourced ledger of what the ad-free, user-funded search engine actually discloses: 50,000 paying members, profitable since 2024, and zero venture capital.
Yoast hit ~$12M ARR in 2019 with zero funding, then sold to PE-backed Newfold in 2021 and stopped disclosing numbers. A sourced, year-tagged ledger, why the 2026 estimates run from $12.4M to $100M, and the operator lesson in why a healthy bootstrapped company sells anyway.
A browser-based Photoshop alternative built by one person with zero funding. Founder-stated revenue hit ~$250K (2019) and ~$1M (2021); third-party estimates now put it near $2.8M to $3M. A sourced, year-tagged ledger and what operators can copy.
Kinsta is private and 100% bootstrapped, so its revenue is genuinely disputed: third-party estimates run from under $20M to as high as $100M. A sourced, year-tagged ledger, and why a company that took $0 in outside money will never let you pin the number down.
Elementor is private, so what does the most-installed WordPress page builder actually earn? Third-party estimates run from $53M to $250M, a near-5x spread. A sourced ledger, why the numbers disagree so wildly, and the stray $761M figure you should ignore.
Kit, formerly ConvertKit, is bootstrapped and private, so what does it earn? Third-party estimates put 2024 revenue near $43.8M on zero capital raised. A sourced ledger from the $1,337 month that nearly ended it to the $200M offer Nathan Barry turned down.
Wistia is private and bootstrapped, so estimates of its revenue swing from $24M to $67M. But the number that really explains it is the $17.3M of debt the founders took on in 2017 to buy out investors and force themselves profitable. A sourced ledger and the operator lesson.
Ahrefs is private and bootstrapped, so what does it actually earn? Credible estimates put revenue near $100M to $149M a year (2023 to 2024) with zero outside investors, and the company's own blog disclosed ~$257M over three years. A sourced ledger, plus the revenue-per-employee number that really explains it.
Tailwind CSS is free, so where did the money come from, and where did it go? Tailwind Labs crossed ~$2M in revenue by 2020, then in early 2026 founder Adam Wathan said revenue was down close to 80% as AI gutted docs traffic. A sourced ledger on what a bootstrapped open-source business actually earns.
A composite founder diary (2026): at $44K MRR my blended monthly churn had been flat for a year, so I finally built a cohort retention grid. It showed my newest cohorts leaving at nearly double the rate of my loyal old customers, who were propping up the average. What cohort analysis is, why logo retention and revenue retention tell different stories, what SaaS Capital and ChartMogul data say about net revenue retention, and why I now judge the business on the leading edge instead of the mean.
Calendly's revenue estimates range from $70M to $349M for the same year. A source-typed ledger reconciles the conflict and traces the real $270M-plus ARR back to founder Tope Awotona's $200,000 personal bet. (2026)
A composite founder diary (2026): the month I rebuilt my pricing page around annual billing at $32K MRR. Why monthly billing was quietly leaking customers, what ChartMogul's 2024 data shows about annual retention (62% versus 41% for monthly), the discount and cash-timing tradeoffs, and why I defaulted new signups to annual instead of forcing everyone onto it.
HighLevel, the private company behind GoHighLevel, never confirms its revenue, so estimates scatter from $82.7M (GetLatka) to $311.5M (ZoomInfo). A source-typed ledger, plus why the '$1 billion' everyone quotes is customer revenue, not the company's.
A composite founder diary (2026): the month I deleted two of my four pricing plans at $37K MRR. Why a 'Business' tier only 6% of customers ever chose was taxing everyone, what the research says about three tiers (and why I went to two), the scary first-week conversion dip, and how simplifying took trial-to-paid from 3.1% to 4.4% and MRR from $37K to $40K without new traffic.
37signals, the private company behind Basecamp, never publishes its revenue, so estimates scatter from $25M to $280M. A source-typed ledger, plus the one 2006 Bezos check that complicates the 'never took a dime' legend.
A composite founder diary (2026): the month I broke a five-month plateau at $29K MRR by repositioning a horizontal 'scheduling for everyone' tool into a product built for one vertical. The panic of a shrinking funnel, the six-week trial dip, and why narrowing took MRR from $29K to $41K without a single new marketing tactic.
1Password says it passed $400M ARR while free cash-flow positive, yet trackers quote everything from $200M to $782M. A source-typed ledger, plus the 13-year bootstrap that ended in a $920M raise.
In 2026, the endpoint you optimize for growth is the same one fraudsters optimize for abuse. A composite founder diary about the morning a card-testing bot turned a frictionless signup form at $35K MRR into a stolen-card validation service, and the 48-hour scramble to shut it down.
Mailchimp hit about $800M in revenue with zero venture capital, sold to Intuit for roughly $12B in 2021, and in May 2026 Intuit confirmed its revenue is now declining. A source-typed ledger, and what actually transfers.
In 2026, the fastest way to lose your biggest customer is a backup you never tested. A composite founder diary about the night a botched production migration erased roughly 22 hours of customer data at $40K MRR, when the newer backup had been failing silently for months and the freshest clean copy was almost a day old. It covers what was lost, the email that had to go out, the agency that churned, and the restore-drill and point-in-time-recovery changes that cut the recovery point objective from about a day to five minutes.
Craigslist has never disclosed its finances, so every revenue figure is an estimate. The famous $694M stat is a frozen 2018 peak; AIM Group data shows it has fallen to about $302M by 2024. Here is the sourced ledger and the operator lesson.
In 2026, one four-sentence email can undo a year of growth. A composite founder diary about the month a core integration partner raised its API price roughly 10x and gutted its free tier overnight at $46K MRR, echoing the real 2023 Reddit and Twitter/X API shocks: the triage, the customer email, and the re-architecture that cut our single-vendor dependency.
Midjourney has never disclosed revenue, so 2026 estimates run from $150M to $500M ARR. Here is the sourced ledger, the pricing math, and the number that actually holds.
At $42K MRR I finally added up what an AppSumo lifetime deal I ran years earlier was really costing me: about 1,150 active lifetime accounts paying $0 a month while eating support and server costs. A 2026 founder diary on the long tail of a lifetime deal, and how I cleaned it up.
Obsidian has never disclosed revenue, so 2026 estimates run from about $2M to $25M. Here is the sourced ledger, the pricing math, and the number that holds.
At $38K MRR my revenue looked healthy, but my gross margin was only about 62%, well under the 75% SaaS benchmark, mostly because of AI feature costs. Here is the month of margin surgery that took it to 76% without touching what customers valued.
Lemon Squeezy hit $1M ARR in nine months, turned down a $50M Series A, then sold to Stripe for an undisclosed sum. An honest, source-typed revenue ledger, and why the disclosed number is the least interesting part.
At $36K MRR I realized one customer was nearly 40% of my revenue. Why single-customer concentration is dangerous even when they stay, and the boring playbook that cut it to under 30%.
ClickFunnels revenue peaked at a reported $265.3M in 2023, then fell to about $170M in 2024. An honest, source-typed ledger, plus the raise it never announced.
At $34K MRR, running solo, a 214-question security questionnaire and a SOC 2 ask nearly froze my biggest deal. The interim packet that closed it, and what it really cost.
Calendly is private and stopped confirming its revenue, so third-party estimates now range from about $70M to $349M. Here is the source-typed ledger, the bootstrapped-then-$3B-raise story, and what actually transfers to operators.
At $33K MRR an agency offered to white-label and resell my product, projecting twelve to fifteen thousand in new MRR. A composite founder diary on why I said no, the discount math the pitch hid, and the referral deal I offered instead.
PostHog publishes its salaries, strategy, and board decks, yet keeps its revenue deliberately vague. The company says 'low $10s of millions'; analyst Sacra estimates ~$57.5M ARR (Feb 2026); Stripe says 'on pace for $100M.' A sourced, typed ledger of what PostHog actually earns, and why the most open company in dev tools blurs its own top line.
At $31K MRR I stopped burying cancellation and stopped fighting refunds. The chargebacks that friction was quietly manufacturing fell away, and MRR kept climbing.
Zapier's revenue is estimated near $310 million a year, reached on a single $1.2 to $1.4 million seed and profitability since 2014, then valued at $5 billion via a secondary sale. A source-typed ledger of what is confirmed, what is estimated, and what transfers.
At $10K MRR my growth stalled, and it was not the top of the funnel. Only about 22% of new signups ever reached value. A composite founder diary on rewriting onboarding around one activation milestone, and the eight weeks that followed.
Zoho crossed about Rs 12,313 crore (roughly $1.4 billion) in FY25 revenue, up 17.8%, having never raised a cent of venture capital. A source-typed ledger of how, and what actually transfers.
At $27K MRR one account offered $45,000 for a custom build plus $3,000 a month, if I would build a bespoke module only they would use. A composite founder diary on why I said no, the arithmetic that changed my mind, and the roadmap test I use now.
Todoist has crossed $100M in cumulative revenue, fully bootstrapped, and its founder says the number is the boring part. A dated, source-typed revenue ledger for Doist, and what actually transfers from a 17-year customer-funded climb.
The month I crossed $23K MRR, a card-testing attack slipped a few hundred fraudulent charges past my checkout. Weeks later the disputes rolled in, my dispute rate crossed 0.75%, and my processor froze payouts. A composite founder diary on chargebacks, watchlists, and frozen cash.
Ahrefs is a fully bootstrapped, nine-figure SEO business, but it has never published an official revenue figure. The one hard number it let slip, ~$257M, and the operator lesson behind it.
The month I crossed $26K MRR, my Firebase bill jumped almost six times with no new growth. A composite founder diary on Firestore read costs, migrating the hot paths to Postgres, and why the pattern, not the tool, was the real problem.
Buttondown's founder Justin Duke is radically transparent about growth and profitability, but stopped publishing his revenue. The best estimate is roughly $392K a year and rising 61%. Inside the bootstrapped newsletter tool that hides its own number.
Everyone asked for it. Almost nobody used it. The month I reached $14K MRR, I killed the most-requested feature on my roadmap and learned the difference between what customers ask for and what they actually use.
Simple Analytics never took a cent of funding and grew almost entirely by word of mouth. The dated ledger runs from $30K ARR in 2019 to a ~$500K/year third-party estimate in 2026. The number is real. The lesson underneath it is slower than it looks.
By the numbers it was my best month ever. By the way it felt, it was the month the job stopped being building and started being answering. A support-overwhelm diary at $25K MRR.
Fathom Analytics has never published its revenue, on purpose. The only figure in circulation is a third-party estimate of $5M to $10M ARR. Inside the privacy company that keeps its own numbers private.
Baremetrics made its revenue public for years, then sold to Xenon Partners for $4M cash in 2020. The real numbers behind the exit, and why the transparency pioneer walked away.
I hired my first employee at $9,000 MRR and it was the best decision I had made, until revenue slid from a $27K peak back to $22K and a fixed salary became a slow leak. A 2026 diary on the math I avoided, why I waited four months too long, and how to do a first layoff with decency.
Balsamiq's founder publishes exact revenue every year: $6.58M in 2024, down on purpose. The real number, the 2008 arc, and why data brokers get it wrong.
A solo founder pours three months into a version 2 and a Product Hunt launch at $13K MRR, and it moves revenue by about $190. The real story is the month after, and what actually grew the business.
Carrd founder AJ confirmed about $100K MRR and $1.2M ARR on The SaaS Podcast, then went quiet. The real revenue number, the $2M raise he took anyway, and why a $19-per-year, one-person business staying flat is the whole point.
The month my SaaS hit $21,000 MRR I signed my first enterprise deal: one $40,000-a-year contract worth 16% of my revenue. A first-person 2026 diary on the 280-question security review, the SSO build I pulled off the roadmap, net-60 cash delays, concentration risk, and the three rules I wrote afterward.
beehiiv is cited at around $30M in annual revenue for 2026, but founder Tyler Denk's own numbers show only ~$20M is recurring software, with ~$10M from ads and a marketplace. Here is what the figure actually counts, the growth curve, the venture backing behind it, and the operator lesson.
The month I finally quit my day job my SaaS read $11,200 MRR. A first-person 2026 diary on the number that actually let me leave (it was not the gross one), the test I used before giving notice, the runway math, and why my first full-time month grew slower, not faster.
Hypefury, the bootstrapped X (Twitter) growth tool by Samy Dindane and Yannick Veys, rode from about $300/mo to a $20K MRR peak, then publicly posted that revenue was declining. Third-party trackers now estimate roughly $567K ARR by 2024. The real lesson sits next to the number.
Third-party trackers peg Transistor.fm at roughly $4.5M ARR. The founders only ever confirmed the milestones: $1,400 in month one to "millions in ARR." The real story is the 50/50, no-VC structure two people chose on purpose.
A competitor cloned my SaaS the month it hit $30K MRR. The real 90-day revenue ledger, what the lawyer said, the three weeks I wasted, and the moat a copy cannot take.
Ghost publishes its finances live: $10.8M ARR, $907K MRR, 3.05% churn, 30,441 customers. The real story is not the number. It is the non-profit structure that means the company can never be sold.
The month my SaaS crossed $4,300 MRR I was finally ramen profitable. A first-person 2026 diary on what actually cleared my account after fees and tax, why the milestone felt smaller than the fantasy, and the checklist I use now.
Plausible crossed $1M ARR in 2022 from $64 MRR in 2019, bootstrapped and profitable. The real growth lever was not privacy. It was hiring a distribution co-founder.
One acquired customer walked out with almost a third of my $24K MRR in a single month. A first-person 2026 diary on customer concentration risk, the cashflow scramble, and the concentration cap I live by now.
A bootstrapped founder removed a freemium tier at about $16,200 MRR in 2026. The honest 90-day ledger: signups halved, support fell ~38%, and MRR grew ~21%.
The month the chart was green and she was not. A solo founder's diary of burning out at $15,000 MRR in 2026, and the five boring boundaries that pulled her back without crashing the numbers.
A bootstrapped founder's MRR chart went flat at $28K. It was not a retention problem. It was involuntary churn, the revenue that leaks out when cards fail and no one notices. Here is the ledger, and the boring fixes that recovered two thirds of it.
A bootstrapped founder's honest ledger to her first $1,000/month in 2026: 35 customers at $29, why she waited too long to charge, and why ads returned almost nothing.
A bootstrapped founder raised prices at ~$18k MRR in 2026. The real 90-day ledger: who churned, who stayed, and the segmented plan that lifted MRR 16%.
My SaaS MRR climbed 27 percent in a quarter while gross churn quietly doubled and net revenue retention slid below 100 percent. A first-person 2026 diary on catching the leak.
ConvertKit was three months from dying at $1,337/mo. The bootstrapped playbook Nathan Barry used to reach $29M ARR, every figure sourced.
Everyone copies Justin Welsh's audience. The number that actually explains his one-person business is the margin.
As of 2026, Gumroad's own revenue was about $21 million in 2023 with roughly $9 million in profit, up from $11 million in 2022 (Sacra estimates). The twist in the Gumroad revenue story: the company Sahil Lavingia publicly called a failure in 2019 became quietly, durably profitable, while the eye-catching $142 million figure is what creators earned on the platform, not what Gumroad made.
One person, no team, no VC, no exits, and about $3.1 million a year. The sourced numbers behind Pieter Levels' portfolio, and the never-sell structure that compounds them.
Arvid Kahl now teaches audience first, but FeedbackPanda hit $55K MRR in two years with no personal audience. Here is what actually drove the growth.
HeadshotPro is cited everywhere as a $300K-a-month solo business. Danny Postma stopped sharing revenue in 2023. Here is what is actually sourced, and what it teaches operators.
Bannerbear's first eight months earned about $400 a month. It took 16 months to reach $10K MRR and a patient year of technical tutorials to reach $27K. The real lesson of Bannerbear's revenue is the shape of the curve, not its size.
Testimonial.to revenue hit $800K ARR bootstrapped by one founder, out-earning his AI product PDF.ai four to one. Damon Chen's real numbers, and the lesson for 2026.
Tally reached $4.3M ARR in 2025, bootstrapped, by giving unlimited forms away free and charging just 2% of users. Why free is the growth engine, not the discount.
Marc Lou launched about twenty startups that mostly failed, then built ShipFast in a week and made $6,000 in 48 hours. A sourced look at how a Next.js boilerplate outearned everything he built for end users.
TypingMind launched five days after OpenAI opened the ChatGPT API and crossed $500,000 in revenue within a year. Here is what solo operators can copy from Tony Dinh's build-in-public playbook.
Naomi Tate built a quiet $7,400 MRR tool for solo bookkeepers, no paid ads, in 18 months. In June 2026 she sent the wind-down email. Not because she failed. Because her infra bill, her toddler, and her own client roster all added up to a different answer than the spreadsheet she started with.
A composite vertical-SaaS founder hit $9,120 MRR running entirely on four AI agents, then made her first paid hire in March 2026, not a developer, an operations contractor at $22/hr through Deel. The 60-day result: NPS +14, churn -1.2 pp, founder hours -11/wk. With the math, the agent she retired, and what she'd do the same.
Inés Vargas, a solo founder in Bilbao, hit 100 paying customers in 90 days for $1,470 in total spend. The itemized cost breakdown, the LinkedIn voice-DM channel that beat cold email 27x, the pricing flip at customer 38, and the channel she'd skip if starting again. Told as told to Joaquín del Río, with the spreadsheet on the table.
Marta del Sol runs a one-person operations studio from Valencia and crossed $4,120 in MRR across nine clients, with three AI agents doing the delivery and a $612 monthly software bill. Here's the real arc: the underpricing she's embarrassed by, the month two clients churned at once, why her clients renew for the Friday report and not the robot, and the caveats she insisted we print next to every number.
Cadence reached $12,400 in MRR with 140 accounts and an up-and-to-the-right graph, then the founders shut it down on purpose. This is the post-mortem of the most dangerous number in startups: too much to walk away from, too little to live on. The retention they didn't track, the customer they optimized for and shouldn't have, the fork they took too late, and the unusually honest way they ended it.
Tomás Iglesias turned six years of hourly freelancing into a productized service business doing $51K MRR with a team of five, and his hours dropped from 60+ to 45 a week while revenue doubled. A real Tuesday, with the numbers: the utilization metric he checks before coffee, the $28K over-hiring scare, why he still does every sales call himself, and the $13K of actual monthly profit hiding behind the headline.
The app worked. The demo hit 140,000 views and 412 people signed up on launch day. Ninety days later: 11 paying users, ~$209 MRR, and an $8,000 monthly burn. This is the boring middle where AI products actually die: the retention cliff the launch hid, the month spent building the wrong thing, the five user calls that came too late, and the specific trap of an AI demo that's too good to be true.
The two people who ask the questions and write them up.
Interviewer at OperatorBook. Sits founders down and asks the awkward question about the numbers, then prints the answer.
Narrative writer at OperatorBook. Turns spreadsheets and Slack logs into stories you actually finish.
