A weekend tool for one teacher, sold two years later at US$55,000 a month

In 2017 Danielle, an opera singer, was teaching English online to Chinese children as a side job. 3 Arvid was a software engineer at a German company in the IoT space. 2

The teachers were paid only for the time in front of the camera. 3 After a day of teaching, Danielle spent two more hours writing feedback for the parents, the same sentences again and again, in Excel sheets and Word documents. 2

Then Danielle saw that the teaching platform showed the student and course ID numbers. That made it possible to keep each student's records in order. Danielle asked Arvid to check it, and Arvid said: "Sure, I can build that." 3

Arvid built the first version in about a week, at night and on weekends, next to the full-time job. It already cut Danielle's feedback time from hours to minutes. 3 Arvid and Danielle called it FeedbackPanda. 2

For about three months Danielle used it every day while Arvid improved it. 3 Arvid chose the tools already in use at work, Elixir and Vue.js, because a business is not the place to learn new technology. Payments, customer chat and logins went to Stripe, Intercom and Auth0. 3

Arvid built only what was needed to take money. Account management and plan changes were left out: "There's no point in working on features for customers you don't even have yet." 3

The teachers met in Facebook groups and Instagram communities. Danielle was one of those teachers. The groups were where Arvid and Danielle checked the problems, and where all the marketing happened. 4 By Arvid's account, the marketing came down to "a number of comments on some Facebook posts". The teachers did the rest by sharing it. 4

The product itself spread the same way. Teachers wrote feedback templates and could share the templates in what FeedbackPanda called the cloud. Other teachers found, changed and shared the templates back. 4

FeedbackPanda also rode a wave. The largest Chinese school had 10,000 teachers when Danielle started in 2017, 20,000 by December 2018, and 75,000 when FeedbackPanda was sold in July 2019. 4 FeedbackPanda sold to each teacher directly, never to the schools. 4

Arvid had one goal from the start: a business that makes over US$50,000 in a month. Arvid split the road into three stages, each with its own target. 3

The business paid for itself 45 days after the marketing started, at about US$2,500 a month. Nine months after the launch it reached US$20,000 a month, enough for Arvid and Danielle to work on it full time. Shortly before SureSwift called, it passed US$50,000. 3

The 30-day free trial converted over 25% of teachers, and churn stayed below 1%, by Danielle's account. Most teachers who left had changed jobs or taken parental leave. 3

The price started at US$5 a month. The teachers converted so fast that it doubled to US$10, and the US$5 plan went away. After a year at US$10, the price rose by half, to US$15 a month. 4

With that rise came a referral program, so referred teachers could keep the US$10 price. Teachers on the old plan kept it for good. Arvid says many of these choices came "not from a place of optimism but of severe pessimism". 4

Arvid ran the engineering and operations. When a Chinese school changed its website, a FeedbackPanda integration broke, about once every two weeks, sometimes on a Sunday night. 1

Support was answered once and then written down. Every first answer became a help article, and after a few months over a hundred articles answered most questions with no human reply. 1

The low point was the servers. A small, cheap German cloud company kept failing, first for minutes a day, then for a whole day. The move to Google Cloud took a few hours with a friend's help, and then there were zero seconds of downtime for the next 400 days. 4

Hiring never happened. Arvid thought a hire had to be a full-time job, so Arvid took the support alone. 4 After the sale, the replacements had to be trained, and Arvid learned it was "easy and enjoyable". 4

Arvid and Danielle never planned to sell. But Arvid had read Built to Sell before starting, so the company was built to run without its founders. 4 When Danielle gave an interview to Indie Hackers at US$35,000 a month, SureSwift Capital noticed and reached out. 3

Why sell: everything Arvid and Danielle owned sat inside one business. 4 Danielle saw the founders making choices out of comfort and safety, not good business, and wanted an owner with more appetite for risk. 3 In July 2019 FeedbackPanda joined the SureSwift portfolio. 3

Arvid will not say the price, only that it was "a life-changing amount of money". 4

Once replaced, Arvid felt without a purpose, and started to write down the lessons of FeedbackPanda. 5 Arvid started a blog, The Bootstrapped Founder, committed to one article a week, and announced it to maybe 500 followers on Twitter. 5

The posts fell into the shape of a full guide. A reader offered US$10 for a PDF of it, many others asked for a book, and Arvid saw it as a validated product. 5 A copy editor trimmed it from 598 pages to 500, and a proofreader checked it again. The editing cost about US$3,000. 5

Arvid hoped to sell 50 copies in the first week. It sold almost exactly 1,000, for about US$8,500. In the first month, 1,571 books brought in US$12,871, and after US$4,000 of costs, over US$8,000 of pre-tax profit. 5

Arvid then wrote The Embedded Entrepreneur and made Find your Following, a Twitter course. 2 Today Arvid runs Podscan, a podcast intelligence platform. 40

1. What did SureSwift pay, and how was the price worked out?
2. How did revenue split between monthly and yearly plans, and between the US$10 and US$15 prices?
3. How much did FeedbackPanda cost to run each month at US$55,000 in revenue?
4. What happened to FeedbackPanda after the handover?