FeedbackPanda: A Couple, Zero Employees, Two Years, $55K MRR, a Seven-Figure Exit — the “Teachers-Only” SaaS

Zero employees, a couple of two, two years from founding, $55K MRR, a seven-figure exit. The case is often cited as the ideal form of indie SaaS, but the inside story founder Arvid Kahl later told holds both the sound logic of selling at the peak and a deep regret about “never hiring.” Because the numbers of success and the confession of failure are both on the record, this is the first thing anyone considering a two-person SaaS should read.

FeedbackPanda in Numbers

The Origin: “No Feedback Within 12 Hours, No Pay”

FeedbackPanda was a SaaS that let teachers teaching English online (mainly instructors on Chinese online English platforms) auto-generate and manage individualized feedback for their students. A peculiar rule of the market was the business opportunity itself: instructors must submit written feedback for each student within 12 hours of the lesson, or they don’t get paid. For an instructor teaching 20+ twenty-five-minute lessons a day, more than two hours of nightly writing awaits after ten hours of teaching.

Co-founder Danielle Simpson, a trained vocalist and a working online English teacher, suffered hours of feedback-writing every night herself, and software engineer Arvid Kahl, while holding a day job, turned it into a tool in three to four weeks. The first customer was Danielle.

The product’s core was a browser extension. It read the student ID, course ID, and classroom ID embedded in the URL of the teacher’s classroom screen and generated feedback in one click from templates. Two hours of nightly work shrank to about five minutes. The stack was Elixir + Phoenix, machine learning handled automatic pronoun conversion, and the 30-day free trial was designed with inspiration from Nir Eyal’s “Hooked.” It supported instructors across multiple platforms including VIPKid and Italki.

What Carried It to $55K with Zero Ad Spend

The launch began with Danielle naturally commenting on a “how do you handle feedback?” post in a teachers’ Facebook group. About 100 people signed up the first day, and roughly 30 new signups a day continued uninterrupted for two years. Paid advertising was zero. SEO was deliberately not pursued.

What they did instead, continually, was “celebrate the customer”: amplifying customer testimonials on Twitter and Facebook, and featuring customer stories in the weekly newsletter “VIPanda.” Testimonials like “this saves me two hours every day” propagated on their own through the Facebook groups where teachers socialize daily. The market was a tailwind too, the population of online English instructors, about 5,000 at launch, had expanded to about 75,000 across multiple schools by the time of the sale.

Our Take

A profession with a dense community of peers is the best possible market for a niche SaaS. Online English teachers (1) all share the same pain, (2) interact daily in Facebook groups, and (3) are paid hourly, so the value of saved time converts easily into money. When these three conditions align, word of mouth substitutes for a sales force. It is a textbook case of starting from “pain inside a professional community”, with parallels to Japan’s SaaS for hairdressers and MENTA.

Kahl himself explains the success factor with a phrase he sets alongside product-market fit: “product-workflow fit.” Placing the product inside the instructor’s workflow (a browser extension in the classroom screen) created a level of frictionlessness that competitors offering the same features from the outside could not match. Because Danielle, an insider, was there, that resolution on the workflow was in place from day one.

The Biggest Failure — “We Never Hired Anyone”

The value of this case isn’t only the success story. Kahl talks about “zero employees” not as a virtue but as his greatest regret.

Supporting 5,000 customers with two people, even with Intercom and automation built in for routine cases like login problems and reactivations, still left 10 to 20 individual tickets a day. When something broke, only he could respond, even in the middle of the night, and after continual emergency callouts every few weeks, he developed anxiety symptoms where his body reacted to notification sounds. He calls it “a traumatic experience,” and says the stress response hadn’t faded even a year after the sale.

The reason he didn’t hire was the assumption that “hiring can only be justified full-time (40 hours a week)”. He rejected it even though Danielle had proposed part-time hiring. His conclusion in hindsight: “We should have hired part-time customer support and a DevOps engineer.” Zero employees helped margins and sale-readiness, but the founders’ minds and bodies paid the price.

The business also carried a structural risk: most customers were instructors of one specific online English company, whose policy changes were a business risk. The decision to sell at the peak ($55K MRR and churn under 3%, before the risk materialized) is the “pre-crisis exit,” the mirror image of Black Magic’s post-crisis sale.

After the Sale — Articulating the Experience Became the Next Business

In 2019, the company sold to SureSwift Capital for seven figures. Kahl’s next business became “articulating the experience”: two books, “Zero to Sold” and “The Embedded Entrepreneur,” the podcast and newsletter “The Bootstrapped Founder,” and a Twitter presence (later exceeding 190,000 followers). He is now building Podscan.fm. The exit was not an ending. He modeled a career pattern in which the experience itself becomes the next content asset.

Conditions for Replication — and the Limits

Japanese readers can take the playbook itself: use repetitive pain enforced by an institution’s rules, like “submit within 12 hours”, as the starting point of market discovery. Put an insider on the team. Embed the product inside the work’s flow. And put celebrating your customers at the center of marketing. All of these are questions of design, not capital, and a duo can execute them.

The market’s growth belongs to this case alone. A tailwind that multiplies the instructor population roughly 15x in two years (5,000 → 75,000) cannot be engineered, and with a static market the same setup would not reach $55K MRR. Also, the single-platform dependency was not solved by the sale, merely transferred to the buyer, and this exit only opens while MRR and churn are at their best. And anyone tempted to imitate “zero employees” should reference it together with the fact that Kahl regrets it.

Further Reading

• MENTA — the Japanese version of starting from a professional community
• TypingMind/Black Magic — the contrasting case of selling “post-crisis”
• Tweet Hunter — a sale at the same two-year speed that suffered through the earnout