Canny: $3.5M to $3.3M in Year 7, the First Decline the Feedback SaaS Published Anyway

“Year 6 ARR: $3.5M. Year 7: $3.3M.” Canny, the feedback-management SaaS that has written a numbers-filled annual review every year since founding, published its first-ever revenue decline in year 7, as is. Outside funding: $0, consistently. Team: 16 at the time of the report, 17 shortly after. The company explains the drop with two factors, backed by actual paying-customer counts: a direct hit from the economic downturn, and the recoil from a free plan it launched itself.

Canny is a tool that lets SaaS companies collect and organize user feature requests and feed them into development prioritization. It launched in March 2017, built by two people, product designer Sarah Hum and engineer Andrew, and has grown fully remote ever since. Annual reviews written in the same format in both up years and down years are one of the few primary sources where you can observe a bootstrapped SaaS’s growth and stagnation at fixed points in time.

Seven years of numbers

• March 2017: Launch, 2 people; $10K ARR and 50 customers within 2 months
• ~August 2018: $200K ARR, third hire (customer success)
• September 2020: $1M ARR, 7 people — 3.5 years after launch
• Year 5 (2021): $2.2M ARR, 9 people
• Year 6 (2022): $3.5M ARR, 13 people, 607 paying customers
• Year 7 (2023): $3.3M ARR, 16 people, 583 paying customers — first decline
• March 2024: over $3.4M ARR, 17 people; 6,000+ teams on the free plan

The route to $1M ARR is also on the official blog. $10K ARR and 50 customers two months after launch, then $50K within six months. Average price started at $20/month and tripled to $60/month by the $50K ARR mark. Growth was “all inbound”: Product Hunt, Hacker News, search, Quora, and word of mouth via the “Powered by Canny” link on the widget. No salespeople, the product itself doubled as the distribution machine.

The hiring order is distinctive too. The founders’ first hire was not an engineer but a customer success manager, at around $200K ARR (roughly August 2018). An engineer and a marketer followed in 2019, and the team was still just 7 at $1M ARR. Headcount consistently lags ARR, including the fully remote setup spread across 3 continents and 4 countries, fixed costs are made to trail revenue, a pattern consistent across all seven years.

Decomposing the decline

For the $200K drop in Year 7, the report lists multiple causes. One is macro: interest rates hit multi-decade highs in several countries, and funded companies rushed to cut costs. Canny’s customers are precisely those SaaS companies, so it took the cancellation wave head-on. The line “the world was/is still struggling and we felt that” sums up B2B SaaS’s sensitivity to the business cycle.

The other was its own decision. When the free tier launched, “around half of our paying customers overnight” moved down to free, while ARPU more than doubled. Letting small accounts parked on cheap paid plans flow to free and concentrating billing on customers who truly pay is textbook freemium restructuring, but it wounds the top line immediately. The way pricing and customer-mix changes surface in churn numbers echoes ScreenshotOne’s case of hitting a churn ceiling.

Cross-referencing the numbers: paying customers fell from 607 to 583 (24 accounts, just under 4%) while ARR fell about 6%. That the ARR decline outpaced the customer decline suggests the remaining customers were also shrinking or downgrading. You can estimate whether accounts or price-per-account eroded only because both actual figures are published.

The countermeasures are concrete: shifting to annual and multi-year contracts, a 3-week product update cadence, AI features (smart replies, comment summaries, a beta that auto-extracts requests from customer conversations), and added hires in operations and customer success. In its down year, Canny grew headcount from 16 to 17 rather than cutting.

Disclosure as a sales machine

What makes Canny’s annual reviews interesting is that disclosure itself is designed as an acquisition channel. The company’s early growth was “all inbound,” and detailed numbers-filled blog posts are at its core. The post charting the road to $1M ARR and the year-7 decline report both keep drawing mentions, year after year, to a tool in the unglamorous category of feedback management. Like Plausible’s “SaaS sold on principles” reaching $1M ARR with zero ad spend, transparency functions as a substitute for a marketing budget.

At the same time, published numbers are chosen numbers. ARR, headcount, and customer counts appear; profit, individual salaries, and actual churn rates do not. The “Raised $0” banner also lowers external growth expectations relative to VC-backed SaaS. Disclosure is proof of honesty and, simultaneously, an expectations-management device, the safe way to read it.

What transfers, and what doesn’t

Three takeaways. The cost of freemium: “half the paying customers went free overnight; ARPU more than doubled” is a rare benchmark for any SaaS weighing a free plan. The macro linkage: B2B SaaS revenue tracks its customers’ economy, obvious but rarely quantified, and here shown in a concrete band, $3.5M to $3.3M. The channel: a documented case of detailed disclosure functioning as an acquisition route for seven straight years.

The limits are equally clear. Canny’s takeoff relied on product quality it itself billed as “built by FAANG-caliber engineers and designers” and on the amplification power of Product Hunt and Hacker News circa 2017. There is no guarantee the same inbound-only playbook works in today’s crowded category. Indeed, in the same category there are cases like Usersnap, sold for seven figures after ten years of operation, where founders chose exit at a growth plateau. Bootstrapped for seven years, adding headcount even in a down year. That is the consequence of choosing not to raise, and building the financial position that allows that choice is the precondition for this playbook.