Loading page…
A Berlin company sells a €200 door controller at 50% margin, then charges €10 per door per month for the platform on top — and has never lost a customer, because ripping it out means visiting every door in the building.
≈€500K of that recurring (≈$574K ARR by the host's conversion) · 5,000 doors live across ~20 logos · ~€10/door/month list price before heavy discounts · ~50% hardware margin · zero churn since the 2016 launch · >€5M raised, not yet profitable
"I call our door controllers the Trojan horses to the buildings — that's by the way why I give so much discount on the recurring, because I know once I have them we have zero churn."
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Michael von Roeder is not Sensorberg's founder. He was an investor in the Berlin company and stepped in as CEO — by his own description, "in a difficult place" — inheriting about 30 people. The relationship with the founder survived: the founder sits on his board and they were having lunch the following week. The product had launched in 2016. It starts with hardware door controllers that turn a building's access points into something programmable, and then layers a cloud platform on top that brings in sensors and actuators — heating valves, temperature sensors, lighting control. The promise is that you unlock the building, book a meeting room, and take delivery of an Amazon package or your groceries through one system, on a deliberately open platform rather than the proprietary stack von Roeder expects from GE or Siemens. The wedge is co-working spaces. They already run a member app for community, and Sensorberg makes that app control the physical space as well. Corporates then adopt co-working concepts internally, and that is the road into corporate buildings. The numbers he gave Nathan Latka in January 2019: roughly €2.5M of total revenue that year, of which about €500K is recurring — Latka converts that to about $574K of ARR. Five thousand doors live across roughly twenty logos, at a list price near €10 per door per month with heavy discounting. Hardware carries about a 50% margin. The company loses money, has raised more than €5M, and was raising a low-single-digit-million round to reach break-even in late 2019 or early 2020 at around a €20M pre-money target. The number that makes the whole thing worth studying: zero churn, across the entire history of the company.
Sensorberg does not start with the landlord. It starts with co-working operators, because they already run a member app and already treat their space as a product.
The quick-reference card — acquisition channels, replication playbook, risks & traps, and our verdict — is members-only.
Join the Unicorn Club — $5/moCancel anytime, in one clickPaste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself. The exact words are quoted below, with a timestamp that jumps to the second each figure is said.
“…50 grand per month right now in pure place sas revenue 2.5 million in terms of last 12 months revenue so significant portion of hardware they're looking at obviously scaling the…”
13:32 · hear it in the videoData credibility: Founder-reported, in a January 2019 interview on Nathan Latka's show, where the host does most of the arithmetic out loud and von Roeder confirms or corrects it. Nothing here is third-party verified: there is no dashboard, no filing and no analyst data. Two cautions about the numbers. First, the auto-caption garbles the top line — the transcript reads "25 million euros" but both men immediately work from €2.5M, and von Roeder corrects Latka's €250K recurring estimate up to "about 500, 500ish," so ≈€2.5M total with ≈€500K recurring is the reading we use. Second, the widely quoted "$50K per month in SaaS" is Latka's list-price arithmetic (5,000 doors × €10), and von Roeder explicitly answers that it only holds theoretically because the discounts are heavy. The zero-churn claim, the 50% hardware margin, the €200-plus unit cost, the €5M-plus raised and the €20M pre-money target are all his own statements. The figures are from early 2019 and have not been re-checked since.
The first money is hardware money. A customer buys door controllers — devices that cost Sensorberg more than €200 each to make and carry roughly a 50% margin — and the €10 per door per month starts once the hardware is on the wall.
Expansion, not acquisition, is what compounds here. A co-working chain opens a new location every month and Sensorberg walks in with no additional sales effort — Unicorn went from one space to six in two months, with twenty planned for the following year.
A year-long sales cycle means that for most of any quarter nothing measurable happens, and Sensorberg carries a hardware bill while it waits. The company loses money; von Roeder was raising a low-single-digit-million round specifically to reach break-even in late 2019 or early 2020, on top of the €5M-plus already in.
Two layers. The bottom layer is a door controller — standard, open hardware, a quad-core processor built on the Raspberry Pi Compute platform, which von Roeder stresses because it means nobody is locked into a proprietary board. It costs the company more than €200 to produce at current batch sizes of two to five hundred units. The top layer is an IoT platform. Once the controller is on the door, the rest of the building connects wirelessly: heating valves, temperature sensors, lighting. The customer can source those sensors themselves — he explicitly mentions customers buying sensors in China and bringing them into the building — and Sensorberg detects them and starts billing per connected unit. What sits on top of that is other people's software. Von Roeder's bet is that the future of building automation is not one vendor doing everything the way GE or Siemens would, but many contributors. Sensorberg works with other startups who provide the apps and integrates with them. The end-user experience is expressed in use cases rather than features: enter the building, book a meeting room, receive an Amazon package, receive your groceries. Our read: the product's real shape is a bus, not an app. The door controller is the piece that has to be Sensorberg's, because it is the one that is physically screwed into the wall and therefore impossible to swap on a whim. Everything above it can be someone else's code, and that is by design — it is what lets a thirty-person company claim a platform story without building a platform's worth of features.
Three lines, in the order the customer meets them. First, the hardware sale, at about a 50% margin. Second, the subscription: around €10 per door per month, falling sharply with volume — a customer with two hundred doors pays less per door, and heavy discounts are also handed out simply to grab market. Third, per-unit fees for every additional sensor or actuator connected to the platform, which the customer can add themselves. Five thousand doors at list price is the €50K per month that Latka does out loud, and von Roeder's answer is careful: theoretically yes, but the discounts are heavy. Pushed for a real number, he gives the top line instead — roughly €2.5M of revenue for the year, of which about 20% is recurring, so about €500K. Latka converts that to roughly $574K of ARR. Across roughly twenty logos, that is an enterprise motion: a year-long sales cycle per new name, several hundred doors per account. There is a fourth, stranger line. Because European real estate does not buy subscriptions, Sensorberg lets customers pre-pay seven years of the recurring fee, and some have. Our read: the pre-pay option is the most copyable idea in the interview and the most dangerous one on the balance sheet. It converts a conservative buyer's objection into cash today, and it also means a chunk of "recurring revenue" has already been spent by the time the service is delivered. If you copy it, track deferred revenue separately from anything you call ARR.
The controller costs more than €200 to build. At 50% margin, that sets the hardware price, and the hardware price is what makes the subscription feel like an afterthought to the buyer. Von Roeder's stated ideal is to give the hardware away entirely and be a pure SaaS business, and he is precise about the condition: batch sizes need to go from two-to-five-hundred units to five thousand, at which point he moves production to China through sourcing partners he already has from a previous business, and the buying cost needs to fall below €100 per unit. Under €100, the mixed model becomes possible. The team is about thirty people, most physically in Berlin, some in Poland and elsewhere in Germany. There is no prescribed work time and no prescribed location — his stated principle. Berlin is the choice because, in his view, it is where you find the best engineers in Germany and probably in Europe. The company is not profitable. More than €5M raised, a low-single-digit-million round in progress, break-even targeted for late 2019 or early 2020 — and he is explicit that break-even is what makes the company free to do what it wants. Our read: this is the structural tension of the business and Latka names it at the end of the interview. Investors will price the company off the hardware line, where multiples are low, while von Roeder wants to be priced off the €500K of recurring revenue. He says the fight has already started. A solo builder copying the model should notice that every euro of hardware revenue drags the company's valuation down even while it buys the zero churn.
Copy the Trojan horse. The reason Sensorberg has never lost a customer is not the software; it is that removing the system means sending an electrician to every door in the building. Any software that becomes physically annoying to remove inherits the same churn profile, and the version of that available to a solo builder is not hardware at all — it is being the system of record that other tools write into, or being the thing printed on a sticker at the point of use. Copy the wedge. Selling to co-working operators works because they already have a member app, an appetite for member experience, and a location roadmap that expands your account for you. The same shape exists in student housing, serviced apartments, gyms, and clinic chains. Copy the pre-pay answer. When a conservative buyer says they do not want a subscription, "then pre-pay seven years" is a better response than a discount. Change the hardware. Von Roeder needs €100 controllers and five-thousand-unit batches before his preferred model is even legal on the P&L; that is a €5M-plus problem. In 2026 the same access layer can be rented: commercial smart locks and controllers ship with APIs, and a software company can sit on top of installed hardware instead of manufacturing it. Our read: the copyable business here is the platform layer without the factory — a booking, access and sensor dashboard for flex-space operators that integrates with whatever locks are already on the doors, priced per door per month. You lose the 50% hardware margin and the physical lock-in. You also lose the €5M raise, the year of manufacturing lead time, and the valuation drag Latka warns about at the end of the interview.
Zero churn is a young company's statistic. The product launched in 2016 and this interview is from January 2019; nobody has reached a renewal decision after a full building refit cycle, and seven-year pre-payments guarantee that several accounts cannot churn yet even if they want to. Treat the number as evidence that removal is painful, not as proof that customers are happy. Concentration. Roughly twenty logos and five thousand doors means a handful of co-working chains carry the revenue. Co-working operators are cyclical tenants of somebody else's building; when they stop opening locations, the expansion flywheel — the part that costs nothing — stops with them. The hardware cost curve. The mixed model requires sub-€100 controllers at five-thousand-unit batches. If that batch size does not arrive, the company keeps selling €200-plus devices at 50% margin and stays a hardware business wearing a SaaS label. Funding dependency. The company loses money, has raised over €5M, and needs another round to reach break-even. Break-even is the stated goal precisely because it removes this risk. Incumbents. Von Roeder names GE and Siemens as the proprietary alternative. They do not need to be better; they need to be the safe choice for a conservative European facilities buyer with a year-long procurement cycle. Our read: the quiet risk is the valuation fight. If investors price Sensorberg as a hardware manufacturer, the round gets done at a number that does not reflect the recurring line, and the cheapest way out becomes an early sale — which von Roeder says he is already being asked about and does not want.
The four founder playbooks — first customers, first dollar, the flywheel, the silent stretch — are in the Unicorn Club, on this idea and on all 1,435 others. The story, the numbers, the sources and every later section stay open to everyone.
Join the Unicorn Club — $5/moCancel anytime, in one click