From €500 Market Test to World Market Leader: 10 Years of Multi-Cloud Hosting for KRUU

A few weeks ago, Markus Seidel invited KRUU co-founder Philipp Schreiber as a guest on his podcast mittelstars (in German). The conversation covers punk rock, a German Shark Tank episode, a no-nonsense US border officer and a market that has no clear global counterpart. Between the anecdotes, however, you can hear a lot about the mindset a self-funded hardware D2C business needs.

We've actually been in this story longer than KRUU itself has existed: Philipp's earlier agency “Schreiber und Freunde” was already one of our hosting customers — so when the photo booth idea took shape in 2015, we were on board from day one. From that first single server to today's multi-cloud architecture, we've had a hand in every IT and cloud decision along the way. It's a position we're genuinely proud of.

From €500 in Google Ads to the first order

November 2015. Philipp Schreiber and his two co-founders have no finished photo booth, no production hardware. What they do have: a prototype, a WordPress site, a form plugin, a stock photo, a price two euros next to the existing competitor — and a €500 Google Ads budget. Ten minutes after switching the ads on, the first lead drops in. An hour later they switch the ads off again because they can't keep up.

Only after that — in February 2016 — does the first actual photo booth get built and shipped. Pre-payment only, because the existing competitor was doing the same.

It's not just a nice founding anecdote. It's the operating system of the company: test, learn, then put money into hardware. We've translated that same mindset into the IT architecture for ten years now. Instead of renting expensive managed services for scenarios that don't exist yet, the stack grows with the business — step by step from the single server in 2016 to today's multi-cloud setup across Hetzner, Gridscale, Digital Ocean and Cloudflare. If you start with the big hyperscaler bet, you're stuck with a cloud bill from day one that usually scales well ahead of the business.

We consult on exactly this question regularly — and consciously chose against the hype with KRUU.

Seedstrapped: why bootstrap-ready IT architecture is not a luxury

In 2018 KRUU did a seed round with business angels and one family office. Since then, no further funding. Philipp coined the term himself on the podcast: “seedstrapped” — one round, then back to bootstrapping. The liquidity to build 5,000 photo booths comes from end-customer pre-payments.

This isn't just a financing decision. It has direct technical consequences. If you don't have a 100-million Series B behind you, you can't afford a hyperscaler bet that burns five percent or more of revenue on cloud bills. Multi-cloud capability is no longer a theoretical architecture exercise — it becomes cashflow protection. The Managed Cloud Servers behind the KRUU platform today aren't an end in themselves — they are the consequence of a very concrete entrepreneurial choice: put money into equity, product and hardware, not into platform fees. The result after ten years: KRUU's IT spend is below 0.5 % of revenue — a number you rarely see, and one of the drivers behind the business model itself.

If you're standing in a similar spot today, our startup offering is the same idea on a smaller scale: a setup that isn't more expensive than necessary in the early phase and doesn't get in its own way later.

100,000 orders, 11 countries, two warehouses

KRUU today has over 5,000 photo booths in rotation. In peak season, around 2,000 of them are shipping every week via UPS — Europe is served from KRUU's own logistics centre in Heilbronn, the US side from KRUU's equally own warehouse in Detroit, Michigan. We run the fulfilment servers on-site at both locations, steering shipping and returns directly from the warehouse floor. On the podcast Philipp is explicitly asked why KRUU doesn't add further warehouses in southern Europe or Texas. The answer, paraphrased: for that warehouse rent and admin headcount, you can build a lot of additional photo booths instead.

That same logic — the simplest viable setup, then more units — maps one-to-one onto the IT architecture. We serve Europe and the US with a stack that runs centrally and only gets decentralised where it really has to. Concretely:

• Booking backend centrally hosted in Germany, internationalised for 11 markets

• Inventory tracking across all 5,000+ devices, automatically interlocked with the warehouse processes

• KRUU app since 2023 with over one million end-user registrations — photo upload, gallery sharing, push notifications

• Photo delivery via Cloudflare as the edge cache plus a self-operated Imgproxy for on-the-fly image processing, so guests in Lyon, Detroit or Berlin load the galleries equally quickly

• US stack in Michigan running parallel to the European setup, on the same code base — so no branch drifts apart over time

What we actually host

The full tech-stack overview is laid out structurally in the KRUU case study. Just the components that make a daily-business difference here:

• Booking and inventory backend in Symfony on our managed servers — the same stack we now offer as a managed service to other customers

• App backend for over a million end users with photo upload, gallery sharing and push notifications

• Fulfilment servers on-site at the KRUU logistics centre in Heilbronn and the KRUU warehouse in Detroit

• 24/7 operations across multi-cloud (Hetzner, Gridscale, Digital Ocean), Cloudflare at the edge

• Patching, monitoring and incident response — resolved quickly and proactively, together with KRUU or independently, depending on the topic

This complexity has to remain invisible. Booths go out and come back during the week, but the events happen on the weekends: bookings, app logins, photo uploads. None of that knows a weekly pause.

Deeply embedded in development

In the day-to-day reality, we're not the hosting provider who calls once a quarter. We're a sparring partner for KRUU's engineering team — from architecture decisions to tool choices (CRM, ERP, monitoring) to whether a new workflow holds together technically at all. Philipp puts it like this in our joint case study: “What impresses me is the depth of expertise. CRM, ERP, a tool tip — whatever it is, I just ask Andy. He’s like a Swiss army knife for tools and cloud systems.”

This position only works because both sides share an engineering attitude: the philosophy of boring technology. Proven building blocks, stable platforms, clear routines. Spectacular engineering isn't a goal — it's usually a sign that not enough has been filtered out yet. Across ten years, that shared caution is one of the reasons the KRUU platform stays above 99.99 % server uptime.

And it's the reason the collaboration runs asynchronously: because both sides trust the other to do the right thing at the right moment. Trust isn't a buzzword here, it's the prerequisite for on-call nights working at all.

What ten years of hands-on teaches you

For all the tech, the most important takeaway from ten years with KRUU is something else entirely: a bootstrapping mindset on the customer side forces honesty on the provider side. If you watch every euro monthly, you don't buy plausible-sounding architectures whose real purpose is optimising our margin. Philipp's own mantra on the podcast — “don’t work with assholes, don’t work for assholes” — describes the same thing from the other direction.

As Philipp puts it on our own website: “What I like about Andy and his team is their focus on the best solution – for us as customers. Even if it sometimes means less revenue for ‘We Manage.’ I don’t know this kind of approach from other service providers.”

Building a similar scaling story?

If you find yourself in a similar spot to KRUU in 2018 — bootstrapped, growing fast, IT bills starting to bite — our startup offering is built for exactly that phase.

For the full KRUU tech story with architecture, FAQ and KPIs: the KRUU case study.
Prefer to write directly? This way.