Rocket Internet: The Clone Factory That Predicted the Holdco Era
In 1999, three brothers from Cologne built a German copy of eBay called Alando and sold it to eBay itself for a reported 43 million dollars — less than a hundred days after launch. That trade taught Oliver, Marc and Alexander Samwer the lesson that would define European tech for the next two decades: you do not need the original idea. You need the market, the speed, and the operators.
In 2007 they turned that lesson into a factory. Rocket Internet did not build one company. It built a system for building companies.
- Rocket industrialised company-building: pick a proven model, localise fast, fund aggressively, exit or list.
- It listed in Frankfurt in October 2014 at 42.50 euros a share, valuing it around 6.7 billion euros.
- It delisted in October 2020 at roughly 18.57 euros a share — the public experiment did not work.
- The failure was not the operating model. It was the mismatch between a portfolio of long bets and a quarterly reporting cycle.
- Every modern holdco — including ours — is running a version of Rocket’s thesis with the public-market wrapper removed.
What Rocket actually got right
The clone label was always a little lazy. Yes, Rocket copied Amazon, Uber, Airbnb and Groupon into markets those companies had not reached. But the genuine innovation was not the products. It was the machinery.
Rocket treated launching a company as a repeatable industrial process. Shared recruiting. Shared capital. Shared playbooks for warehousing, payments, performance marketing and hiring country managers. A new venture in a new market did not start from zero — it started from the accumulated operating knowledge of every venture that came before it.
Rocket’s real product was never the startups. It was the compounding operational knowledge that made the next one cheaper to launch than the last.
That is the exact thesis behind every serious holding company operating today, ours included. We do not think of Canvas & Ivy as a brand we happen to own. We think of it as the place where the playbook gets proven before it touches anything else.
Why the public markets rejected it
Rocket went public in October 2014 at the top of its range. Six years later it left, telling shareholders it was "better positioned as a company not listed on a stock exchange" and that public capital was no longer an essential financing source.
| Moment | Datapoint | Reading |
|---|---|---|
| Alando exit, 1999 | About 43 million dollars in under 100 days | The founding insight: speed beats originality |
| Frankfurt IPO, Oct 2014 | 42.50 euros per share, roughly 6.7 billion euros | Peak narrative — a factory for internet companies |
| Delisting, Oct 2020 | Around 18.57 euros per share | The narrative did not survive quarterly scrutiny |
| Today | Private holding and family office | The model persists; the wrapper is gone |
The structural problem is easy to state. A portfolio of early ventures produces lumpy, unpredictable, occasionally embarrassing results. Public markets pay for legible, repeatable, quarterly results. Rocket asked shareholders to hold a venture portfolio while judging it like an operating company, and shareholders declined.
The lesson we actually use
Rocket is not a cautionary tale about copying. It is a cautionary tale about capital structure. The operating thesis — build a machine, reuse it across assets, let each one make the next cheaper — was correct, and is more correct now than it was in 2014, because AI has collapsed the cost of the machine itself.
What Rocket could not do was run that thesis in public. So the modern version runs private, with permanent capital and no obligation to explain a slow quarter to anyone.
- Permanent hold horizon. No fund clock, no forced exit window.
- One proving ground brand before any playbook touches the portfolio.
- Shared infrastructure across brands is the asset — not any single brand.
- Stay private long enough that lumpy quarters are a feature, not a crisis.
Rocket Internet was right about almost everything except where to keep the shares. That is a very specific kind of failure, and a very useful one to study.
Sources: CNBC (September 2020); TechCrunch, "As it delists, Rocket Internet’s ill-fated experiment with public markets is over" (September 2020); public filings and contemporaneous reporting on the 2014 Frankfurt listing.
Own a brand this applies to?
We talk to founders at every stage — long before they're ready to sell.
Start a Conversation →