"I want to build a company to last in the market. To last, I need knowledge assets." 22:17
From the episode
- His own path (19:22). He published papers after his master's, around 2009 to 2011. A conference then rejected one of his proposals for being "too practical, more of a project than research". That pushed him into building companies for a while.
- Salla's research centre (20:26). In 2022 Salla founded a research and innovation centre, which later spun out as KnowledgeX.
- Why the board hesitated (20:41). Venture investors measure returns as multiples on their investment. A research centre needs 7 to 8 years of steady spending before it pays back, at an annual return he puts at 15% to 30% (21:04). The board saw that as too early for a startup.
- Why companies do it anyway (22:17). Any product can be copied, even by your own customers while they use it. A product built on patents cannot be copied without legal risk (23:00). That, he says, is what makes Silicon Valley companies last.
- The German story (23:39). He met a researcher from a large German-language e-commerce marketplace. The researcher's job was publishing papers, for example on reordering warehouse stock by season so that in-season products sit at the front (25:19). The research department alone had about 1,800 people, he says (26:08).
- "Reserve intellectual space" (27:22). A company that wants to lead a sector stakes its claim through published papers, filed patents and licensed patents. These are knowledge assets even if they do not show up in a product straight away.
- Research is everywhere (41:43). Everyday rules come from studies: EU bottle caps that stay attached (42:55), local slaughter policy in Finland (43:48), and bidet sprays in Finnish hotels (45:01). Inside companies too, people click "OK" on the same error for years instead of asking why it happens (42:14).
The idea, explained
A company has two kinds of value. One is what it sells today. The other is what it knows that others don't. A research centre turns the second kind into assets that can be owned, defended and traded:
| Asset | What it protects | How long |
|---|---|---|
| Published paper | Priority: proof you got there first, plus reputation | Permanently public |
| Patent | The right to stop others using the invention | About 20 years from filing in most countries |
| Trade secret | Know-how you never publish | As long as it stays secret |
| Data and models | Advantage from what only you can measure | As long as you keep collecting |
The trade-off described here is about timing. A product team gets paid next quarter, while a research team pays off in years. That is why research often loses budget fights in startups, and why the companies that keep funding it tend to be the ones that last.
Context and evidence
- E-commerce companies that publish. Stories like this are common in Germany. Zalando, the country's largest fashion marketplace, has a science group that publishes on forecasting and replenishment. In 2025 it published a paper in Scientific Reports on a replenishment engine tested on about 2 million articles from about 800 merchants (Zalando Engineering).
- Bottle caps. Under EU Directive 2019/904, single-use plastic drink containers must keep their caps attached from 3 July 2024, because caps are among the items most often found on EU beaches (IFLScience).
Try it
An illustrative model, not real company data. Move the annual return to see when a research centre overtakes a quick product launch.
Key terms
Research and innovation centre · Knowledge assets · Venture capital multiples · Trade secret.