Thread by @Ben_Reinhardt

1/ Individual Incentives + Corporate R&D: A Thread

The built-in incentives mean that corporate R&D orgs can only get and keep the best people in a narrow range of circumstances.

2/ People are generally motivated by one or more of money, glory, autonomy, or impact.

Let's see how each of those interact with corporate R&D

3/Money

If an individual researcher is motivated by money and create something valuable, they can capture more of the value they create if they spin off a startup.*

Therefore, researchers who stay in R&D labs long-term aren’t motivated by capturing value.

4/ * The ability to capture more value by starting a company is especially true in the world of software but may hold less true in areas with higher capital costs.

It also depends on the industry’s willingness to acquire.

5/ Glory

Researchers can usually capture more glory in academia. Companies have the incentive to keep many more things secret. In academia, the only thing that keeps people from publishing is fear of being scooped.

6/ In a company you want to keep the research secret until you are already in a position where you can execute on the research better or you have no interest in executing on the research.

7/ Therefore, you would expect researchers who are doing things relevant to the company’s core business to have restrictions on what they can do and therefore less publishing and glory-capture than if they were in academia.

8/ One exception to the ‘glory-restrictions’ on business-adjacent research may be if the company feels like they’re the only ones who can take advantage of the research.

9/ One of the reasons Bell Labs may have been able to publish so much was because their monopoly position assured them they would be the only ones able to execute on the research.

Google Brain may fall into this same category - an effective monopoly on search and video hosting

10/ Autonomy

Instead of money or glory people might want to work on interesting or novel work that would otherwise require months of grant writing.

11/ Novel, interesting work is often not aligned with the parent org because the value of innovations is not always clear and big companies cannot start small things - you need to create something that can add millions of dollars of value within a year to even move the needle.

12/ If research doesn’t fit into an existing business unit, only outliers are incentivized to keep it going which means that researchers in corporate R&D can only be autonomous for a finite amount of time.

13/ Impact

Corporate R&D can potentially have large impact *if* you’re working on something that’s aligned with the main company.

For example if you’re working on research that can become part of an existing Google product, it will affect billions of lives.

14/ However, because R&D efforts need to be connected to a core product to maintain support, any research not connected to a core product will have a finite lifetime and probably won’t have a large impact.

15/ One way that Corporate R&D *can* get the best people is if they provide resources that would be hard to come by otherwise - the massive compute at OpenAI and Google Brain for example, or access to world-class labs at pharma companies.

16/ Incentivizing the best people to stay may also be part of the reason why R&D orgs are expensive. You need to provide hard-to-get resources, which usually translates to $$.

You could make them feel extremely comfortable - call this a ‘local optima’ - which is also expensive.

end/

This is part of a bigger piece I'm working on about constraints on corporate R&D as an innovation organization to finally continue the series I started with benjaminreinhardt.com/startup-constr…