Bridgewater culture

Every org should try to be more like Bridgewater if they can pull it off. Radical transparency leads to better decision making, higher resource allocation efficiency, more meritocracy, and some really fun weird shit in the long run. It’s also incredibly hard to pull off, so most are better off not even trying.

There are 2 things you need to get right for this kinda org to work: a decision tracking system and the right culture.

Here’s one theoretical example of the Dalio system working: an email marketing manager sends a re-engagement campaign with A, B, C variants in the headline. They predict that A is the most promising. A support agent predicts B instead based on their customer intuition.

Support analyst wins. Their correct prediction is logged on their profile. Next time, the email marketing manager will consult the support analyst. Better campaigns ensue.

It gets more interesting when you also gate access to decision making on track record (as opposed to role and title). It’s been a year, this support agent is consistently on point in their predictions. They are now invited to decision making meetings higher up the chain, giving input into the email marketing yearly budget.

If you take this to the extreme, you get the unbundling of traditional jobs and levels. Each person’s role boundaries are determined purely by what they have a track record on. There are no more pre-defined jobs. People allocated their time and energy on problems they have advantage on, building a track record, and ladder into more and more high impact decisions.

Add some common record keeping standard and data portability, and you get massive labour market liquidity. Maybe you’ll even see the rise of micro-contracting, e.g. instead of having your pretty-good-support-agent do email marketing, you contract that decision making part to external workers with a track record on this decision in the top 1% percentile. Today, this is only possible for rote decisions where qualifying a worker’s competency is reliable and cheap aka transcriptions, translation, labeling — the MTurk tasks. The better you can assess decision competency, the more you outsource high impact and complexity problems.

With enough people with tracking standards floating around, you could run functional Decentralized Autonomous Organizations. So why aren’t we on this path? It requires some crazy skillful culture & org building. More on this next time.

27 Jan 2020