How to Operate with Keith Rabois (How to Start a Startup 2014: Lecture 14)

You want to do as less actual work and trust your junior partners to make decissions. How do should you know when to step in and when to let them go with it:

• Sort your level of conviction on a scale. Are you very sure something is a mistake or are you unsure about the outcome?
• Same with consequences: How dramatic are the effects? Is it a reversible or ireversible decision?

Where does the decission fit? 

• If low conviction & low consequences, delegate
• high conviction & consequences: step in & over-rule. The best way to do it is to explain the why rather than using authority

There are two categories of people: ammunition & barrels. Most people are ammunition. But you need barrels. You can add all the ammunition you want, but if you have only five barrels in your company, you can literally do only five things simultaneously. 

Barrels are the people who can take an idea from conception to shipping and bring the whole team together around that objective.

How to work work with barrels? Expand their responsibilities until they break. 

How to find them in the company? See who is a resource to other people.

Aiming the people in the company should be the main focus of a leader. Ideally, as Peter Thiel said, everyone should have one thing they are focusing on. Sounds simple but it is counter-intuitive as people feel insulted to be working on ONE thing, especially since they take pride in juggling different tasks.

Most people will solve problems in their area of expertise: B+ problems. A+ problems are hard, so most people procastrinate by focusing on B+ problems. It's your job not to let that happen.

The metrics help people to make decisions, hopefully as good as you, the CEO would make yourself. Transparency starts with access to metrics. Some ideas:

• Make a simple & intuitive company wide dashboard. Measure its success by how many people in the team visit it each day.
• Review board packs with employees.
• Take & share meeting notes throughout the company. Because people will want to be in every meeting.

The main goal for a metric is to predict output. Not activity.

Be aware that people will optimise what is measured. So always ask “What will people do to maximize that number?” If there is a false positive rate, make sure you track it too: If recruiters are optimising for hires, they will fill roles. You need to counterbalance it with quality of hires.

Look for anomalies. The insights is in the small deviations from the expected trend. for example: LinkedIN noticed that 35% of the users opened their static profile. It was vanity, they were looking themselves in the mirror. As they expanded on vanity features (content, endorments) stickiness grew.

One strategy, employed by companies like Paypal, is to only promote from within:

• The design lead is the best designer
• The dev lead is the best developer

This creates a performance based culture. The core assumption is that management can be thought. 

The best way to turn a high performer is through a mentor, which does not have to be the boss. The boss is focused on progress, the mentor or a coach is focused on you, the beginner manager.