How to Operate

Keith Rabois, Partner at Khosla Ventures and former executive at PayPal, LinkedIn, and others talks about how to forge a company, which in many ways is more difficult than forging a product.

The talk is full of great advice and probably the most densly packed of this series so far, which will explain the quantity of notes below. One drawback is not being able to cover the nuances of the advice. The answer to the final question at 44m52s demonstrates how the seemingly contradictaroy advice of Insist on Focus and Details Matter is resolved. I found the analogy of barrels and ammunition at 14m11s to be very vivid and an important concept.

Keith Rabois – @rabois


  • 0m18s – forging a company more difficult than forging a product becaues people are irrational
  • start with cleanly aritechtected engine but reality is messy and full of hacks in the beginning
  • ultimately an effecient well organized and self-running organization is the goal
  • build a company that idiots could run because eventually they will
  • 1m38s – manager’s output = Maximize output of organization and organizations around you
  • recommended book: High Output Management by Andy Grove
  • motion is not same as progress
  • 2m43s – # Triaging: in beginning every day there is a new problem
  • something is low priority; others look low priority but could be potentially fatal
  • 3m19s – # Editing: an editor is best metaphor for job of CEO (as opposed to writing)
  • first thing an editor does is take out red pen and eliminate things; clarify and simplify for empoloyees
  • don’t accept excuse of complexity; you can change the world in 140 characters
  • “think different”, “1000 songs in your pocket”, etc.
  • simplify every marketing initiative, product, everything you do
  • 5m14s – # Clarify: find ambiguity and ask questions to clear up understanding
  • when you can eliminate unnecesary steps, you can improve performance 30 – 50%
  • 6m20s – # Allocate resources: can be top down or bottom up
  • people who work with you should come up with their own initiatives
  • goal overtime is to use less ink over time
  • 7m41s – # Ensure Consistent Voice: any product communication should feel like it was written by the same person
  • initially it is ok for founder to be that voice, but overtime you don’t want to be the one doing all the voice editing yourself
  • 9m04s – # Delegate: you shouldn’t be doing most of the work, but you are still responsible for everything
  • Abdicate v. micromanagement – both sins
  • 9m55s – Task-relevant Maturity – low v. high
  • your management style needs to be dictated by your employee
  • 11m15s – 2×2 decision grid: level of conviction (high v low); Consequence of Decision (low v. high)
  • low, low = Delegate fully; high, high = step in. decide. overrule.
  • high, low or low, high = ?
  • important to try to explain the why of your decisions
  • 14m11s – # Edit the team
  • Barrels and ammunition – most people, even great people are ammunition; what you need in your company is barrels to increase performance, velocity, firepower. you can only shoot through the number of unique barrels you have.
  • Add barrels, then stock them with ammunition
  • Barrels are incredibly difficult to find; when you do, give them lots of equity, promote them, take them to dinner every week; they’re virtually irreplaceable
  • Barrels are very culturally specific (barrel at one company might not be a barrel at another)
  • 16m01s – How do you figure out who is a barrel and who isn’t?
  • start with a small task as a test; smoothie test; expand scope of responsibilities until it breaks; keep pushing the envelope
  • watch who goes up to other people’s desk, particularly people they don’t report to; if you see that consistently, those are your barrels; they are building network to get things done
  • 18m12s – # Scaling: when do you hire someone above somebody; when do you mentor somebody?
  • every company has it’s own growth rate; every individual has they’re own growth rate
  • LinkedIn was slow (20 to 50 employees after 2yrs); Square was faster (20 to 300 after 2yrs); if employee’s learning rate steeper than companies growth rate then you can keep them in that role, else need to hire above
  • 19m17s – # Insist on Focus: where you point those barrels
  • Peter Thiel would insist on focus of individual; focus on one thing until you conquer it
  • A+ problems are hard and people procrastinate and instead solve B+ and lower problems; adds value and growth but not as much as if everyone were working on A+ problems 100%
  • 21m07s – # Metrics and Transparency: can’t make all decisions yourself so need ways to measure and track; dashboard
  • you as CEO need to draw out what metrics of success are; key metric of success is what percentage of employees use that dashboard
  • transparency – metrics first step, board decks review with employees (less compensation ok); every meeting with 2 or more people summarized in notes and sent to notes email alias so everyone could track what was going on; every conference room at square has glass walls; stripe give entire company access to all emails; compensation transparency (experiment)
  • 24m51s – # Gathering and Simplifying Information: goal is to predict output to enable you to adjust; measure output not activity; pairing indicators – measure effect and counter effect (e.g. fraud rate v. false positive rate to force innovation)
  • want to look for anamolies; paypal found ebay power users asking to be paid with paypal; at LinkedIn found high click rate of own profile tied to emotional vanity, helped clarify what users wanted;
  • 28m30s – # Details Matter: don’t focus on building billion dollar business, etc; that is by-product of getting all the details right
  • Bill Walsh took 49ers from worst to 3 superbowls over 10 years by starting with teaching receptionist how to answer the phone in 3 page memo
  • Details matter; debate is on items that don’t face the user
  • Example of Steve Jobs’ insitence of perfection of Mac curcuit board even though it couldn’t be open
  • Example serving bad food results in gossip and complaining instead of working, collaborating, brain storming
  • task rabbit leader: take distractions away and give people tools to be successful
  • you need to look for office space yourself; where people work everyday dictates culture and many other factors, how hard people work
  • effort: need a lot of effort to build a company and you need to lead by example
  • 32m32s – # Q&A
  • # How do you implement transparent compensation?
  • Few ways; one is with bands or tiers such as low experience ($85K) and high experience ($130)
  • 32m27s – # What kind of details do people care about?
  • Nice laptops, best possible tools, how do can you make people more successful
  • 34m15s – # How do you manage for best tools when have scare resources as start up?
  • Start with having your own office for company (not shared with other company) to create “cult”-like experience; figure out what’s most important; office very important
  • Can tell a lot about a company by their office space; how hard they working; how distracted they are
  • 35m55s – # What is best way to to gain street cred for a new manager?
  • Being excellent at something, then being promoted by merit; PayPal didn’t hire general managers, instead promoted best of each department to not demoralize other employees; can teach them to manage but hard to transition for individual contributor to manager (hardest is time allocation); get a mentor to focus on you
  • 38m03s – # How to have single voice?
  • Look at every piece of copy in every department; recruiting website; customer support (treat customer support like a product); cross-train different high-level executives (hires from Apple v. Google)
  • 39m12s – # What are some tactics of how do you manage people?
  • Have one-on-one roughly every week or two; only have 5 to 7 direct reports; agenda should be crafted by employee not manager, benefit is for employee; for strong competent employee could push to once every two weeks but never less frequently than monthly
  • 40m28s – # When do you compromise and hire more ammunition rather than a barrel?
  • You will hire more ammunition than barrels, question is ratio; engineering 1 barrel to 10-20 ammunitions; designers little different
  • avoid empire building philosophy in manager/barrel making hires by gauging performance as output divided by direct reports and be explicit; effect is number of direct reports will not grow until absolutely necessary
  • 41m56s – # How often do VC meet with company?
  • When leading a series, generally meet with founders every 2 weeks; and on ad-hoc basis for good and bad news
  • Meeting: tell me your problems; then have you tried this/ that; have you talked to this person; etc
  • 43m19s – # How do you prioritize recruiting?
  • Depends but if #1 priority then about 25%;
  • calendar audit for CEOs: ask them to rank priorities, then look at calendar; never matches
  • 44m52s – # How do you harmonize details matter with only focus on single A+ objective?
  • there is some tension; underlying philosophy of getting details right important to install in very beginning then new hires have and grow so that CEO is never the one doing it; culture is like that; culture is framework for making decisions

Additional Reading



Trackbacks / Pingbacks

  1. How to be a Great Founder | - 2014-11-10
  2. How to Manage | - 2014-11-18

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: