OKRs: How Andy Grove's Intel System Reached Google Through John Doerr

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

Objectives and key results, or OKRs, grew out of the hardware labs of Intel in the 1970s under Andrew Grove. The dominant system of the day was management by objectives, pioneered by Peter Drucker in the 1950s, but its goals were handed down from the top, set once a year and tied to pay. Grove, working in a microprocessor industry that moved at breakneck speed, built a version called iMBOs that updated quarterly, ran from the bottom up as much as the top down, and was divorced from salary and bonuses. He documented it in his 1983 book High Output Management. In 1975 a young Intel salesperson named John Doerr took Grove's internal training course, and by 1999, as a venture capitalist at Kleiner Perkins, he pitched the system to a chaotic startup in his portfolio: Google.

Larry Page credited OKRs for keeping Google on track and for its tenfold growth, and Doerr's 2018 book Measure What Matters spread the method to LinkedIn, Twitter, Uber, Microsoft and GitLab. The framework splits work into inspirational objectives, three to five measurable key results, and initiatives. Its strangest rule is scoring: Doerr recommends a 70 percent success rate on aspirational key results, so a team that always hits 100 percent is setting goals that are too easy. Committed key results, such as a legally required privacy feature, still target 100 percent. The method has also had to evolve, as cascading individual OKRs turned into task lists tangled up with performance reviews.

• Grove's quarterly, pay-free design was meant to stop employees from sandbagging their goals to protect their bonuses.
• A key result must allow a binary answer at quarter's end: the number was hit or it was not.
• Organizations are told to track leading indicators such as weekly active users rather than lagging ones such as quarterly revenue.
• Vague verbs like help or consult, and business-as-usual objectives like maintaining server uptime, are named as common pitfalls.
• Rick Klau popularized cascading OKRs in a 2014 Google Ventures workshop, then reversed himself in 2017 and advised skipping individual OKRs.