Measure what Matters: John Doerr

Book cover

Below are some notes while reading Measure what Matters for my own future reference. I understood shortly before reading this book how important goal setting is, as an individual contributor. No one can know what I should work on better than myself. OKRs seem to be a proven methodology for establishing goals and how one will achieve them, but they take time to take effect. First few quarters will have various problems such as the OKRs being too rigid for unseen adaptations.

• OKRs invented by Andy Grove, who was VP and then CEO of Intel 1978-1980s
• Google, Larry Page and Sergey Brin, met John who invested a lot of money into startup Google
• They accepted John's present of OKRs as a trial
• Page was convinced that their search engine was better than all competitors by a long-shot and that their company would change the world from the start
• OKRs are transparent: a frontline engineer can look up the CEOs OKRs and analyze how her own goals ladder up and align
• OKRs should not be tied to promotion or performance evaluation. They are a tool of the contributor, judge the contributor not on the tool.
• Use CFRs instead for performance
• Google flips management:employee ratio to at most 1:7, smaller ratio is desirable to create flatter hierarchy
• 4 superpowers of OKRs:
• Focus and Commit to priorities
• align and connect for teamwork
• track for accountability
• stretch for amazing
• John called Bill Davidow, highest ranking Intel person while trying to find an internship, and he was invited to meet him because John knew how to write benchmarks
• Gordon Moore: shy and revered deep thinker (moore's law)
• Robert Noyce: co-inventor of Integrated circuit
• Andras istvan Grof: ? 

It doesn't matter what you know. It's what you can do with whatever you know or can acquire and actually accomplish that tends to be valued here

• ancestor of OKRs was MBO (management by objectives)
• it was good, theorized by Peter Drucker as a wrecking ball to Taylor-Ford (henry) model of top-down scientific explicit micromanaging of work goals
• did not include KRs or how to accomplish objectives
• was annually reviewed
• private and siloed, top-down, tied to compensation, and risk-averse
• john started from intern, went to finish at Harvard, worked as PM for 18 months, then went out and did sales in Chicago for Intel
• returned as marketing manager and tried to apply OKRs on his management crew
• Operation Crush:
• Intel facing competition in microprocessor chip from motorola 68000 and Zilog Z8000
• a sales employee noticed the competition and brought it up to his manager, who got it in front of Andy Grove. This launched Operation Crush
• Shows how OKRs but also healthy company culture allows bottom-up development and alignment because the trenches first figure out what's going wrong and what's working before C-class execs
• not everything a manager or director does can be encompassed by OKRs
• Within 3 weeks, Intel made a plan on how to approach sales of their 16-bit microprocessor 8086 and beat Motorola
• Motorola couldn't respond as indicated by John's contact in the comp:

I couldn't get a plane ticket from Chicago to Arizona approved in the time you took to launch your campaign

• Intel did not have to out-engineer their competitors for Crush. They had to sell a narrative to their customers. i.e. broad product family and superb technical support along with Intel design aids. Motorola is big and diverse company whereas Intel focused solely on microprocessors, so they can probably help or do better
•  leaders have to model OKRs and not just say "my goals are my team's goals"
• esprit de corps: a feeling of pride, fellowship, and common loyalty shared by the members of a particular group.
• Jeff Weiner on OKRs every quarterly all-hands meeting:

When you are tired of saying it, people are starting to hear it

• Pairing key results; it is important to not just have quantitative KRs but also KRs that check/audit them
• case and point: Ford Pinto. "under 2000 lb and $2000 automobile" meant safety was not on the OKR so hundreds of people died.
• the more ambitious the OKR, the greater the risk of overlooking a vital criterion
• have "effect and countereffect" KRs
• example of strong OKR, green indicates the additional KRs that make this strong
• Objective: Win the Indy 500
• KR: increase speed by 2 %
• KR: Test at wind tunnel 10 times
• KR: reduce stop time by 1 sec
• KR: reduce pit stop errors by 50%
• KR: practice pit stops 1 hr/day
• good company leadership know that there are a million things to do, but you can't do everything in 3 months. Prioritize 3-4 things and "put more wood behind fewer arrows" as Larry Page says.
• Superpower of OKRs 1- Focus, the Remind story
• CEO Brett Kopf dedicated to helping teachers provide a platform through text to teach and provide extra help to students outside the 8am-3pm school schedule
• He has ADHD/dyslexia, was not confident in school himself, but worked hard thanks to mentorship from one HS teacher and went to college.
• Remind is Twitter for Education
• Only measure and care about what is essential. For example, number of registered users for Remind is inessential whereas weekly active teachers, monthly AT, retention
• John pitched OKRs to Brett and his brother and example of focus on what matter

Teachers want to be able to remind a 5th grade class to bring novel every Monday morning so this would be a nice engineering feature, but is this top priority of the quarter? Based on OKRs, the answer was no so we decided to shelve it- a tough call for a teacher-centric organization

• Ease into OKRs when trying them in an organization first time. Need buy in from executives before the ranks agree
• Commit: the Nuna Stora:
• Jini Kim, a former Google PM, was inspired to launch a healthcare startup to help people like her brother who is diagnosed with severe autism.
• no orders till year 2- crashed HR conferences and talked to benefits directors to learn what is important to them
• mistake with Nuna was to project OKRs on whole team of 20 ppl instead of starting it small with just leadership team
• Can give a number between 0-1.0 to evaluate how close you came to accomplishing the KR (a google tactic)
• They received huge government contract to create federal database of 74M American health records in central place

We don't hire smart people to tell them what to do. We hire them to so they tell us what to do- Steve Jobs

• Alignment for OKRs is important so should it be done top-down or bottom-up? 50-50
• If done top-down, lack of agility, marginalized contributors who will fear trying to change up the orders from "up above", one-dimensional linkages preventing horizontal links across depts, 
• orgs can be more directive when urgent times require it, but allowing your contributors to work and decide for themselves how and what is important to work on can help propel the team more. They buy-in which is all around better than manager overhead
• My fitness Pal Story- Mike Lee and his brother
• it got acquired by Under Armor for 450M, but challenges there once Mike joined and took leadership of multiple apps beyond MyFitnessPal included that multiple C-suite execs had expectations or dependencies already assumed from Mike's org. They said you have to do this this quarter...
• Mike learned OKRs from John Doerre in 2013; more people in org (10 -> 30) meant slow down in progress because need to train people and prevent overriding of duplicate work

OKRs can be we well-crafted, but implementation fell short. You have to make the dependency explicit. Coordination was hit-and-miss , with deadlines blown on regular basis.

• OKRs should have a single owner; co-ownership weakens accountability
• It's ok to start a project with somewhat ambiguous OKRs without metrics, but that should be your first OKR (establish a metric for future OKR-setting improvement)
• Under Armor did not use OKRs so Mike engaged in 1-way transparency by sharing his org's OKRs and getting buy-in from leaders
• Zume Pizza, Coursera, CFRs = Conversations, Feedback & Recognition
• End of book is about culture, the switch from annual performance reviews to continuous feedback on a monthly basis, using Betterworks for that system
• The end of the book has an awesome summary of the entire book. For me, here is a checklist to measure my OKRs every time:

Objective

• Does achieving this objective provide clear value to company?
• Do you differentiate between committed vs. aspirational goal [lower threshold of completion 1.0 vs 0.7]
• If you ask your customers what they really want, does your aspirational objective meet or exceed their request? 
• (a litmus test against timid aspirational objective)
•  Could the OKR get a 1.0 under reasonable circumstances without providing direct end-user or economic benefit? 
• Don't commit to low-value objective, litmus test. 
• If so, reword objective to focus on the tangible benefit. Instead of "define x", "Establish definition for x and publish on public with versioning"
• Does your objective fit on one line?

Key results

• Does your KR exclude any team-internal terms?
• Impact of work > actual work
• Are you using real dates to qualify your KRs?
• Is your KR measurable and the metric is unambiguous?
• Does achieving all your KRs complete your objective?
• If it doesn't, that means you are not writing down the difficult commitments required to finish your objective (taking easy way out)
• Does your KR describe end-user impact and not activities. 
• not "assess Colossus latency" but "publish average and tail latency measurements from 6 Colossus cells by March 7"
• Did you include evidence of completion for your KRs?