2/5
The first part of this book is flash fiction about founders. It's easy to read, but I did not find it particularly useful as examples went.
The rest of the book is considerably denser, and I found it harder to parse. There are good ideas in here (one objective for each team, looking at the 5 reasons goals aren't achieved, examples of good and bad objectives, arguments for limiting to one objective, and advice for taking a hands-off approach).
There are also inscrutable idioms ("We all know you don't pour whiskey into a flask through a leaky funnel in order to get drunk. For the same reason, you can't set 12 Objectives" (???)). Often examples would be provided with one being "good" and one being "bad". To me, the naive reader, the examples looked very similar.
My biggest piece of pushback is that the author suggests goals be graded two weeks before the end date as "Except in sales, it's unlikely you can pull off a hail Mary or miracle in the last two weeks". I don't think that's true! If the goal is "Get to 0 backlog items" and the team has been doing an average of 20 a week and there are 45 remaining, that one can go either way! I would say that's too early to call. I can think of a dozen other examples where calling it two weeks early is giving up. "Average rating of 8.0 with 100+ users" and we have a 7.8 with 100 users. A concerted push could make that happen! "$200k revenue" and we're at $188k.
I also personally would find it continually demotivating to set objectives with a 50% chance of success. If you're always stretching, you will get burned out. The book's answer to this is to have a "health index", where anyone on the team can raise an emergency. I have never worked on a team where someone would say "Hey I'm getting burned out, this is an emergency for everyone", especially if the design of the team is to continually be stretching. Health metrics in general were poorly explained.
I did not get a lot of out this book, but it did have some good ideas I will list below.
One OKR for the org
One for each team that ties to org's
KRs that tie to the OKR. COnfidence out of 10, starting at 5, reviewed weekly
Quadrant.
Top right O, KRs with rating
Top left top 3-4 priorities p1 and p2 only. These move towards the KRs that week. Weekly scope. Tasks.
Just big stuff, stuff should be aware or, or need help with
Bottom left: four week out heads up. Pipeline coming
Bottom right: health metrics
Why aren't goals achieved?
- Lack of priority
- Not communicatd clearly and often enough
- No plan in place
- Haven't made time
- Give up instead of iterating
Need to frequently (e.g. weekly) track progress towards OKRs. If you set them and ignore until end of quarter, you'll be surprised at how little progress has been made.
Pick which goal matters the most and prioritize it. Don't be greedy and try to do everything.
Clarify what that goal looks like: when is it accomplished?
State the goal clearly until everyone undersatnds and pursues it
Make a plan to pursue it even when disheartened
Dedicate time to accomplishing the goal. Tomorrow never comes.
Be ready to fail and try again.
Objectives are qualitiative
Key Results are quantitative
Key results should be outcomes not tasks (user complaint calls down, not sales team completes training)
BAD OKR
O: New self-serve help area
kr: better search
kr: new FAQ
kr: Forums
GOOD OKR
O: The company helps customers succeed when they are strugglign
kr: "Did this help" rating up 15%
kr: "Problem resolved" rating on FAQ up 30%
kr: Peer-to-peer help forum has DAU of 2k (down from 10k)
If a result is a result, you can switch tactics. If a result is a project, you are locked into a project even if it's not helpful
Simpler is better. Want to hold them in memory.
What's a good metric?
Comparative. Compare vs other time periods, users, competitors. Increased conversation from last week is better than 2% conversion
Understasnable. If people can't discuss or remember it, it's hard to turn it into culture change (what is a sticky script?)
Ratio or rate.
Changes the way you behave. What will you do differently based on changes in the metric?
Choose ONE objective at a time and give it your focus
"In most cases an engineer should spend an extra week reducing the complexity of an application vs making millions of users spend an extra minute using the application due to complexity"
It's more important to be effective than efficient. Efficiency kills innovation
OKRs only work if you set the objective and surrender control of tactics. Let your team choose the approach
Learning can be done 3 ways, most people skip the last.
6. Instruction. Read a book or be taught. Least useful!
7. Action. Do the thing. Very useful!
8. Reflection. Discuss, analyze, write an essay about it. Overlooked! Very useful!