This document discusses strategies for distributed decision making and innovation. It advocates for empowering small, cross-functional teams to drive their own strategies and priorities. It also discusses the importance of establishing a culture of risk tolerance by focusing on "painful risks" rather than risks that only cause embarrassment. The document highlights the success of LinkedIn's Hackdays program in fostering inspiration and innovation by dedicating time every month for employees to experiment with new projects.
7. Distributed Systems
High parallelization
Loose coupling
N-tier architectures
Fault tolerance
Massive scalability
Why donât we build organizations this way?
8. What do we do instead?
Linear processes, without parallelization
Tight coupling, often with single points of failure
Fault tolerance generally limited to under utilization
Extremely limited scalability, even with winning teams
9. What do we do instead?
Linear processes, without parallelization
Tight coupling, often with single points of failure
Fault tolerance generally limited to under utilization
Extremely limited scalability, even with winning teams
Who would architect a system like this?
12. Distributed Organizations
Small, cross-functional teams empowered to drive their
strategy, prioritization and execution
Common, harmonized goals with unambiguous definitions
of success
High degree of independence between teams in terms of
both process and technology
Constant, redundant and consistent communication
between teams and up-and-down the organization
Clear definitions of exceptions, and protocols for escalation
14. Optimus Prime
âSometimes even the
wisest of men and
machines can be
in errorâ
-- Transformers (1984)
15. High Risk, High Reward?
Most common incorrect statement made about
investment decisions
The market does not reward risks that can be
diversified away
Stupid risk does not add value
In a rapidly changing technology landscape, the risk of
not moving fast enough is huge
16. High Risk, High Reward?
Most common incorrect statement made about
investment decisions
The market does not reward risks that can be
diversified away
Stupid risk does not add value
In a rapidly changing technology landscape, the risk of
not moving fast enough is huge
Whatâs the right framework for risk in high tech?
17. Three types of risk
Mortal risk
If this goes wrong, the venture literally is over.
Painful risk
If this goes wrong, you face a significant issue with a
key metric (market share, revenue, usage)
Embarrassment risk
If this goes wrong, you have to admit you were wrong.
18. Three types of risk
Mortal risk
If this goes wrong, the venture literally is over.
Painful risk
If this goes wrong, you face a significant issue with a
key metric (market share, revenue, usage)
Embarrassment risk
If this goes wrong, you have to admit you were wrong.
Most companies spend too much time on
embarrassment & donât take enough painful risks.
19. Painful Risk is Reality
Technology is hyper-competitive
Great projects often include commitments to deliver
solutions to problems that havenât been solved yet.
The landscape changes so often, the âright solutionâ to
a problem is often sub-optimal within a few years.
Moderate risk with experimentation, iteration, and
minimally viable products.
Reward risk by sharing learning & owning the decision
21. Optimus Prime
âFreedom is the right of all
sentient beingsâ
-- Transformers (1984)
22. LinkedIn Hackdays
One Friday, every month
Projects due by Monday,
9am
Projects displayed on intranet
Projects publicly judged by
executives
Prizes. Glory. Real Success.
23. Why does it work?
In software, you donât learn deeply by reading or
discussing. You learn by doing.
It takes dedicated time to experiment and learn a new
platform, new architecture, new algorithm.
Small teams learn together.
Inspiration is contagious.
24. Inspiration â Innovation
Most of the limits that we face we place on ourselves.
Competition challenges and inspires
Leadership behavior and incentives matter, as they can
easily confirm or deny prioritization of innovation
Culture of innovation requires peer-based inspiration
25. Inspiration â Innovation
Most of the limits that we face we place on ourselves.
Competition challenges and inspires
Leadership behavior and incentives matter, as they can
easily confirm or deny prioritization of innovation
Culture of innovation requires peer-based inspiration
We inspire each other to do our best work.
26. More on LinkedIn Hackdays
http://blog.adamnash.com/2011/05/05/why-linkedin-hackdays-work/
28. Organization, Risk & Culture
Innovation begins with empowerment. Small, cross-
functional teams have the data, the passion, and the
ability to execute. Empower them.
Itâs easy to talk about being risk tolerant, but which
types of risks, and when? Beware spending too much
time on embarrassment risk.
Donât run systems at 100% of capacity. Engineers
need some amount of bandwidth to learn, experiment
and execute on new ideas.