Video: Why Scaling Agile Doesn't Work - Jez Humble
Why Scaling Agile Doesn't Work • Jez Humble
ROI - cost estimates don't matter
prioritize based on cost of delay
"how much would we have earned if this was in production"
We're more wrong than not in guessing which features will be valuable
(90% wrong in uncertain/new markets, xx>50% in well known markets)
So what do we do?
impact mapping
work backwards from the outcome
- Target
- Stakeholders
- ways in which the stakeholders could help/hinder this outcome
- what you could do aka. "Story"
key thing = shared understanding by creating the impact map together
Different types of user research
Generative = brainstorm
Evaluative = decide
(Cf diamond)
ways that features cost
1) opportunity cost
2) complexity maintenance cost
3) complexity slows down adding new features
======
CASE STUDY
hp laserjet firmware
goal:
- get firmware development off the critical path
- 10x productivity increase
1) goal: firmware off critical path + 10x productivity
2) current condition:
3) target conditions
"these are the difficult ones"
(eg. plan for one month for entire 400 ppl org)
4) define outcome and let people experiment
daily questions (for the ppl)
======
How do you introduce this in your org?
1) Guerillia experimentation
Find someone in your org you think is behaving stupid.
Take them out to lunch.
See why they do what they do.
Improve a small bit.
eg. That person from UX has yet again sent a crappy screen
2) Book: leading change - john kotter
No lasting change without executive support.
Only through disaster.
"find someone feeling the disaster, they are more receptive to change"
sense of urgency