Product Over Project
Three TWers and Customers
Inception
Prototypes
Master Story List
Estimates
Release Plan
First release in 10 months
After two months
Velocity is low
Team working harder
..and is accumulating debt
Customer is nervous
Another month goes by
Finger pointing
Developers: A lot of this was not part of original scope
Customer: That is because of bad analysis
Analysts: Only if we were building stuff fast enough
Release in 20 months at current velocity
Estimation, Replanning and Contract Negotiation
Customer: You guys should fix bid it
Development Team: Looks like we have screwed up big time. We should
work on weekends
TW: Guys lets make this happen
TW: We are giving discount
Almost five months!
Developer: Worried about quality of code
Can fill a whole newspaper with technical debt
QA: Worried about lack of automation
Analysts: Not keen talking to customer
TW: Do we have to give more discounts?
WHAT ARE WE DOING WRONG?
1
We are not good at this game?
Every product is different
Software development is discovery process
Delivery is the only test of estimation
Everything else is just opinion
2
Advice needs to be better than…Think out of box
Be focusedBe creative
Stay commited
“Trying harder is not an improvement strategy” – Jason Yip
3.
Planning (estimation) is important
Plans (estimates) are not
4. Most Important
This is the wrong game?
What are we building anyway?
Why is no one talking about that?
Tail is wagging the dog
Project is defining the product
Don’t trust me
Look at your project status report
Listen to people in standup
So, what’s the proposition?
1.Customer pays for software
10 Months is too long
Do Shorter releases
Find ways to do it. Don’t accept customer’s word for it.
Internal release if nothing works
Internal release should also meet all criteria
Customer Signed-off != Done
Focuses everyone on the next release
Its not how fast we are going
but… where we are going?
Effective over efficient-- Dan North
2.Build the right thing
At the right time
Again..software development is discovery process
Yes, even for the customers
Need onsite presence to define the product
Not only to……to get story signed-offfor backlog creation
for feature estimation
Send best people onsite……instead of keeping them to keep
velocity high
Creating a product which no one wants is worst kind of waste
Building things ahead of time is waste
Play the right game!
