Evermore people are talking about Discovery and Hypothesis-driven approaches. But where do you start? What do they really mean?
Pedro will share with us how he moved away from a 2-year delivery roadmap by enabling his Engineering teams to do a Dual Track Agile. A real case-study!
Key Learning Points:
- Understand what Dual Track Agile is
- Learn why Pedro and his team decided to use it at OutSystems
- Know what was the strategy in place for the Change Management
- Understand their failures and what they have learned with it
- Identify some Common Pitfalls
- Understand the importance of cadence for alignment and trust
- Understand the importance of building (truly) autonomous teams
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
Implementing Dual-Track Agile :: Lessons from the trenches @ITSpring.by May 2019
1. Dual-Track Agile
Lessons from the trenches
Pedro Teixeira
May 31, 2019
@_pteixeira
https://linkedin.com/in/teixeirapedro
http://napkin-talks.com
2. About
● 15 years working in Product-led companies
● First encounter with Agile was in 2008
● Co-organizer of AgileConnect Lisbon
● What this talk is not
Pedro Teixeira
Product Agility Coach
Engineering
160+ people
@_pteixeira
https://linkedin.com/in/teixeirapedro
http://napkin-talks.com
7. Dual-Track Agile
Earlier and Faster Feedback-Loops
Fast Learning and
Validation
Can they use this?
(Usability)Do our customers
need/want this?
(Value)
Can we build this?
(Feasibility)
Shippable working
product
credits to Jeff Patton and Marty Cagan
8. Do the users need/want this?
(value)
We assume ours
customers wants this
Let’s experiment!
If we have 1000 clicks our
assumption is right
#1
9. Can they use this?
(usability)
Can we build it?
(feasibility)
#2
25. Experiment in your process
High-Performance does not come for free
Celebration is important … but don’t forget to be leaner in your ways of work
(e.g. invest in achieving the same outcomes with less process)
26. Be a team of grown-ups
Scale if you are in trouble
Decision-making deadlock
No continuous-improvement
unfit process
28. Outcome over Output-factory
Problem Validation before Optimizations
Adaptive Plan over Strict Roadmaps
Speed over perfection
Learning over right and wrong
Right Data over opinions
Build while doing Fast Discovery
Dual-Track Agile
29. Discovery is not optional
But do it fast
Be intentional in your experiments
They should be actionable, time-boxed with a clear question to be answered
Plan everything for a given cadence
You keep the customer interest, the constant feedback flow-in, measure progress and have
better alignments on dependencies.
Hi everyone
My goal for today is to
Give you a small introduction on Dual-Track Agile
Share some lessons I’ve been learning in my teams that in Dual-Track Agile
Who is not familiar with Dual-Track Agile?
I’m one of the co-organizers of the AgileConnect Lisbon that happens every month,
If you happen to be in Lisbon
have a look on the meetup and feel free to come by
And I love to share my experience and thoughs either on meetups, conferences like this one and on writing on LinkedIn and my blog
Currently I help a Engineering department to be deliver the right product and on time.
Come with me to a typical journey of a Development Team. Let’s call them the orange team
Orange Team, Scott, Paul and Peter
6 Months ago they started Working in a new Product
Running perfect sprints where their velocity was steady and increasing over time
Over nights to deal with perfrmance issues
First release after few months - experience
New apikes, PoC and embarked in a serie of code refactorings.
This is the approach #1 that I call Single-Track
This is what I call the Single-Track Development
They started with their best idea. Perhaps the only idea that came out from their most senior person in the room
Validate the idea with product
Leads to technical waste, misfit product and a demotivated team that lost the trust of their leadership.
On the top layer the we have the fast learning and validation of ideas.
First you talk with a customer, you take 2 months to have a stable version at the ends of the customer, and then we wants something different
You need to do Continuous Discovery to lower the risk and validate those three aspects. Specially the revelante of what you are building
Discovery for everything => Everyone wants to get backed-up with data to validate all their assumptions
Decision-Making is hard and slow => This causes having too many data pointers to analyse, process, debrief and act up on
Development is in idle, waiting => Leading to delay releases, because we are aulted waiting
Dual-Track duel => and because managers start to ask, why is this not done yet we become having finger-pointing between people, specially between the tracks. And this breaks the team in two.
WHen
Let’s put the designations Single-Track and Dual-Track aside. What do we really want?
Now I want to share my lessons until now, and I want to start by trying to use what you really want instead of using this metaphor that can be misunderstood
We don’t want to Discover. Must sometimes we need to validate some ideas sooner.
You had data a certain feature was not being used
You assumed the experience was not good
You changed the exprience and the data was still low
If you had spent time with the customer you could verify the link was hard to see on the homepage - and that was the reason!
I got introduced to this approach by Mike Burrows at Aginext London this year and aí found it a simple and elegant way to ensure we are being outcome driven.
Why? Because we first neet to identify what are the outcomes, how we validate their accomplsiment and then we go backwards creating the backlog to ensure this happen.
Let me explain with Legos.
The minimum working-wow-Product important but it will allows us to focus our work and then take a decision based on the the customer acceptance.
There many representations for what work cycle, I like this one because is simple and captures the important steps.
I think we all agree that a development team has a limited capacity to work and focus. To do all the activities they need to fulfill all this cycle.
Incorporating Fast Discovery is not easy
The best teams I know are the ones that do experiments in their process. Try new ways to store the information from the interviews of yours customers
They have someone who is accountable to make it happen and brief the results with everyone.
New way of estimate
New format for Sprint Reviews
A new virtual board for remote teams
A new technique for A/B testings
High-Performance does not come for free.
Brings a early stage called Product Discover
That Marty Cagan, one of the most inspiring and though Product Coaches define as a …”
Basically to help us avoid having a misfit product after 6 months coding.
Let’s put the designations Single-Track and Dual-Track aside. What do we really want?
Two tracks, two mindsets that work side-by-side within the team
One that is focused in discovering which are the most important problems to solve today
And other, that once those problems are captures, make sure they are built right, performant, securely and with a great experience.
Everything happening at the same time.
Orange Team, working in their product backlog in a serie of Sprints.
This is what I call the Single-Track Development
They started with their best idea. Perhaps the only idea that came out from their most senior person in the room
Validate the idea with product
Leads to technical waste, misfit product and a demotivated team that lost the trust of their leadership.
This is the approach #1 that I call Single-Track