After a pilot with 20–30 teams, organizations often ask:
“Did it work?”
“What should we roll out next?”
Both questions are premature.
At this scale, a pilot is not validating practices.
It is revealing how your system works—and what it lacks.
A Better Question
Instead of asking what worked, ask:
“What capabilities and conditions should be present—and where are they missing?”
Two Things a Pilot Reveals
A large pilot reveals two different dimensions:
1. System Conditions
Flow of work
Dependencies
Work in process
Feedback timing
2. Organizational Capabilities
Ability to define clear objectives
Ability to prioritize what matters
Ability to solve problems quickly
Ability to learn and adapt
Most organizations focus only on the first.
But improvement depends on both.
Start With Two Critical Questions
Before analyzing flow or performance, ask:
1. Do We Understand the Essential Stakeholders?
Who are the critical stakeholders?
What do they value?
What are their success criteria?
What constraints matter to them?
If this is unclear:
Teams may be executing well—but on poorly defined improvements
This is not a flow problem.
It is a problem definition problem.
2. Is Work Being Done in the Equivalent of Quickest Valuable Releases? (QVRs)
Are we sequencing work to deliver value quickly?
Are we making trade-offs explicit?
Are we learning from small, meaningful increments?
If not:
The system is likely batching work, delaying feedback, and obscuring value
This undermines both flow and learning.
Patterns to Look For
Once those are addressed, examine the system.
1. Flow Patterns
Where does work slow or accumulate?
What conditions are preventing smooth flow?
2. Demand Alignment
Are teams working toward shared improvements?
Or reacting to competing requests?
3. Work in Process
Is too much being started?
Are teams able to finish?
4. Feedback Timing
How quickly does the organization learn?
Where is rework discovered?
5. Objective Clarity
Are teams solving clearly defined problems?
Or:
Working from vague requests
Delivering output without measurable improvement
Often:
Teams that outperform others are not better at execution
They are working on better-defined improvements
6. Variation Across Teams
Why are some teams more effective?
Look for differences in:
Stakeholder clarity
Alignment
Dependencies
Access to shared services
Variation is not noise.
It is signal.
A Simple but High-Leverage Check: Sprint Length
If your pilot used:
3-week sprints (or longer)
Then:
Shorten them to 2 weeks or less
Why this matters:
Faster feedback
Earlier detection of problems
Reduced batching
Improved coordination across teams
Longer cycles often hide issues that shorter cycles expose.
This is not about Scrum compliance.
It is about improving the speed of learning and adjustment
From Symptoms to Capabilities
Instead of listing problems, identify:
What capabilities are missing?
For example:
Misalignment → lack of prioritization capability
High WIP → lack of disciplined work selection
Delays → lack of flow management
Poor outcomes → lack of stakeholder clarity
What To Do Next
Step 1 — Identify Key Gaps
One primary system condition (i.e., how our system is holding us back)
One primary capability gap (i.e., where we don’t have sufficient capability)
Step 2 — Improve Both
System → improve alignment, flow, or WIP
Capability → improve how problems are defined and prioritized
Step 3 — Use Enabling Improvements Carefully
Improve local areas—but ensure they:
Reveal system behavior
Do not hide underlying issues
When You Can’t Change Direction
Often, teams cannot change what they are asked to do.
In these cases:
Improving flow and visibility becomes a way to demonstrate the cost of current decisions
This creates:
Evidence
Awareness
Pressure for change
From Building to Improving
The most important shift:
Don’t ask what teams delivered
Ask what was improved
This:
Anchors work in value
Enables prioritization
Aligns decision-making
Final Thought
After a large pilot:
You are not deciding what to scale
You are learning how your system behaves
And improvement begins when you see:
What is missing—not just what is broken
See the previous post that this builds on When There Is Not A Single Constraints
AmpGPT now has a diagnostic mode.
If you like my work, you will find AmpGPT of interest. It is ChatGPT, specially trained in Amplio, to interact with you to help you learn. Amplio is not a framework but is built on a solid foundation of system, cognitive, and learning dynamics. It can assist you regardless of the approach you are taking. You can learn more about it here.
I have recently added a ‘diagnostic mode’ which can guide you in performing a diagnostic on your situation. To learn more, and to use it, go here.