Most organizations plan a transformation as if go-live is the finish line. The AMIGA Framework structures transformation into five stages instead — Clarify, Create, Implement, Test, and Harvest — and the fifth stage, Harvest, is the one most programs skip almost entirely, which is a major reason benefits never get proven.
The Five Stages
Clarify. Defining the actual problem, scope, and success criteria before any solution design begins. Programs that skip or rush Clarify tend to discover disagreements about what the program was even supposed to achieve much later, when they're far more expensive to resolve.
Create. Designing the future-state processes, technology architecture, and organizational changes. This is where governance structures, data migration approach, and change management plans get built — not improvised later under schedule pressure.
Implement. Building and configuring the actual solution — the phase most people picture when they think "the project."
Test. Validating that what was built actually works, covering unit, integration, and user acceptance testing against realistic data and scenarios, not just a clean happy-path demo.
Harvest. Benefits realization and sustainment — extending 12 to 24 months beyond go-live. This is the stage most organizations incorrectly treat as the program's endpoint rather than its actual conclusion.
Why Harvest Gets Skipped
By the time a program reaches go-live, the team is usually being reassigned, the budget line is closing out, and organizational attention has already moved to the next initiative. Nobody is explicitly staffed to keep tracking whether the promised benefits actually materialized — which is exactly how roughly 73% of organizations end up unable to prove ROI on a transformation that, by every earlier measure, looked successful.
A Roadmap Is Not Just a Timeline
Treating these five stages as a simple linear timeline understates what each one requires. Each stage has its own governance gates, deliverables, and success criteria — a program that treats "Implement" as just "the middle chunk of the schedule" tends to under-invest in the specific decisions that stage actually requires.
Plan for Harvest From the Start
The single highest-leverage change most organizations can make to their roadmap is staffing and budgeting for Harvest explicitly, at the same time the rest of the program gets planned — not deciding on it after go-live once attention has already moved on. The Benefits Tracking Template is built specifically for this stage.
Get the full AMIGA Framework Quick Card →
Frequently Asked Questions
How long should each of the five stages take?
It varies significantly by program scope and complexity, but Harvest specifically should be planned for 12 to 24 months beyond go-live — a duration many organizations don't budget for at all.
Can stages overlap, or do they have to run sequentially?
Some overlap is common in practice — Create and early Implement work often run in parallel — but Harvest specifically can't meaningfully begin until go-live, since it depends on measuring real post-launch behavior.
What's the biggest risk of rushing the Clarify stage?
Disagreements about scope and success criteria that weren't resolved in Clarify tend to resurface during Implement or Test, when they're far more expensive and disruptive to address than they would have been at the start.
Instagram