A transformation roadmap executives can actually steer.

Five phases with decision gates

Diagnose → Target → Foundation → Build → Scale — each phase ends with an explicit go/no-go decision and the evidence it requires, so momentum never substitutes for judgment.

Capability-anchored workstreams

Workstreams organized around what the business does — not around vendors, departments, or whichever system shouts loudest.

The metrics that keep it honest

Leading and lagging indicators per phase, plus the three failure signals that tell you a transformation is drifting months before the budget review does.

Phase 1 · Weeks 1–6

Diagnose — know what you actually have

Transformation fails most often at the start, by skipping the unglamorous work of knowing the current state. This phase produces the three artifacts every later decision depends on.

• Capability map (level 1): the 8–12 things the business does, stated in business language, validated by executives outside IT.
• System & data inventory: every application mapped to the capability it serves, with annual run cost and a named owner. Include the shadow IT.
• Friction register: the top 10 places where work stalls — re-keyed data, swivel-chair processes, reports nobody trusts — each priced in hours or dollars.

Gate 1 — proceed only when: executives agree the capability map describes their business, and the friction register is priced. If you can't price the pain, you can't prioritize the cure.

Phase 2 · Weeks 6–12

Target — decide what good looks like

Now — and only now — define the target state. Anchor it to capabilities, not products: "order fulfillment quotes real-time inventory" survives vendor changes; "implement System X" doesn't.

• Capability targets: for each critical capability, one sentence describing how it performs in three years, with the measure that proves it.
• Disposition per system: invest, maintain, tolerate, or retire — every system in the inventory gets a verdict and a rough date.
• Workstream slate: 3–5 workstreams (not 40 projects), each owning a capability cluster, each with an accountable executive sponsor.
• Sequence by dependency: data and integration foundations before the experiences built on them. The order is architectural, not political.

Gate 2 — proceed only when: every workstream has a named sponsor, a capability target, and a first-90-day deliverable. Unowned workstreams don't start.

Phase 3 · Months 3–9

Foundation — build the plumbing once

The least visible phase and the most decisive. Integration, data authority, and governance foundations determine whether everything after ships in weeks or in quarters.

• Integration backbone: a standard way systems talk (APIs/events), so every future connection is a pattern, not a project.
• Data authority: one named authoritative source per core entity — customer, order, product — and a plan to demote the pretenders.
• Delivery governance: environments, pipelines, and security guardrails standardized, so teams ship safely by default.
• First visible win: pick one priced friction item and remove it end-to-end. Foundations earn patience by paying early rent.

Gate 3 — proceed only when: a new integration takes days not months, core entities have one authoritative source, and the first win is live and measured.

Phase 4 · Months 9–24

Build — transform capability by capability

Workstreams now modernize their capabilities in slices — each slice shipping to real users, each retiring something as it lands. Never a two-year build with a big-bang cutover.

• Ship in 90-day slices: every slice delivers working capability to real users and produces evidence for the next investment decision.
• Retire as you build: every new capability names what it replaces, with a decommission date. If nothing retires, you're adding, not transforming.
• Automate where it pays: layer AI and automation onto capabilities with clean data and measured processes — in that order, not the reverse.
• Quarterly rebalance: re-rank the slate against capability targets each quarter. Kill or pause what the evidence no longer supports.

Gate 4 — checked quarterly: each workstream shipped user-facing value this quarter, retired at least one thing this year, and its capability measure is moving. Two misses = restructure the workstream.

Phase 5 · Month 18 onward

Scale — make it the operating model

Transformation ends when its disciplines stop being a program and become how the organization runs. This phase dissolves the program into the operating model — deliberately.

• Capability scorecard: capability health and run cost on the executive dashboard, trended quarterly, owned by business leaders.
• Intake guardrails: every new initiative names its capability, its owner, and what it replaces — the portfolio can no longer sprawl silently.
• Continuous modernization budget: a standing allocation (not a special program) for retiring, consolidating, and upgrading — so the next transformation is never needed.
• Annual re-diagnosis: re-run the Phase 1 diagnostics every year. Drift is inevitable; noticing it within a quarter is the discipline.

Exit criteria: the program office dissolves, the disciplines persist, and the capability scorecard — not a steering committee — steers investment.

The three drift signals

Watch for these monthly. Any one of them predicts failure earlier than any budget review will.

1. Activity replaces outcomes: status reports count workshops and milestones instead of capability measures moving.
2. Nothing is retiring: the estate is growing. You are funding an expansion, not a transformation.
3. Gates stop being decisions: every gate passes automatically. Momentum has replaced judgment, and the roadmap is now a schedule.