Making Organisational Change Irresistible: Phil Gilbert's Product Approach

Executive overview

Most change programmes fail not because employees resist, but because leaders treat change as a mandate rather than a product. Phil Gilbert led IBM's design transformation across 400,000 people without direct authority over any of them — by applying startup go-to-market thinking to internal change. The core reframe: if people don't adopt the change, it is the change leader's fault, not the employee's. Sustained cultural adoption, not improved competency, must be the north star of every change decision.

Change as a product

• Change programmes share the identical problem every startup faces: nobody knows the thing exists, inertia resists adoption, there is no mandate to use it.
• Treating change as a product shifts accountability entirely — if teams don't embrace the change, the programme team must diagnose why, not blame employees.
• The phrase "they just don't get it" was explicitly banned; non-adoption was treated as a product-market fit failure, not an employee failure.
• Mandates create compliance theatre: measurement systems show compliance, but outcomes don't actually improve.
• Agency — the genuine freedom to opt out — makes adoption 1,000x more sustainable than compliance.

Selecting the first teams

• Seed the market exactly as a B2B startup would: identify two or three friendly, early-adopter teams and work with them intensively before scaling.
• Tiger teams and innovation teams are a trap — their wins don't prove anything to mainstream employees because the context is artificial.
• The first teams must be real, fully funded, mainstream line teams working on high-profile projects in production systems.
• When those teams tell their success stories, peers respond "that's my world" — and the social proof becomes credible.
• IBM's programme was named and branded: the Hallmark Program. Branding creates a neutral container that doesn't activate pre-existing biases around words like "AI" or "design thinking."

The one-way door

• Every team entering the Hallmark Program committed fully: no split attention, no "feet in both worlds."
• At the start of each team's week-long onboarding, Gilbert personally offered a no-questions opt-out: if the project hadn't advanced 10x by Friday, the team could leave and never hear from the programme again.
• Nobody ever left — but the offer itself was the mechanism that made participation feel like a genuine choice.
• Teams that stayed received visible senior leadership recognition, dedicated support, and the removal of bureaucratic roadblocks.
• High-profile projects were chosen deliberately: public failure would deliver fast, definitive feedback rather than slow ambiguity.

Quick wins done wrong

• The standard definition of a quick win — using the new method and getting a better project outcome — is incomplete.
• A good outcome on a non-representative team proves nothing about scalability to the rest of the organisation.
• A true quick win must demonstrate that the change works in the mainstream context, on real systems, with real funding, by real teams peers can identify with.
• Most programmes celebrate outcomes; IBM measured and communicated adoption patterns and cultural indicators alongside outcomes.

Adoption vs. competency

• The critical decision point in every change programme: optimise for sustained cultural adoption of the thing, not for immediate competency in the thing.
• Competency gains fade when people return to old environments. Cultural adoption is self-reinforcing.
• Allowing teams to revert after initial training is one of the most common structural failures Gilbert observes when coaching CEOs and chief transformation officers.
• Once a team was in the Hallmark Program, it only operated in the new way — permanently. There was no return path built into the system.

Cupcakes, birthday cakes, and wedding cakes

• When cutting scope to meet a delivery deadline, teams typically deliver 80% of every use case — a half-baked slice of wedding cake that serves no one completely.
• The cupcake principle: deliver 100% of a smaller problem rather than 80% of everything. A cupcake is holistic, delightful, and complete in itself.
• A wedding cake may require ten cupcakes released sequentially — each solving one use case end-to-end.
• The metaphor became embedded in IBM's product teams as a shared heuristic for scope decisions under time pressure.
• Key test: will this release completely overwhelm its target user with delight? If not, it is a slice, not a cupcake.

Branding the change

• Generic labels — "AI-first," "design thinking initiative" — activate pre-existing opinions and constrain the programme team's own thinking.
• A neutral brand name (Hallmark) freed the team to change systems, incentives, and environments that fell outside the apparent scope of a "design" programme.
• IBM's Hallmark Program was built on three explicit values layered beneath the tactical methods:
  1. The user is the north star.
  2. Restlessly reinvent — everything is always a prototype.
  3. New, diverse, empowered teams that move quickly.
• Values that no one wants to argue against provide a motivational foundation that pure method training cannot.
• Branding also enables consistent storytelling: teams speak about "the Hallmark Program" rather than a shifting set of practices.

The role of middle management

• The frozen middle — the assumption that middle managers are the primary resistors of change — is a widespread leadership bias, not an empirical finding.
• Gilbert admits he held this bias and built his early programme architecture around it, inadvertently marginalising middle managers.
• Once the team understood what was actually constraining middle managers (unclear expectations, dual accountability, no air cover), unlocking them became the single biggest scaling accelerant.
• Middle managers are the determining factor in whether any change sticks at the team level; getting them wrong defeats every other element of the programme.

Structural principles summary

• Cover every base at reduced scale: start small, but make the small version complete — a cupcake, not a slice.
• Sequence matters: nail team selection before scaling; nail the onboarding experience before broadening the brand.
• Systems around the change (performance management, incentives, physical environment) must reinforce the new behaviour — neglecting these is why change reverts.
• The programme team's job is product management: listen to why adoption fails, iterate on the offering, and go back to market.
• Success metrics must include cultural adoption signals, not just project-outcome wins.