Why most digital transformations stall — and why that is the wrong question
Too many leaders still treat transformation as a programme: a three-year plan of projects, PowerPoints and KPI dashboards. That’s why we keep seeing the same headline — most transformations fail. A useful read on the economics of these efforts is the Harvard Business Review piece on the value (and under-delivery) of digital transformation, which found companies capture far less of the expected gain than they plan for. HBR: The Value of Digital Transformation.
Reframe transformation as product discovery — not project delivery
When you treat change as a temporary programme you leave the business with a portfolio of completed projects but no persistent capability. Product thinking flips that: you build continuous value streams that users pay for, and you staff them with long-lived, multidisciplinary teams whose job is to discover outcomes not just deliver outputs.
- Outcomes over outputs: shift goals from “deliver feature X” to “improve metric Y for this customer segment”.
- Long-lived teams: keep people together to learn, iterate and own end-to-end results.
- Experimentation as default: instrument everything and make hypotheses cheap to validate.
Example: ING’s agile reorganisation — structure to learn
ING’s move to squads and tribes is often cited for a reason: it was less about Agile ceremonies and more about creating teams that could iterate with autonomy and focus on customer value. The McKinsey write-up of ING’s agile transformation highlights how structure + leadership attention enabled faster learning cycles. McKinsey: ING’s agile transformation.
For product leaders the lesson is clear: organisation design matters. You can copy ceremonies, but without stable ownership and a measurable customer outcome, teams will default back to local optimisation and feature-factory behaviour.
Three practical moves every executive can make this quarter
1. Convert top initiatives into product rivers
Stop launching six-month projects and start funding continuous product teams. Re-badge the work as a product, give it a north star metric, and provide a small, sustained budget for experiments. That lets teams learn and pivot — the exact thing programmes refuse to do.
2. Measure learning as a first-class KPI
Traditional KPIs reward shipping. Instead, measure validated learning: experiments run, hypotheses proved/disproved, customer outcomes improved. Use lightweight guardrails — safety, legal and privacy — but make learning the currency of progress.
3. Protect new bets from the rot of the core business
Large organisations are terrible at protecting nascent products. Create incubation boundaries: separate deployment pipelines, different success metrics, and a limited-life sponsorship from an executive sponsor who can shield teams from BAU demands. This isn’t theatre — it’s structural protection for the discovery cycle.
AI, talent gaps and governance — the new constraints for product leaders
The latest wave of AI has raised expectations and risk simultaneously. Leaders are tempted to “AI-enable” existing projects, but without product-led ways of working this becomes an expensive cosmetic. McKinsey’s guidance on combining leadership and digital transformation suggests that capability and governance must keep pace with ambition. McKinsey: Leadership and digital transformation.
Product organisations must therefore pair AI investments with three things: accessible training for staff, clear ethical guardrails, and a fast experiment pipeline. That reduces the pandemic of stalled AI pilots and turns experiments into learned improvements for users.
Governance that empowers, not micromanages
Effective governance looks like funded outcomes, not approvals. Give product teams a budget envelope, a clear success metric, and a fortnightly review where the conversation is about evidence, not excuses. This shifts governance from a permission machine into a learning accelerator.
Where to start next week
- Pick one stalled programme and convert it into a product with a north-star metric.
- Form a product trio (PM, Tech Lead, Design Lead) and seed a two-week discovery sprint focused on measuring one customer behaviour.
- Run a governance review: replace one monthly approval meeting with a fortnightly evidence review.
These are low-friction, high-impact changes that push organisations away from the project mentality that creates short-lived wins and long-term technical debt.
Final thought
Transformation will never be a one-time programme. The organisations that win will be the ones that treat digital as a product practice: small autonomous teams, relentless measurement, and governance designed to accelerate learning. If your board still asks for a three-year rollout plan with milestone Gantt charts, you have a cultural mismatch to address. Start with one product river, show the learning, and let the results do the persuading. The work is less about technology and more about creating a system that continually discovers what customers value — that’s where real transformation lives.
Leave a Reply