The Big Bang Transformation Myth: Why Small Steps Win
Large-scale digital transformation projects in the public sector fail more often than they succeed. That’s not opinion — it’s documented. McKinsey data shows 70% of digital transformation projects fail to meet their goals. For large public-sector IT projects specifically, the numbers are worse: 81% overrun their schedules, with cost overruns averaging three times higher than private-sector equivalents. (https://www.runn.io/blog/it-project-management-statistics) The Standish Group found that government tech projects over $6 million succeed only 13% of the time. (https://www.belfercenter.org/publication/government-tech-projects-fail-default-it-doesnt-have-be-way) BCG surveyed large-scale tech programs across ten industries, including public sector: only 30% met timeline, budget, and scope expectations. (https://www.bcg.com/publications/2024/most-large-scale-tech-programs-fail-how-to-succeed) These are not outliers. This is the baseline.
Why large projects fail The pattern is consistent across studies. Large transformation projects fail because of:
Scope creep without checkpoints No shared understanding of current state before starting Ownership gaps — no one person responsible for outcomes Underinvestment in change management (only 10% of transformation budgets, on average) Unrealistic timelines set at the start, before constraints are understood Vendor relationships structured around deliverables, not outcomes
BCG’s research points to shortcomings in planning (unclear scope and business case), management (internal resources and external partners), and delivery (failing to identify risks and dependencies early). (https://www.bcg.com/publications/2024/most-large-scale-tech-programs-fail-how-to-succeed) The common thread: projects try to solve everything at once, with fixed scope and fixed timeline, before anyone fully understands the problem.
The alternative: phased delivery with checkpoints Projects that succeed look different. They start smaller, validate assumptions early, and build momentum through visible results. McKinsey found that projects using iterative principles improve delivery speed by 20–30% on average. (https://www.proprofsproject.com/blog/project-management-statistics/) BCG recommends an incremental implementation roadmap — following a minimum viable product mindset — where everyone involved understands the current project status at any given time through predefined reporting. (https://www.bcg.com/publications/2024/most-large-scale-tech-programs-fail-how-to-succeed) What does this look like in practice?
Baseline first. Before defining scope, document what exists: services, workflows, responsibilities, constraints. This prevents planning against assumptions. Phased delivery with named ownership. Each phase has a defined outcome, a responsible owner, and a checkpoint before the next phase begins. Quick wins early. Small, visible improvements build stakeholder confidence and internal capability. Staff learn the new ways of working before the stakes get high. Documentation and handover built in. Every milestone produces something that can be maintained independently — not a dependency on the original team. Measure against baseline. If an improvement doesn’t outperform the starting point, it doesn’t ship.
This approach doesn’t guarantee success. But it reduces the risk of catastrophic failure — the kind where millions are spent and nothing works.
Why institutions still choose big bang If phased delivery works better, why do large projects keep getting funded? A few reasons:
Budget cycles reward big asks. It’s easier to secure one large budget than to return for incremental funding. Political timelines. Leaders want visible results before elections or term limits. A three-year roadmap doesn’t fit a two-year mandate. Vendor incentives. Large fixed-scope contracts are more profitable than phased engagements with ongoing accountability. Optimism bias. Every project team believes they’ll be the exception.
These pressures are real. But they don’t change the data. Large projects fail at predictable rates, for predictable reasons.
What this means for your next project If you’re planning a digital transformation — or recovering from one that stalled — consider:
Start with a readiness snapshot, not a requirements document. Understand the current state before defining the future state. Define phases with outcomes, not just deliverables. “New website launched” is not an outcome. “Citizens can complete permit applications without visiting the office” is. Assign ownership by name. If no one is accountable for a result, the result won’t happen. Build in checkpoints. Each phase should produce something usable — and a decision point on whether to continue, adjust, or stop. Measure against baseline. Track what changes, not just what ships.
The 70% failure rate is not inevitable. It’s the result of a specific project structure — one that can be changed.
DIGIPART helps public institutions plan and deliver digital transformation in phases, with clear ownership, documentation, and measurable outcomes. If you’re facing a stalled project or planning a new initiative, we can help you assess where you stand and define a realistic path forward.
Sources:
McKinsey / Runn: https://www.runn.io/blog/it-project-management-statistics Standish Group / Belfer Center: https://www.belfercenter.org/publication/government-tech-projects-fail-default-it-doesnt-have-be-way BCG: https://www.bcg.com/publications/2024/most-large-scale-tech-programs-fail-how-to-succeed ProProfs / McKinsey: https://www.proprofsproject.com/blog/project-management-statistics/
