Canada Doesn't Have a Consultant Problem. It Has a Classification Problem.

When a major government initiative fails, the post-mortem almost always converges on the same explanations: the vendor underdelivered, the procurement was mismanaged, the political oversight was inadequate. These explanations are not wrong. They are just downstream of a more fundamental error that rarely gets named.

Canada's federal government grew its public service to approximately 448,000 full-time equivalents between 2015 and 2025 — an 80 percent increase in the annual cost of the bureaucracy, from $39.6 billion to $71.4 billion. Under ordinary administrative logic, that expansion of internal capacity should have reduced reliance on external advisors. Instead, federal spending on professional and special services rose from $9.5 billion in 2015 to a projected $23.1 billion annually by the mid-2020s, with management consulting alone accounting for approximately $5 billion in 2024–2025.

The standard political explanation is corruption, favoritism, or incompetence. The structural explanation is more precise, and more useful: Canada is systematically reaching for the wrong tool.

Two types of problems. One default response.

The Cynefin Framework, developed by organizational theorist Dave Snowden and formalized in a landmark 2007 Harvard Business Review paper, draws a critical distinction between two classes of problems that are routinely conflated in public sector procurement.

Complicated problems are technically dense but ultimately knowable. The cause-and-effect relationships exist and can be uncovered through rigorous expert analysis. A bridge has optimal engineering specifications. A tax code has optimal compliance architecture. In these environments, hiring subject-matter experts to analyze the system, identify the best options, and implement a recommendation is exactly the right approach. Traditional management consulting was designed for this domain, and it delivers genuine value here.

Complex problems are categorically different. Cause and effect cannot be determined in advance. The system involves multiple autonomous actors with divergent incentives, deeply entrenched behavioral patterns, and no stable best practice. The solution cannot be found on a desktop; it can only emerge through the structured convening of the stakeholders who inhabit the system, iterative probing, and continuous adaptation. In this domain, no consulting firm — regardless of prestige or billing rate — can hand over a functional answer, because the answer does not yet exist in a form that analysis can retrieve.

The critical failure in Canadian public procurement is not that the government hires bad consultants. It is that the government consistently hires consultants — a Complicated-domain tool — and deploys them against Complex-domain problems. The failure is baked into the choice of intervention before the contract is signed.

The pattern across four cases

Phoenix Pay System

The government treated federal payroll modernization as a sophisticated software engineering problem and contracted IBM to build and implement a solution. Federal payroll is not a software problem. It encompasses over 80,000 unique pay rules, dozens of distinct collective agreements, and the deeply variable human behaviors of 300,000 employees across more than 100 departments. IBM's contract grew from $309 million to $545 million across 50 amendments. When the system collapsed on launch, the government contracted McKinsey to provide process optimization services — a second Complicated-domain intervention on the same unresolved Complex problem. That contract grew from $4.9 million to $27.7 million. By late 2025, the backlog still exceeded 233,000 unprocessed transactions. The Auditor General's assessment identified the root failure as the government's inability to simplify rules and achieve consensus before implementing technology. That is not a technical finding. It is a stakeholder alignment finding.

ArriveCAN

The pandemic border management app was estimated at $80,000 and ultimately cost between $54 million and $60 million. GC Strategies, a two-person intermediary firm, collected between $11 million and $19 million without building any part of the application — its value was navigating the government's own procurement labyrinth. The Auditor General found 18 percent of tested invoices lacked sufficient documentation to confirm whether expenses were even related to ArriveCAN. The government, lacking internal capacity to navigate a rapidly evolving, multi-jurisdictional crisis, outsourced the complexity to a firm whose primary skill was procurement navigation. The complexity did not go with it.

PrescribeIT

Canada Health Infoway invested $298 million in a digital prescription platform. The technology worked. Fewer than 5 percent of prescriptions ever flowed through it. The program was shut down in May 2026. The failure was not in the pipes — it was in the behavioral coordination challenge that no consulting mandate was designed to address.

McKinsey and IRCC

Federal spending on McKinsey grew from approximately $2.2 million across nine years of the Harper government to over $117 million under the Trudeau administration. A 2024 Procurement Ombudsman review found that 25 of 32 reviewed contracts were sole-sourced, producing what the report described as "a strong perception of favoritism." Immigration processing — entangled with shifting geopolitical realities, legislative constraints, and intense domestic political pressures — is a structurally Complex challenge. The IRCC contracted McKinsey for $24.8 million in "transformation strategies." Process optimization cannot resolve structural legislative and geopolitical constraints.

Why the default persists

Three structural forces sustain the pattern. First, hiring a globally recognized firm provides political and bureaucratic cover. If a $300 million initiative fails, decision-makers can point to the engagement of IBM or McKinsey as evidence of due diligence. The optics of expert engagement substitute for actual problem diagnosis.

Second, federal procurement rules demand premature certainty. An RFP requires a defined scope, specific deliverables, and a fixed timeline — epistemologically impossible for a Complex problem whose solution must emerge through interaction. Public servants are structurally compelled to flatten Complex challenges into Complicated technical specifications to satisfy procurement requirements. The real-world complexity then asserts itself through endless, expensive amendments.

Third, continuous outsourcing erodes the internal diagnostic capacity required to classify problems correctly. As external consultants absorb more of the high-level analytical work, the public service loses the institutional muscle to frame problems accurately — and becomes increasingly dependent on firms whose primary expertise is navigating the contracting processes that the government itself created.

The Auditor General's reports, parliamentary committee findings, and ombudsman reviews consistently interrogate the execution of contracts. They ask whether the money was spent correctly. The prior question — whether the money was spent on the right thing — goes largely unexamined.