The Buy Mandate Just Caught Up with the Build-It-Yourself Pitch

Some agencies are being pitched a custom, homegrown entity resolution build right now, positioned as more control and a system built “for us.” It’s an appealing pitch. It’s also arriving at the exact moment federal procurement policy has moved hard in the opposite direction, and at the exact moment the government’s own data shows what building identity infrastructure in-house actually costs.

The Ground Has Shifted Since April 2025

Federal agencies have had a commercial preference on the books since 1994, under the Federal Acquisition Streamlining Act. For most of three decades, it was more aspiration than enforcement. That changed starting in April 2025.

Executive Order 14271, Ensuring Commercial, Cost-Effective Solutions in Federal Contracts (signed April 15, 2025), directed agencies to justify any non-commercial procurement in writing. The order itself states that past administrations “evaded statutory preferences and abused the Federal contracting framework” by procuring custom solutions where a commercial one would have worked.¹ Executive Order 14275, Restoring Common Sense to Federal Procurement, signed the same day, is driving a full rewrite of the Federal Acquisition Regulation, with the FAR Council’s proposed changes out for public comment through July 23, 2026.²

The enforcement mechanism landed on April 17, 2026, in OMB Memorandum M-26-12, Increasing the Acquisition of Commercial Products and Services, signed by OMB Director Russell T. Vought. It requires every agency to report all non-commercial contract awards, and for anything over $10 million, the agency’s senior procurement executive must provide market research and a written explanation for why a commercial solution wasn’t used at the next contract option.³ Separately, if an agency wants OMB’s input before soliciting a non-commercial acquisition, a step the order makes available but doesn’t require, the senior procurement executive first needs concurrence from the agency’s Chief Acquisition Officer or an equivalent senior non-career official before that request even reaches OMB.³

OMB’s own justification for the crackdown: in fiscal year 2024, more than two-thirds of total federal contract spending reported through the Federal Procurement Data System was for non-commercial products and services, including over $130 billion for common services like professional support and IT, much of it under cost-reimbursement contracts.³

The Mandate Exists Because the Track Record Is Bad

This isn’t a policy preference invented in a vacuum. GAO has kept federal IT acquisition and management on its high-risk list since 2015. Its most recent assessment found 463 of 1,881 IT-related recommendations to agencies and OMB, issued since 2010, still unimplemented. GAO’s own words: “federal IT investments too frequently fail or incur cost overruns and schedule slippages” while delivering little on their intended mission outcomes.⁴ Entity resolution is a specific, narrow case of a pattern GAO has documented at the portfolio level for over a decade. It’s a mistake to assume any given agency’s homegrown build will be the exception. It’s also worth noting why: entity resolution is one of the most complex algorithmic challenges in data management, data quality, prediction systems, and AI. Building industrial-strength, commercial-grade entity resolution technology has historically required investments in the tens of millions of dollars, which is exactly what the next section lays out.

What Building It Actually Costs, and What Buying It Has Delivered

Separate from the policy question, there’s the plain economics. Based on more than 350 person-years of engineering experience building entity resolution technology, an organization attempting to build in house should expect to spend roughly $1 million to reach about 70 percent of the feature set found in a mature commercial solution. Reaching 80 percent moves that estimate to around $5 million. Reaching 90 percent, the estimate rises to roughly $30 million, and even then, the result isn’t considered competitive with a purpose-built commercial engine. Pushing further, from roughly 92 percent to 95 percent, can require a ground-up rewrite rather than incremental investment.

The Control Need Answered Directly

The real appeal of a homegrown pitch to a government buyer usually isn’t cost. It’s control: data sovereignty, no vendor lock-in, an audit trail the agency owns. Those are legitimate concerns, and the answer isn’t to dismiss them. A commercial engine deployed on premises or in an agency’s own cloud, with no data ever flowing back to the vendor, delivers the same sovereignty without the multi-year build risk. Any mature commercial entity resolution engine, ours included, gives every resolution decision a full audit trail showing why records matched and what drove the decision. That is the explainability government compliance and oversight functions actually need, and it is a property of buying a mature commercial solution generally, not something unique to any one vendor. That directly answers the reason a homegrown pitch sounds appealing in the first place.

Sources