Build, Buy, or Broker: Rethinking the Modernization Decision Before It Makes Itself
Photo: executive boardroom decision meeting modern office, via images.stockcake.com
The Question Every Technology Leader Answers by Default
Somewhere between the quarterly board review and the next security audit, most US enterprises answer the build-versus-buy question without ever formally asking it. A legacy order-management system gets wrapped in APIs rather than replaced. A SaaS contract renews for the sixth consecutive year. An internal platform team balloons to forty engineers building capabilities no competitor would ever pay for. These are not decisions; they are drift.
The stakes have never been higher. Enterprise software spending in the United States now routinely consumes 4–6% of revenue at mid-cap companies, and the margin between a well-chosen modernization path and a poorly chosen one compounds annually. What follows is not a verdict on which path wins — it is a framework for choosing deliberately.
Three Paths, Three Philosophies
Build means owning the codebase, the roadmap, and the failure modes. It maximizes differentiation and control but demands scarce engineering talent, multi-year capital commitments, and executive patience. When Capital One announced its exit from the public cloud data center model and committed to building proprietary banking infrastructure, it accepted a decade-long obligation in exchange for capabilities competitors cannot license.
Buy means renting speed. A SaaS procurement can compress an eighteen-month internal build into a ninety-day rollout. The trade is sovereignty: your workflows bend toward the vendor's data model, your pricing renews under their terms, and your roadmap quietly becomes their roadmap.
Broker — the path least discussed and most practiced — means orchestrating a hybrid stack: commercial platforms for commodity functions, custom code where differentiation lives, and integration middleware holding it together. Nearly every large US enterprise already brokers whether it admits to or not. The question is whether the orchestration is intentional or accidental.
The Cost Nobody Budgets For
Sticker price is the least honest number in any of these conversations. Build projects routinely exceed estimates by 50–100%, and the true cost includes the opportunity cost of engineering hours diverted from product innovation. Buy decisions hide their expenses in integration labor, per-seat escalation at renewal, and the switching cost that accumulates with every year of vendor lock-in. Broker arrangements carry the stealthiest cost of all: the integration tax, paid in brittle point-to-point connections and the salaries of the few engineers who understand how it all hangs together.
A useful discipline is to model total cost of ownership over seven years — long enough to capture at least one migration, one vendor renegotiation, and one major platform re-architecture.
A Practical Scoring Rubric
Score each candidate capability from 1 (low) to 5 (high) across five dimensions:
- Differentiation. Does this capability make customers choose you? Score high, lean build.
- Velocity pressure. How quickly must this ship? Score high, lean buy.
- Talent availability. Can you realistically hire and retain the engineers required? Score low, lean buy or broker.
- Regulatory sensitivity. Does data residency, auditability, or compliance make vendor dependence risky? Score high, lean build.
- Ecosystem maturity. Are there credible, financially stable vendors serving this need? Score high, lean buy.
A capability scoring high on differentiation and regulatory sensitivity but low on talent availability is a candidate for a brokered approach: buy the commodity layer, build the differentiating edge, and invest heavily in the integration seam. A capability scoring low on differentiation across the board should almost never be built — no matter how much the internal team enjoys building it.
The Talent Ledger
The rubric's most frequently ignored line is talent. The US engineering market concentrates senior platform talent in coastal hubs and well-capitalized technology firms; a manufacturer in the Midwest competing for the same Kubernetes specialists as a FAANG-adjacent startup will lose that auction. Brokered strategies acknowledge this reality: they reserve scarce senior builders for the two or three capabilities where engineering excellence translates directly into revenue, and rent commodity competence everywhere else.
Decide on Purpose
The CTOs who modernize successfully are not the ones who picked the right path by luck. They are the ones who made the choice explicitly, documented the reasoning, scored the trade-offs, and revisited the decision when the assumptions changed. Build, buy, or broker is not a one-time referendum — it is a portfolio discipline. Treat it like one, and the drift becomes direction.
