A sound architecture decision is not correct forever. It is correct under a set of assumptions: expected volume, risk profile, available skills, budget, and pace of change. When those assumptions stop being true, preserving the decision for consistency can be riskier than revisiting it.
Yet many organizations document what they decided and omit how long the decision should remain valid. The consequences are familiar. A shortcut chosen to meet a deadline becomes a permanent platform. A temporary integration ends up supporting critical processes. A constraint from five years ago continues to shape current investment.
My position is straightforward: every important architecture decision should have a review date.
Review does not mean replacement
Adding a date does not mean changing technology every six months. It means examining the assumptions again before reality invalidates them at a high cost.
A review may conclude that the decision remains appropriate. In fact, that should happen often. The difference is that the organization confirms it with current information instead of preserving it through inertia.
For this to work, a decision record needs more than the selected option and the alternatives that were rejected. I would include four elements:
- Assumptions: what must remain true for the decision to make sense.
- Review signals: which change in volume, cost, regulation, or risk should trigger an earlier assessment.
- Owner: who watches those signals and initiates the review.
- Date: when the decision will be reviewed even if no signal has fired.
The date is not a commitment to change. It is a mechanism against organizational memory loss.
The hidden cost of permanent decisions
Platforms rarely become difficult to modernize because of one poor choice. The issue emerges as decisions accumulate without being reconsidered. Each one adds a dependency, an exception, or a specialized skill that the next initiative must accept.
Over time, the team stops discussing the architecture the business needs and starts discussing what the inherited architecture allows. Technology is no longer enabling strategy; it is constraining it.
Transparency also suffers. When the original assumptions are unclear, a proposal to change can look arbitrary, while defending the current state can look prudent. An explicit validity period lets leaders compare both options using observable criteria: operational risk, total cost, team capability, and time to value.
Lighter and more useful governance
This does not require another committee. Reviews can become part of planning, risk monitoring, or the preparation of major initiatives. The priority is to focus on decisions with long consequences: shared data, identity models, integration contracts, cloud platforms, payment rails, and boundaries between business domains.
I recommend starting with the ten decisions that constrain the business most today. For each one, reconstruct the rationale, make the current assumptions explicit, and set the next review date. If the team cannot explain why a decision is still valid, it has already found the first one that deserves attention.
Sustainable architecture is not architecture that never changes. It is architecture that can distinguish between a foundation that still protects the business and a constraint that has outlived its purpose. A review date makes that distinction visible before an urgent modernization program exposes it in the worst possible way.
I help organizations reassess technology decisions and turn them into a defensible modernization roadmap. View services.
