Architecture Debt Is a Cost Model, Not a Backlog
Most organisations track technical debt as a list of deferred tasks. Treating it as a compounding cost model changes which decisions get funded.
Most organisations track architecture debt as a backlog: a list of deferred refactors, each with a rough estimate, each perpetually outranked by feature work. The framing guarantees the outcome. A backlog item competes for capacity. A cost model competes for budget — and budget conversations happen at a level where architecture decisions actually get made.
What the backlog framing hides
A deferred refactor has a carrying cost that compounds across three dimensions at once:
- Infrastructure — over-provisioned services that were sized for an architecture that no longer exists.
- Delivery — every feature routed through the compromised boundary pays a coordination tax.
- Risk — the compliance and resilience surface widens quietly.
None of these appear on a ticket. All of them appear on a P&L.
Reframing the conversation
When I work with leadership teams on modernisation, the first artefact is rarely a target architecture. It is a model that attaches a monthly number to the current one. Once the carrying cost is visible, sequencing stops being an engineering argument and becomes an economic one.
That shift is what makes modernisation fundable — and what keeps it funded after the first quarter of work stops producing visible features.