Lead Time to Value
The elapsed time from identifying an opportunity to delivering measurable benefit, which is usually dominated by waiting rather than by building.
Time to market is discussed as an engineering property and is mostly not one. Measure the whole path and the picture changes: idea to prioritisation, prioritisation to start, start to code complete, code complete to production, production to measured benefit.
In most organisations the build portion is a small fraction. The rest is queueing — prioritisation cycles, dependency waits, approval gates, environment contention, release windows, and the delay before anyone measures whether it worked.
Architectural properties that genuinely move it: independent deployability, so a release is not gated by another team; a fast deployment pipeline, since anything slower than about ten minutes changes behaviour and causes batching; loose coupling, so a change touches one service rather than five; feature flags, which decouple deployment from release and remove the release-window queue; and a platform with a paved road, which removes the setup work from every new initiative.
The trade to be explicit about: speed can be bought with debt, and doing so deliberately for a genuine market window is legitimate — provided it is recorded with a repayment trigger. Doing so habitually produces an architecture where nothing is fast any more, which is how organisations arrive at a modernisation programme.