Cost·Sep 10, 2026·4 min read

Cost and complexity takeout

Most legacy platforms cost more to keep alive than to run. The savings are real, but they hide in dead code, duplicated services and infrastructure sized for a peak that stopped happening years ago.

MKMichael Kofman · Founder, ARC

Where the money goes

Run costs are visible: hosting, licences, support contracts. Keep-alive costs are not. They are the engineers who can't be moved because only they understand a module, the change that takes six weeks because nobody knows what it will break, and the licence renewed because no one can prove the feature isn't used.

What ARC finds

  • Dead and duplicated code. Screens no journey reaches, services no one calls, and three implementations of the same calculation.
  • Change hot spots. Files that change together in git history, revealing the coupling that makes every release slow.
  • Maintainability, measured. Each component scored, with the debt that would pay back fastest called out first.
  • Infrastructure against usage. What the system actually does at peak, so the target is sized for reality.

Then a plan with a number on it

Discovery ends with a fixed-price plan for the rest. You decide what to remove, rebuild or leave alone with the cost of each choice in front of you, not after a two-year program has started.

You can't cut what you can't see. Discovery makes it visible.

What you get

Inventory of wasteDead code, duplicated services and unused screens, each cited.
Debt rankingComponents ordered by maintainability and payback.
Target sizingA target architecture sized against real usage.

Timelines depend on access to code and people; we confirm them on the scoping call.

Under pressure to cut run costs?

Bring last year's run cost and the system you suspect most. We'll scope a discovery that shows where it goes.