Modernization governance / May 2026
The modernization case finance will actually fund
A modernization proposal built on a lower run cost competes against growth projects and loses to them. The case that gets funded prices what standing still will cost, shows measurable value inside the first two quarters, and names one business owner accountable for the benefit. Governance is what keeps those numbers honest after the approval.
Why Good Programs Die
The savings story is honest and weak.
Nearly every large organization has already agreed, in principle, to replace a system it depends on. Very few can produce a case that survives an hour with the CFO. The engineering argument is usually fine. The money argument is built the wrong way round, asking for certain spend today against a cheaper operating bill two or three years out.
Finance discounts that shape heavily, and correctly. Spend is front-loaded and certain. Savings are back-loaded and conditional. A benefit arriving a year late can erase the return even when every promised line eventually comes true.
A second problem follows. A savings case competes in the wrong category. On the capital plan it appears beside a new product line, a market entry, a pricing capability, each promising to earn money. A proposal framed as spending less will lose that comparison most of the time.
The Baseline
Draw the do-nothing line as a rising curve, then defend the slope.
The defect we meet most often is a flat baseline. The proposal compares modernizing against today's cost of the existing system, held constant for five years. That comparison is fiction and a capable finance team will say so in the first ten minutes.
Three forces bend the line upward, and none are under your control. When mainstream maintenance ends, staying patched becomes a recurring extended-support fee that rises each year you delay. Microsoft publishes exactly that structure for its Extended Security Updates program. The talent pool shrinks each year, and the engineers who remain price their scarcity. Compliance exposure builds with no invoice attached until an audit finding or an incident turns it into one large number on a date nobody chose.
In 2019 the US Government Accountability Office examined ten legacy systems it identified as critical to federal operations. Their ages ranged from 8 to 51 years, and operating them cost roughly $337 million annually in total. The figure is not the lesson. The curve is. Legacy cost climbs, and a baseline worth defending is drawn the same way.
The comparison that matters is therefore never against today. It is against the climbing line. The widening gap between the two is much of what the investment buys, and a savings-only deck leaves it on the table.
The Model
Four streams of value, and three of them are usually missing.
A defensible case accounts for four separate kinds of value. Most proposals quantify one well and gesture at the rest.
- Run cost reduction. The licence, infrastructure and support savings every deck opens with. Usually the smallest of the four, the slowest to arrive, and the simplest for finance to write down. Open with it and the proposal reads as overhead.
- Risk cost avoided. Escalating support fees, the exposure created when the last person who understands a batch process retires, and the liability that grows with each unpatched year. This stream is the climbing baseline with a price attached. In our experience it produces the biggest figure in the model, and it is the figure most often left out.
- Velocity and revenue enabled. The things the business will be able to do afterwards that the current platform made slow or impossible. A capability delivered in weeks instead of quarters. A channel that could finally open. A pricing model the existing system cannot represent. This stream is what allows a modernization proposal to argue on the same ground as a growth proposal.
- Optionality. The worth of owning a platform that can be changed. A current estate makes it possible to add a capability, absorb an acquisition, or leave a supplier without a second migration. You are buying the right to move later, not the obligation. Say so out loud, because a CFO who prices risk professionally understands optionality better than most engineering teams.
Put risk avoided and velocity in front of the committee. Skipping them is why technically sound programs go unfunded year after year.
Cash-Flow Shape
Report when the value arrives, not only how much of it there is.
A case containing all four streams can still fail on timing. Spend arrives first, so the return is sensitive to when the benefit shows up. A program promising everything at a single cutover two years out concentrates its risk in one date, and finance knows how those dates behave.
The fix is structural. Sequence the work so measurable value appears early and keeps appearing. Routing traffic at a seam and retiring old paths incrementally can return value within four to six months. A substantial estate still takes eighteen to thirty months end to end, but when the cash arrives matters as much as how much of it there is. Each early proof point retires a risk the CFO would otherwise price in.
It is also the straight answer to the discount-rate objection. Nobody is asked to trust a benefit that arrives three years from now. You produce one inside two quarters and compound from there.
A specialty retailer had been refused funding twice for the same platform replacement, both times on a run-cost argument. We rebuilt the case around a rising support and compliance baseline, priced the velocity stream against two product commitments the existing system could not meet, and staged the program so the first measurable result arrived in month five. The board funded it, and capital spending fell 24 percent.
Governance
Three failure modes, and the governance that prevents each one.
The first is the padded number. Soft productivity gains and optimistic headcount assumptions get piled up until the payback period looks short. A finance team writes that total down to nothing as soon as one assumption reads as aggressive, and the defensible parts go down beside it. A small set of numbers that survive checking is worth more than a large total that does not. Keep the assumption register as a standing artifact the steering group reviews.
The second is claiming a benefit and never instrumenting it. If the case rests on faster delivery or fewer incidents, capture those measures on the existing system before any new code ships. A benefit with no baseline is an assertion no auditor can test.
The third is scoping the work as an engineering project. With no named business owner to retire capabilities, change the process around them, and protect the decommissioning line, a program tends to rebuild the old system faithfully, unused corners included. Organizational failure modes dominate technical ones. Name a person, not a committee, and give that person authority to remove whatever the replacement renders unnecessary.
What To Do Next
Build it in six steps.
- 1. Draw the do-nothing baseline as a curve that climbs, and defend its slope using dated vendor end-of-life schedules, evidence from the talent market, and specific compliance exposures.
- 2. Model all four value streams, and put risk avoided and velocity at the front of the paper.
- 3. Instrument the current system before a single line of migration work starts, so every claimed benefit has an earlier number to be measured against.
- 4. Sequence so the first measurable value appears inside six months, and report cash-flow shape alongside total return.
- 5. Name one business owner accountable for the benefit, the process change, and the decommissioning line.
- 6. Re-forecast the four streams quarterly, and write the conditions under which the program stops.
The programs that win funding are rarely the ones with the tidiest architecture diagrams. They are the ones whose case treats standing still as the expensive choice it quietly became, and proves that with figures finance can test.
Related Service
Modernization and Transformation Governance
We build the case and set the governance that keeps it honest afterwards.
Start A Conversation