① Guide · Budget and forecasting
A flat run rate hides every risk that matters. Mainframe agreements run multi year, consumption creeps, and the big uplifts land in years you are not looking at. A three year model puts renewals, creep, and uplift risk in one forecast, so finance is never surprised and the negotiation starts while you still have leverage. Here is how to build it.
A one year budget cannot see the renewal that ends year two.
Most mainframe software budgets are a run rate. Take last year's spend, add a percentage for inflation, and submit it. The problem is that the mainframe does not cost money in a straight line. Agreements run on three and five year terms, so the expensive events, the renewals, arrive in bursts that a single year view never anticipates. Consumption creeps a few percent a year and compounds against sub-capacity charges and Tailored Fit baselines. And the vendor's opening uplift on a major renewal can move the number by a margin no inflation assumption covers. A flat budget meets all of this as a surprise.
A three year model replaces the run rate with a forecast that shows where the money moves and when. It places every renewal on its term expiry, projects the consumption that drives the metered charges, and carries a provision for the uplift risk you cannot yet size. The payoff is twofold: finance funds the negotiation lead time instead of absorbing a shock, and the team sees every renewal far enough ahead to prepare it. This guide builds the model in four layers. For the disciplines behind it see cost optimization and renewal advisory.
Current contracted cost by vendor and product, the floor you are obligated to. Build it from the contracts, not the invoices, so the number reflects what you committed to rather than what happened to be billed. This is the only layer you know with certainty, and everything else builds on it.
Every agreement placed on its term expiry across the three years. The renewals that fall inside the horizon are the events the budget exists to anticipate. Flag each one with its lead time so the model shows not just the cost but the date preparation has to start.
Projected MIPS and MSU growth across the period. Creep raises sub-capacity charges directly and inflates any Tailored Fit baseline set during the window. Model it as a managed range, not a single line, so the budget reflects both an unmanaged trajectory and the lower one disciplined optimization can hold.
A held provision for the uplift on renewals you cannot yet size. Set it to the vendor's observed pattern as a range, not a point. A prepared negotiation that lands below the provision releases money back to finance as a visible win; an unprepared one that exceeds it is exactly the surprise the model exists to prevent.
| Layer | Year 1 | Year 2 | Year 3 |
|---|---|---|---|
| Committed base | Known floor | Known floor less expiring terms | Reduced as agreements roll off |
| Renewal layer | Minor renewals | Major renewal lands; prep started year 1 | Mid term, prep next cycle |
| Consumption layer | Managed range | Range widens; clean before year 2 renewal | Baseline locked from cleaned profile |
| Risk provision | Low | Sized to the year 2 vendor pattern | Released or carried per outcome |
The shape above is illustrative. Actual figures depend on your contract mix, term dates, and consumption. The point is the structure: see the events early, fund the lead time, budget the range.
Mainframe agreements run on multi year terms, so a one year budget sees only a slice of the risk. A three year horizon catches the renewals landing in years two and three while there is still time to prepare them, exposes the creep that compounds across the period, and lets finance fund the negotiation lead time rather than absorb a surprise.
Four layers: a committed base of current contract costs, a renewal layer placing each agreement on its term expiry, a consumption layer projecting MIPS and MSU growth, and a risk layer holding a provision for uplift on renewals you cannot yet size. Together they turn a flat run rate into a forecast that shows where the money moves and when.
Carry a risk provision sized to the vendor's observed pattern as a range, not a point number. Budgeting the range rather than the best case means a hard renewal does not blow the plan, and a well negotiated one releases provision back to finance as a visible win.
Audit notice or renewal under 18 months out? We mobilize within 48 hours. Building the multi year budget now? We model the renewals and uplift risk before they reach finance.
The renewal you can see three years out is the one you can still shape.
Related guides: capacity planning with software cost, benchmarking your mainframe software spend, and the CFO guide to mainframe software spend. Explainer: the MIPS to MSU conversion question. Commercial: cost optimization.