📋 PLAYBOOK
Reading the dial does not stop the leak.
📍 IN BRIEF
By the end of this you will be able to run cost discipline predictably between renewals, built on a register of what you actually own, a finance-agreed number for what it must earn, and a rhythm that keeps the two honest. The aim is not to cut harder once a year, it is to make the bill an output of decisions you made deliberately rather than a surprise you explain afterwards.
The platform bill is a lagging indicator. By the time it lands, the decisions that set it are months old and fully priced in, the licences assigned in March, the customisation approved in May, the module bought for a programme that quietly stopped. Renewal is where cost becomes visible, not where it becomes controllable, and reading the dial more carefully has never yet fixed a leak. A sustainable return comes from running cost as a discipline between renewals, and that discipline is what this playbook hands you.
When to use this
This is for platform owners and governance leads whose spend is rising faster than the value anyone can prove, and who are tired of defending the platform with anecdotes at budget time. It assumes total platform cost already has a single named owner. If licensing sits with procurement, additional storage with infrastructure, support with operations and nobody adds them up, appoint that owner first. A cost discipline without an owner of the whole number is a spreadsheet with ambitions.
Stage 1: Build a register of what you actually own
Pull every entitlement into one register, including the ones you did not buy.
Start with the order forms, then go and find the entitlements that never crossed your desk. In some organisations of any size, other teams may hold active licences the platform team has never got into, bought for a project, inherited through an acquisition, or signed regionally and forgotten. Map everything the organisation is licensed for against what is actually deployed, because the register does two jobs at once. It shows you where you are paying for capability nobody switched on, and it stops a business unit buying something the organisation already owns.
The gotcha is the second bill. Optimising the spend you can see while an unwatched entitlement estate grows elsewhere produces a saving on paper and an increase in total. The sweep through procurement is not a formality, it is the difference between optimising the platform and optimising your corner of it.
Stage 2: Give every capability a number it must earn
Agree with finance what each major capability costs in full and what it is expected to return.
For each significant capability, set a baseline the finance team has signed, covering the full cost of ownership, the expected return, and the time value should take to arrive. Define the outcomes in financial terms wherever they can be, cost removed, capacity released, systems retired. This is slower than asserting benefits in a slide, which is exactly why it works. A baseline agreed with finance before the spend is the only evidence that survives scrutiny after it.
The gotcha is the business case as funding theatre. Most platforms have one, written once, polished to win approval, never opened again. An unrevisited case cannot defend anything, because every claimed saving becomes an assertion against a number nobody remembers agreeing. If the case does not get re-opened when scope changes, you do not have a baseline, you have a brochure.
Stage 3: Reconcile what you pay for with what is used
Close the gap between entitlement and usage before every renewal, not after.
With a register and a baseline, reconciliation becomes mechanical. Compare what you are entitled to with what is deployed and what is genuinely used, then reclaim the difference. Access assigned to people who changed role or left, tiers set above what the work needs, capacity provisioned for something that ended. Do it on a clock that runs ahead of renewal, so you renegotiate against real demand rather than last year's high-water mark. Gartner has put the achievable saving from disciplined licence optimisation at as much as 30% of software spend, and most of that comes from reclaiming, not renegotiating.
The gotcha is buying a tool instead of running a process. Reconciliation is mostly discipline and only partly technology, and a tool laid over no process simply reports the waste faster and in better colours. Decide who acts on the reconciliation, on what schedule, with what authority to remove access, before you decide what software will produce the report.
Stage 4: Make new demand state its return before it enters
No licence, customisation or capability enters the platform without a stated, auditable return.
Everything so far manages spend you have already committed. The larger prize is the spend you have not. Put a real assessment at intake, where every demand states its expected impact and its financial return, and the ones that cannot are held until they can. This is not bureaucracy, it is sequencing, the same case you will need at benefit-review time, written while it can still change the decision. The cheapest cost remains the one you never incur.
The gotcha is sponsor-marked homework. If the person requesting the spend also authors its value score, everything scores high and the gate becomes a queue with paperwork. Make the financial return auditable by someone who does not want the project, and let the board decline things. A gate that has never said no is furniture.
Stage 5: Price the cost that never reaches an invoice
Treat every customisation as a recurring charge, and review the estate each upgrade.
Some of the platform's most expensive commitments never appear on any bill. Every departure from the standard platform adds work to every future upgrade, more to review, more to retest, more that can break, and that drag compounds for as long as the customisation lives. Before each upgrade cycle, review the estate and revert what no longer earns its keep. The question is never whether a customisation was justified when it was built, only whether it is still worth what it now costs your platform and business.
The gotcha is invisibility, because no invoice ever arrives, nobody ever prices it. The visible bill stays flat while the platform quietly becomes more expensive to keep current, and by the time upgrades feel heavy the estate is years deep. If you do not put a price on customisation at approval time, the price gets paid at upgrade time, with interest.
Stage 6: Review value on the same rhythm as cost
Embed value measurement in the operating routine, and count the indirect returns too.
A cost discipline that only measures cost ends up cutting value, because savings are easy to see and returns are not. Put the value measures into the same routine as the cost reviews, regular health checks, benefit reporting against the Stage 2 baselines, and governance decisions taken from that reporting rather than from advocacy. Count the indirect returns deliberately, risk reduced, compliance evidenced, resilience improved, because they are real, they are usually large, and they are the first things an unmeasured platform loses. In the method well-run programmes follow, value management is a standing workstream of its own, just under 10% of the overall activities during a project are in this stream, not a slide written at year end.
The gotcha is the platform that looks like a pure cost. If nobody writes down the indirect value, the platform's defence at budget time rests entirely on direct savings, which shrink every year precisely because your optimisation is working. Measured value compounds, unmeasured value evaporates, and both happen quietly.
The six moves. The baseline is the one to protect.
Common failure modes
The renewal sprint. All six moves run once, hard, in the quarter before renewal, then nothing for a year. The estate refills, the baseline goes stale, and next year's sprint starts from scratch. The fix is boring on purpose, run the routine quarterly at modest intensity rather than annually at heroic intensity.
Handing it to procurement. Procurement can negotiate unit price, it cannot know which licences defend outcomes. Delegating the whole discipline gets you the best possible deal on things you should not be buying. Reclaim first, baseline first, then let procurement negotiate from strength.
Optimising into the muscle. With no value baseline and no indirect measures, a cost programme cannot tell fat from muscle, so it cuts whatever is least defended, which is usually the quiet capability holding risk down. That is how a platform delivers a headline saving and a worse year.
The bottom line
If you protect one move, protect the baseline. Everything else here is mechanics for defending a number, and without the number the mechanics degrade into cost cutting, which always takes the wrong things first because it cannot see what anything earns. Agree what the platform must return, in writing, with the people who control the money. Then the bill stops being an annual argument and becomes what it should have been all along, evidence.
Cost optimisation is not the art of spending less. It is the discipline of proving what the spend buys.
P.S. One test before your next budget round. Ask finance what the platform returned last year, in their numbers, not yours. If the answer is a pause, your problem is not the size of the bill, it is that nobody ever agreed what the bill was buying, and this playbook starts there.