Transition programmes budget the state they are leaving and the state they are entering and fund the overlap between them out of neither, which is why the double-running period is the one nobody has authority to extend.
Put one question to the next transition review and watch what happens to the room: who is paying for the weeks in which both operating states are running at once?
The answers arrive in a predictable order. Finance says the cost sits inside the programme. The programme director says the approved budget runs to acceptance, and acceptance is behind us by the time the question bites. The operations director says the operating budget was built on one platform, one roster and one set of vendor charges. Nobody is being evasive. Each is describing their own instrument accurately, and the interval is in none of them.
That absence outlives the meeting. A cost nobody funded is a cost nobody may authorise more of. When the overlap needs another three weeks — because the data will not reconcile, because the new state is behaving correctly but slowly — there is no line to draw against and no signature that releases it. The decision migrates downward to whoever is standing at the cutover with the most information and the least authority, and at that level it can be settled only one way: compress the interval, and absorb what follows into operations, where it is recorded as operational underperformance rather than as the price of a transition the enterprise chose.
This is not an oversight that tighter budgeting discipline would catch. It is a property of the instruments themselves.
The Strategic Context
Consider a regional water authority replacing the platform that reads meters, calculates charges, holds arrears and administers hardship arrangements. The migration cannot be instantaneous, because a billing cycle is not an instant: accounts opened mid-cycle must be finished under the rules they started under, and payment plans agreed under the old arrangements must be honoured under the new. For at least one full cycle, both systems hold live customer positions, both produce numbers, and a group of people exists whose entire function is keeping two records of the same truth in agreement. The specifics are hypothetical; the shape is not — any transition spanning a cycle longer than a day has an interval in which both states are real.
The same shape appears without technology. A theatre company rebuilding its home venue keeps its season alive from a temporary space, running two front-of-house operations, two box offices, two sets of licences and insurances, and one technical crew stretched across both. Its subscribers experience one company. Its cost base carries two.
The interval belongs to the programme alone. The project that installs the new platform is complete when the platform is accepted; the operation is stable when one state is running. Between those conditions sits a period that exists only because of the relationship between them, which is the definition of programme-level work. At portfolio level the exposure is larger: an enterprise running four transitions a year may have three in overlap at once, and almost none computes that aggregate.
What Leaders Commonly Misread: Handover as an Event
The discipline teaches closure as a sequence of acts: verify the deliverables, obtain formal acceptance, transfer ownership to the party who will live with the result, capture what was learned, release the resources. Each happens on a day, and things that happen on days are naturally drawn as a workflow, with handover as a node in it. Where a schedule concedes duration at all, it concedes it to the preparation — an activity carrying a number of days — while the handover itself stays a milestone with no width.
The material used to train project managers contains a sharper version, worth describing because it survives inside the discipline's own teaching. One widely used transition brief pre-authorises a staged transfer: parts of the operation may move to the new site before the old one closes to the public. Two paragraphs later the same brief provides an external grant to sustain the enterprise through the period of full closure while the new site is established. Both sentences describe the same window — in the first, both states are partly live; in the second, nothing is live at all. The brief never notices that these are different intervals with different cost structures, and the work built on it inherits the confusion.
The funding device does a second thing, quieter and worse. Because the money is external, the overlap never enters the enterprise's own cost structure at all: no line anyone owns, no rate anyone computes, no duration anyone defends. It is somebody else's problem — and the teaching model reproduces in miniature the condition it should be inoculating against.
Reframing the Issue
The overlap is not a gap between two funded states. It is a third operating state, and it is the only one of the three with no owner.
Its cost base is both operating cost bases running together. It has labour that exists nowhere else in the enterprise's life — the reconciliation function, whose purpose is to hold two versions of the same reality in agreement, almost always described as temporary and almost always staffed from people who already have other jobs. It has an error class of its own: the customer, order or record that exists in both states or in neither, and can be created only during the interval. And it carries a suppressed-change cost: everything else the enterprise wanted to do to that part of the business is frozen while two versions are live.
Recognising the overlap as a state rather than a boundary changes what must be true before it starts. A state needs an owner, a cost, a duration and a defined exit. A boundary needs none of those, which is why treating it as one is comfortable and expensive.
This article does not price what a day of that state is worth in forgone revenue; the computable price of delay for a revenue-generating asset, and how rarely it is computed, is the subject of [Related article: Every Day of Delay Has a Price]. The concern here is prior to it: whether anyone holds the authority to buy one more day.
How the Overlap Loses Its Funder
The instruments are built at different times, from different objects
The project budget is built during planning from a decomposition of scope, and everything in it is an activity with a deliverable. Running an existing operation is not an activity with a deliverable; it is a condition. The operating budget is built annually from a model of the business as it is expected to be configured, and a transition year is modelled as the old configuration or the new one, rarely both. Neither instrument is wrong. They were never designed to meet.
Nobody sequenced the interval into anyone's authority
Permission to stage a transition is usually given early and enthusiastically, because staging reduces risk. What is rarely given at the same moment is permission to fund staging, which is a different thing. Staged delivery converts a single cutover into a period of coexistence whose length stays a live variable throughout execution, and if authority over that variable is not assigned when the staging decision is made, it defaults to nobody.
This article does not examine why the hardest and least reversible work is scheduled into that window in the first place — how ordinary sequencing rules push difficulty later is the subject of [Related article: From Sequencing Rule to Point of No Return]. Here the question is who pays for the window once it has been loaded.
Compression is the default, and it lands on the wrong person
Because the interval has no funded duration, it has no defended duration. Pressure to end it comes from every direction that can see a cost: operations watching two vendor invoices, finance watching contingency drain after acceptance, the executive watching a benefits date recede. Pressure to extend it comes from one direction only — the people running the cutover, who can see the reconciliation queue. They are junior to everyone pushing the other way.
The result is that the enterprise's most consequential irreversible decision is taken by its least senior participants, under duress, without a framework. Whether it can still return to the old state at that moment is a separate and equally neglected question, treated in [Related article: Reversibility Is an Asset That Decays]; this article assumes the door is closing and asks only who may hold it open a day longer. The interval after the transition entity itself dissolves — who holds the residue when no programme remains to hold it — belongs to Article 16 in this series, and begins where the overlap ends.
Decision Framework: The Overlap Funding Line
The overlap funding line is a single named budget line, opened at the moment the transition strategy is approved and never later, carrying five things and no fewer.
A rate. What one day of double-running costs: both operating cost bases, plus the reconciliation labour that exists only during the interval, plus the carrying cost of change work frozen while both states are live. This is arithmetic, not analysis, and a competent finance team produces it in a fortnight.
A duration in days. Not a date range: dates absorb slippage silently, a count of funded days does not. The count is the enterprise's stated view of how long it will accept being in the third state.
A named owner. One executive who can spend a day from the line without returning to a board. If two people must agree, nobody owns it.
A pre-approved extension allowance. A further number of days released against a stated test rather than a discussion — typically evidence that the reconciliation position is closing at a rate that reaches zero within the extension. An allowance released by argument is not an allowance.
A termination test for the old state. Written before the overlap begins by the people who will operate the new state, not by those delivering it, specifying what must be demonstrably true before the old state is switched off.
The line is auditable in an hour with four questions: what does one day cost, how many days are funded, who buys the next one, and what evidence buys it. A programme answering three of the four is not three-quarters funded. The unanswered question is where the failure will occur, because it is the one the organisation has already decided not to think about.
At portfolio level the line aggregates. Funded overlap days and their rates, summed across all transitions in a period, are the enterprise's real concurrency limit — a more honest constraint than delivery capacity, and one few boards have seen.
From Strategy to Execution
Immediate. For every transition in flight, run the four questions this week and publish the answers unedited, blanks included. Where no rate exists, instruct the programme to produce one within a fortnight. Expect at least one transition to reveal that its overlap is already running and unfunded.
Medium term. Make the overlap funding line a condition of investment approval, so no transition business case is complete without a rate, a day count, an owner and a termination test. Change the operating budget process in parallel: any function receiving a transition in the budget year submits its double-running rate, forcing the two instruments to meet before either is locked.
Long term. Hold the aggregate at portfolio level and treat it as a constraint on how many transitions may be in flight at once. An enterprise that can deliver six transitions a year but fund only two overlaps at a time has a capacity of two, and should plan as though it does.
Signals to Monitor
Watch for a cutover date drifting toward the end of a reporting period, which usually means the interval is being compressed to protect a number rather than an operation. Watch for the phrase "run in parallel for a short time" in a plan with no day count attached, and for reconciliation staff described as temporary in month one and still in place in month six, their cost absorbed into a business-as-usual cost centre.
Watch for project contingency drawing down after formal acceptance — the signature of an overlap funded from a source never meant to carry it. And watch for performance dipping in a function that has just received a transition, then being investigated as an operational failure rather than a transition cost.
Questions for the Leadership Team
- What did the overlap on our last completed transition cost, computed as both operating bases plus the labour that existed only during that interval, and in which cost centre did it land?
- For each transition now in flight, how many days of double-running are funded, and which named executive can authorise the next one without a board paper?
- Which of our current operating cost lines are still carrying residue from an overlap that was never formally closed out?
- When our last cutover was compressed, what was surrendered — which reconciliation was passed to operations, and how long did it take to clear?
- How many transitions will be in overlap simultaneously over the next four quarters, and what is the combined daily rate on the days they coincide?
- Who wrote the termination test for the old state on our largest current transition, and were they the people who will operate the new one?
Closing Perspective
The overlap gets paid for in every case. The only variable is whether the enterprise pays deliberately, through a line someone owns and can extend on evidence, or silently, through an operation that was never asked, staffed by people improvising a reconciliation function they were not funded to run, ending on a date set by whoever could no longer withstand the pressure.
The second route is not cheaper. It is the same money spent without a decision, plus the errors of compressing an interval that needed its length, plus the corrosion of asking a delivery team to make an irreversible call it had no authority to make. What an executive team chooses, in approving a transition with no overlap funding line, is not a lower cost but the removal of its own ability to intervene at the one point where intervention still changes the outcome.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.