Standing up project and portfolio management
A capital-intensive organisation getting control of the largest thing it does — where the problem is not scheduling, it is that the capital programme's financial position is reconstructed rather than reported.
A composite worked example, not a client. It is assembled from patterns that recur in capital-intensive organisations under external funding scrutiny, and from this library's model. No organisation is described, and every derived section is computed from the same model the rest of the site runs on.
Where it stands
Capital is the largest thing the organisation spends and the least visible thing it reports. The programme is planned in one tool, costed in the finance system, delivered through contractors managed in email, and reported in a monthly spreadsheet assembled by hand.
Project costs arrive from three directions: contractor invoices through payables, internal labour through payroll, and materials and plant through the works system. Nothing joins them at the project until somebody does it in a spreadsheet.
Work in progress ages. Projects are practically complete long before they are financially closed, so capital sits in work in progress, depreciation starts late, and the asset register lags the network by months.
Nobody can answer the two questions that matter at once — what has this project cost to date, and what is it forecast to cost — from the same source.
The portfolio is a list, not a portfolio. Projects are approved individually against a funding envelope that is managed centrally, so trade-offs between them are made in conversation rather than against a model.
Why now
- —Capital delivery is reported externally against a funding determination, and the reporting is manual enough that its accuracy depends on a small number of people.
- —Capitalisation lags delivery, which misstates both the asset base and the operating result.
- —Forecast to complete is produced by asking delivery managers, which means it is optimistic and unauditable.
- —Contractor spend is committed in email before it is committed in a system, so commitment against budget is known late.
- —None of those is an end of support date, which makes this the harder kind of business case: there is no clock, only cost.
What that scope actually implies
5 areas picked, 13 topics in scope across 4 modules
The end-to-end it sits on
Plan-to-Deliver (Project)
Projects and Portfolio (PPM)
Bill-to-Recognise
Projects and Portfolio (PPM) → Financial Accounting
Acquire-to-Retire (Assets)
Financial Accounting → Enterprise Asset Management
Acquire-to-Retire (Physical)
Enterprise Asset Management → Financial Accounting
Procure-to-Pay
Supply Chain Management → Financial Accounting
Deliberately out of scope
Exclusions cause scope fights when they are assumed rather than written. Each one here has the reason attached.
The general ledger
Stays. The model flags it as a prerequisite this scope depends on without including — which is correct and is the thing to watch: project costing posts into a chart of accounts nobody is changing, so the project structure has to fit the ledger rather than the reverse.
Detailed scheduling and site delivery tools
Stay with the delivery teams. The programme needs the cost, commitment and completion signals out of them, not to replace them.
Maintenance and works management
Out for this release. Renewal work blurs the boundary between maintenance and capital, so the classification rule matters more than which system the work order lives in.
Payroll
Out of scope, in scope as an interface. Internal labour is a real project cost and it originates in payroll, so timesheet-to-project costing has to be designed even though payroll is not changing.
The boundary
Every system the finance core exchanges data with, and what happens to that interface. An interface with no disposition is an interface somebody will discover during testing.
Contractor commitment and spend
Rebuild
Contracts, purchase orders, receipts and invoices, against project←Procurement and payables
The interface that makes commitment visible before invoice. Without it the programme is reported on spend, which is the number that arrives too late to act on.
Internal labour costs
Rebuild
Timesheet hours costed to project and activity←Payroll
Internal effort is a real capital cost and is the one most often left out, which understates project cost and overstates the operating result.
Materials and plant
Not decided
Issues, plant hire and usage against project←Works and stores systems
Whether these route through the project or straight to the ledger has not been settled. Until it is, project cost is incomplete by an amount nobody has quantified.
Capitalisation and the asset register
Rebuild
Capitalised project costs, componentised, with in-service dates→Finance and asset management
The highest-value interface here. It is where the capital programme becomes an asset, where depreciation starts, and where the financial and physical registers are supposed to agree.
Project postings to the ledger
Rebuild
Cost, accrual and capitalisation journals→General ledger
Into a chart of accounts nobody is changing. The project structure has to fit it, and that constraint should be established before the structure is designed rather than discovered during build.
Capital programme reporting
Rebuild
Delivery and financial position by the funder's categories→Board, funder and regulator reporting
Currently assembled by hand. The replacement either fixes this or inherits it, and inheriting it should be a decision rather than an outcome.
Contributed assets
Not decided
Asset details, valuation and handover documentation←Third party contributors
No invoice, no cash, and usually no designed path. Decide it before the first one arrives in the new system.
Flows the model says straddle this boundary
Derived from the scope, not written here: hand-offs where one side is being changed and the other is not. Compare them against the boundary list above — anything present here and missing there is a gap.
Sub-ledger postings←Accounts Payable
General Ledger · continuous · without it: Payables detail and the control account disagree.
Depreciation journal←Asset Management
General Ledger · monthly · without it: Asset carrying values drift from the ledger.
Net pay file→ABA
Payroll · per pay run · without it: Staff are not paid.
Depreciation journal→General Ledger
Asset Management · monthly · without it: Carrying values drift from the ledger.
What this context does to a generic finance build
Capital versus operating is the decision, not a classification
Where the boundary sits — what is renewal and what is maintenance, which overheads are capitalisable, when a project becomes an asset — determines the reported result, the asset base and what can be recovered through prices. It is a finance decision with an operational input, and it has to be written as a rule rather than judged per job.
Work in progress is a balance with an age
Capital sitting in work in progress is not neutral: depreciation has not started, the asset is not in the register, and the longer it sits the harder it is to substantiate. Ageing work in progress deliberately, with an owner and a threshold, is the control most often missing.
Project structure has to serve two reports
Delivery wants the programme by project and phase; finance and the funder want it by asset class, service and regulated category. One structure has to carry both, which makes the project and work breakdown design a joint artefact rather than a delivery preference.
Commitment matters more than spend
A contract awarded is money committed, and in capital delivery the gap between commitment and invoice is months. A programme reported on invoiced spend is reported on the wrong number, and the overrun is already locked in by the time it appears.
Contributed and gifted assets
Infrastructure contributed by third parties arrives as an asset with no purchase order, no invoice and no cash movement. It needs a designed path in, or it becomes a year-end journal nobody can evidence.
Government procurement applies to the contractors
Most capital spend leaves the organisation through contracts. Government purchasing policy, local content and social procurement obligations and the associated disclosure all attach to that spend, and the system has to evidence them rather than merely permit them.
Funding is determined, not requested
Where an external body sets the envelope, capital is reported against a determination on that body's categories. The chart of accounts and the project structure have to produce those categories natively or the submission stays a spreadsheet.
The order to do it in
The grouping is computed from the model's dependency edges. The goal and the exit condition for each release are a delivery judgement.
Structure and commitment
Procurement · Asset Lifecycle · Project Planning
Project and work breakdown structure agreed against both the delivery view and the ledger, the asset hierarchy it will capitalise into established, and contractor commitment captured at award rather than at invoice.
Done when The capital programme reported from the system on committed cost, in the funder's categories, without a spreadsheet in the middle.
Cost to asset
Accounts Payable · Project Costing
Every cost stream landing on the project, capitalisation running to a written rule, work in progress aged with an owner and a threshold, and forecast to complete produced by a method rather than by asking.
Done when A project closed financially within a defined period of practical completion, capitalised and componentised into the register, with depreciation starting on the in-service date rather than the closure date.
Recovery and reporting
Project Billing
Recoverable and billable work invoiced from project cost, and the external capital submission produced from the system.
Done when The funder submission and the board pack generated from the same source as the ledger, reconciling to it without adjustment.
The model also flags General Ledger, Accounts Receivable as depended on but not included. Fine if they are staying as they are — a gap otherwise.
Decisions to settle
Where is the capital and operating boundary, in writing?
It moves the reported result, the asset base and what is recoverable. Judged per job it will be inconsistent, and inconsistency is what an auditor tests for.
By: Before the project structure is designed. Everything downstream encodes it. · Asset Management
What is the project and work breakdown structure?
It has to serve delivery reporting and the funder's categories at once. Designing it for one and mapping to the other afterwards is the version that ends in a spreadsheet.
By: Before configuration, jointly between delivery and finance. · Project Planning
Do materials and plant route through the project?
If not, project cost is incomplete and nobody can say by how much. If so, the works and stores interfaces are in scope whether or not those systems are changing.
By: With scope, because it determines the interface list. · Project Costing
Is internal labour capitalised, and at what rate?
It is real cost and it is the one most often omitted. Excluding it understates the asset and overstates the operating result; including it needs a defensible rate and a timesheet discipline that does not exist yet.
By: Before the payroll interface is specified. · Project Costing
What closes a project?
Practical completion and financial closure are different events, and the gap between them is where work in progress ages. Without a rule and an owner, projects stay open for years.
By: Before the first release goes live, with a threshold and an escalation. · Asset Lifecycle Management
Who owns the forecast?
A forecast produced by asking delivery managers is optimistic and unauditable. One produced from commitment, actuals and a method is defensible. The difference is a governance decision, not a system feature.
By: At mobilisation, because it decides what the reporting is built to produce. · Project Costing
Risks worth writing down
The project structure is designed for delivery alone.
So: Every external report needs a mapping layer, and the mapping layer is a spreadsheet maintained by one person — which is the problem this programme exists to solve.
Instead: Make the structure a joint artefact of delivery and finance, and test it against the funder's actual reporting categories before it is configured.
Capitalisation is treated as a finance month-end task.
So: Work in progress ages, depreciation starts late, the asset register lags the network, and the asset base reported externally is wrong.
Instead: Design capitalisation as a joint process with asset management, triggered by an in-service event owned by delivery, with work in progress ageing reported every period.
Commitment is not captured until invoice.
So: Overruns are visible months after they are locked in, and the programme is governed on a number that is always out of date.
Instead: Capture commitment at contract award, report committed against budget as the primary measure, and treat invoiced spend as the lagging indicator it is.
Internal labour stays outside project cost.
So: Project cost is understated by a material amount, whole-of-life comparisons are wrong, and the operating result carries cost that belongs in the asset.
Instead: Settle the capitalisation position and the rate early, and build the timesheet discipline before the interface, because the interface cannot create data that is not captured.
The programme is scoped as a scheduling tool.
So: A better plan, reported the same way as before, with the financial position still assembled by hand.
Instead: Hold the business case to the reporting outcome rather than the planning capability, and make producing the external submission from the system a gate criterion.
Delivery managers are asked to become cost accountants.
So: Data entry resented and done badly, which makes the reporting worse than the spreadsheet it replaced.
Instead: Design for what a delivery manager already knows and can confirm; derive the rest. Where a field needs a finance judgement, route it to finance rather than to site.
Readiness
Written capitalisation policy
gap
Including the treatment of overheads and internal labour, and the renewal versus maintenance boundary. Usually exists as practice rather than as a rule.
Work in progress position understood
gap
The current balance, its age, and how much relates to projects that are practically complete. It is almost always larger and older than expected.
Funder reporting categories mapped
gap
The actual categories the external submission requires, obtained before the project structure is designed rather than after.
Chart of accounts constraint accepted
unknown
The ledger is not changing in this scope. Whether the project structure can fit it, or whether that assumption breaks, should be tested early.
Timesheet discipline
gap
Internal labour cannot be costed to projects that nobody books time against. This is a behaviour change, not a configuration.
Asset register condition
unknown
What the capital programme will be capitalising into, including whether the financial and physical registers currently agree.
Delivery and finance sponsorship, jointly
gap
This programme fails when it has one of the two. The structure, the capitalisation rule and the forecast method all need both.
The artefacts this programme has to produce
Each one has a guide behind it. The library is the general case; this list is the order it is needed in.
Capitalisation policy, including overheads, internal labour and the renewal boundaryFinance
Project and work breakdown structure, serving delivery and the funderDelivery and finance
Cost capture design across contractor, labour, materials and plantFinance
Commitment capture at contract awardProcurement
Capitalisation and in-service trigger, with asset managementFinance and asset management
Work in progress ageing control, with owner and thresholdFinance
Forecast to complete method, and who owns the numberDelivery and finance
Contributed asset recognition pathFinance and asset management
Recoverable and billable work rulesFinance
Capital programme reporting pack, produced from the systemFinance and delivery
Stage gate definitions and approval authority for the portfolioPMO
Change impact assessment for delivery managers and site teamsChange
Argue with it
The scope, the phase order and the straddling interfaces are all computed from the model. If a phase looks wrong, the dependency edge behind it is the thing to change — say so here.