The construction specification of a physical asset is set by variables outside the site boundary, controlled by parties the enterprise has no contract with, and can be reset after the capital is committed.
Eighteen months ago the board of a hypothetical regional airport operator approved capital for a terminal expansion. Nothing on the site has changed since; the design is the one approved and the contractor is performing. The design manager has just reported that the external wall, glazing and roof packages must be reworked, that the cost movement exceeds the project's delegation, and that the programme will slip a quarter.
The cause is a revised map. The site now sits in a higher exposure band than when the capital paper was written, and the band determines what the building must withstand. Nothing inside the fence moved. Nobody in the enterprise did anything.
The chief executive's first question is almost always who got it wrong, and the honest answer is nobody. The variable that set the build standard was never inside the enterprise's control, never inside the site boundary, and never fixed. It was treated as an attribute of the project because it arrived in the design brief looking like one.
That is the mistake worth the board's attention, and it is not a design mistake. The enterprise bought land, a design and a contractor; it did not buy the inputs that set the build standard, and it cannot. Which of those inputs are controlled by neighbours, the weather and the state — and can be reset after your capital is committed?
The Strategic Context
Where a jurisdiction operates a banded hazard classification, the band that determines what an asset must withstand is the output of a function whose inputs sit outside the fence: the land use and vegetation around the site, the topography falling away from it, the regional exposure zone it sits within, and which version of the classification that jurisdiction has taken up. Nothing on that list is inside the boundary or owned by the enterprise, and every item can move.
The enterprise's contractual apparatus faces the other way. Contracts bind the designer, the head contractor and the supply chain — parties the enterprise pays to influence. The parties who can move the specification are an adjoining landholder, the body that publishes the maps, and the jurisdiction that decides which version applies. None is a counterparty: no clause to negotiate, no liability to allocate, no remedy when the requirement changes.
The consequence differs by level. For a single project this is a contingency question; for a programme of similar assets, a strategy question; for the portfolio, a capital-allocation question, because the sites carrying the highest exogenous exposure are not those carrying the highest project risk scores, and no portfolio view in general use separates them.
What Leaders Commonly Misread
The first misreading is that the specification is fixed at design freeze. A design freeze fixes the enterprise's intention; it has no purchase on the conditions that set the requirement, which move on their own schedule.
The second is that contingency covers it. Contingency is sized against delivery variance — weather days, labour rates, material escalation — from a distribution of past project performance. A reset of the requirement is not variance around the estimate; it is a change to the thing being estimated, arriving after the estimate has become a contract.
The third is that a favourable condition is an asset. An adjoining paddock kept clear lowers the rating today, and designing to it takes an unpriced dependency on somebody else's land management — no contract, no notice, no compensation when it ends. Enterprises book the saving and never the dependency.
The fourth is that someone in the delivery chain owns the risk. The designer priced against the inputs current when it was engaged; the contractor priced the design it was given. Whoever holds the design obligation when an input moves acquires an entitlement met from the owner's balance sheet. The allocation argument settles who submits the claim, not who pays it.
Reframing the Issue
Stop asking what the asset must be built to withstand. Ask who decides, how often they may decide again, and what each decision costs you.
Three boundaries follow. An enterprise that answers by building everything to the worst credible rating has taken a capital decision in the language of prudence; that abolition of a family of cheaper responses, and the unpriced cost of those that remain, is the subject of [Related article: A Zero-Tolerance Commitment Is a Capital Decision] rather than of this piece. Which version of a classification an organisation is bound by, and how upgrading to the current one can move it out of compliance rather than into it, belongs to [Related article: The Law Pins the Edition; the Standard Moves On]; adoption appears here as one input among several, none of which sits inside the fence. Where no published rating exists at all, the benchmark transfers to whoever wrote the specification, which is the territory of [Related article: Where the Standard Is Silent, the Benchmark Becomes Yours]; this article deals with the opposite case, in which a rating exists and somebody else sets it.
Four Variables, None of Them Yours
What the neighbours decide
Take a hypothetical vineyard and winery estate — cellar door, barrel hall, guest accommodation — sited within woodland because the setting is much of what the guest pays for. The adjoining property is grazing land. The neighbour sells; the new owner plants. The estate's rating rises, and with it the build standard required for the accommodation stage not yet started. The estate had no contract with the outgoing owner or the incoming one, no notice, and no remedy.
The reverse is equally unmanaged. A neighbour clears, the rating falls, and an estate already built to the higher standard has spent capital it cannot recover. Enterprises treat exogenous movement as a downside distribution when it is two-sided, and capture neither tail.
There is also a feedback loop most owners never see: what the estate plants for amenity feeds into its own rating. It is the one input inside the fence, and in ERANORTH's judgement it is rarely on the landscape designer's brief as a cost variable.
What the ground decides
Topography is the exception that proves the rule: exogenous to the owner's control but not to the owner's knowledge. Slope and fall around a site amplify exposure; they are fixed, cheap to determine, and determinable before anyone signs a contract of sale.
Site due diligence in most enterprises covers title, contamination, services and planning. It rarely covers the physical geometry that sets the build standard, so the geometry is discovered at design, after the land price is sunk. The information was available at a fraction of the eventual cost, and nobody asked, because the question belonged to no one on the acquisition team.
What the state decides
Two distinct decisions are routinely treated as one. The first is mapping: the body that publishes exposure maps revises them, and a revision can move a site between bands with nothing physical having changed. The second is adoption: a jurisdiction decides which version applies within it, and when.
For a single site this is a timing risk. For a programme it is a standardisation risk. A hypothetical airport operator running three regional terminals in three jurisdictions, with one module intended to be built three times, may find it sound in one, marginal in the second and needing substantial redesign in the third. The build-once-deploy-many economics that justified the business case quietly evaporate, and what the enterprise owns is three designs sharing a façade.
When the variable moves relative to when the money commits
The same input change costs radically different amounts depending on when it lands. At concept it is a design choice absorbed inside the option set. At detailed design it is rework and a programme slip. After commitment it is a variation on a priced contract, a re-approval, and potentially a fresh consent from a body on its own timetable.
No enterprise marks that boundary on a schedule, yet it is the most consequential date in the project's risk profile. Every exogenous variable resolved before it is the cheapest risk reduction available, and most are resolvable earlier than they currently are.
Decision Framework: The Exogenous Variable Register
The exogenous variable register is one sheet per site, listing every variable that sets or can reset the required build standard. Six fields per entry.
One — the variable, named specifically. Not "surrounding vegetation" but which parcel, on which boundary, in what condition today.
Two — who controls it. A named party, or "physical, nobody". Where the enterprise cannot name a controller, the site is not diligenced.
Three — contractual reach. Whether any contract exists with that party. The count of "no" answers is the enterprise's uncontracted dependency, and belongs on the capital paper as a number.
Four — reset mechanism and notice. What event moves the variable, and how much warning the enterprise gets. Zero-notice entries are the priority set.
Five — cost of one band of movement, at three points. Concept, commitment, post-commitment. The spread between first and third is the value of resolving the variable early, and the number that justifies pre-purchase investigation.
Six — trigger and owner. The event that would reveal the variable has moved, and the named person watching for it. "The design consultant will advise" is not an owner.
| Variable (hypothetical terminal) | Controller | Notice | Who pays after commitment |
|---|---|---|---|
| Land use on the adjoining parcel | Adjoining landholder | None | Owner, via a variation |
| Slope and fall around the site | Physical, nobody | Not applicable | Owner, at design, if it looked |
| The published exposure band | Body publishing the maps | A consultation nobody watches | Owner, via redesign and re-approval |
| Which classification version applies | The jurisdiction | Announced, rarely to the owner | Owner, differently at each site |
| Vegetation the owner plants | The owner | Full control | Owner, by choice |
The neighbour test. For each site, name the adjoining parties whose land-use decision would move the specification, and when the enterprise last spoke to any of them. An enterprise that cannot name them carries an unmanaged dependency inside the largest cost line of the project.
The adoption test. For a multi-site programme, record each jurisdiction's adoption state and whether a change is in consultation. Standardisation is approved only across sites sharing one; elsewhere the standard design is a drawing convention, not an economy.
The threshold. A variable that is zero-notice, has no named owner, and carries a post-commitment cost above the project's contingency line is red, and blocks capital approval until it is resolved, insured, priced into the base estimate, or accepted on the record by a named executive.
The governance test. No capital paper is approved without the register attached and the post-commitment column summed, reported separately from contingency — which is sized for delivery variance and already spoken for.
From Strategy to Execution
Immediately. Run the register on the two projects closest to commitment. Walk the boundary of each site and name the adjoining owners. Commission the topography on any site under offer before that offer becomes unconditional; the three-point cost spread justifies it.
Over the next two to four quarters. Move exogenous variables into acquisition due diligence alongside title and contamination. Change the gate paper template so the register and its summed post-commitment column are mandatory. Re-run the register at every design freeze and report the delta rather than the position.
Over the longer term. Treat the exogenous profile as a site selection criterion and price it into land acquisition, where it is cheapest to act on. Stop assuming standard designs transfer between jurisdictions, and rebuild the programme business case on the adoption states that exist. The enterprises that build well through the next decade will know, for each site, who can change their requirement and how quickly.
Signals to Monitor
Watch land transactions and development applications on adjoining parcels, which are public and almost nobody reads. Watch consultation notices on mapping and adoption, and whether anyone in the enterprise receives them. Watch the proportion of variations attributed to a changed external requirement rather than to design development. Watch whether any design freeze in the past year was accompanied by a re-check of external inputs. And watch the capital papers: if none the board has seen states an exogenous exposure separately from contingency, the exposure is being carried without having been decided.
Questions for the Leadership Team
- For our three largest committed projects, which parties outside the site boundary could change what we must build, and can we name them?
- What would a one-band movement in the hazard rating cost on each of those projects today, and who calculated it?
- When did we last determine a site's surrounding topography before, rather than after, the purchase became unconditional?
- Across our multi-site programme, which jurisdictions share an adoption state, and what did we assume when we priced the standard design?
- Which current design depends on a neighbouring land condition we neither own nor control, and what happens if that condition ends?
- In the last three years, how many variations came from an external requirement moving, and where did that cost land?
Closing Perspective
Capital papers answer whether the enterprise can afford the asset and whether it will earn. Both assume the asset is a known quantity. The specification of a physical asset is not, and it is not the enterprise's to know. It is a reading of the land around it, taken on a particular day, by parties under no obligation to tell the owner when they take it again.
The responsibility now sitting with the leadership team is to decide how much of that reading it will buy at the cheapest moment rather than the dearest. Every exogenous variable can be investigated before land is bought, priced before capital is committed, and monitored afterwards by a named person. None of that is expensive. What is expensive is discovering, eighteen months after approval, that the enterprise committed against a requirement which belonged to the neighbours all along.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.