Value at stake: what losing the data would cost
Value at stake is what the owner loses if the data in a file is gone. For files with maintainable elements ForensicBIM measures it from the same operate & maintain model that gives value in use: four parts, two one-off and two that follow the useful life, with three inputs you can set for your own asset.
The four parts
| Part | What it is | Follows the useful life? |
|---|---|---|
| Re-survey | Field capture of every item that can be surveyed: location, class, type, material, quantities and system, at the same minutes per item the maintenance model uses to price a gap, plus the one-off set-up tasks (setting out, a quantity take-off, mapping the systems). This is the cost of having the facts again. | No, one-off |
| Survey gap | The yearly saving the surveyable part of the data delivers, forgone while the survey takes. The gap is the capture time divided by one surveyor-year (1,600 hours), and never shorter than half a year or longer than two. | No, one-off |
| Concealed items | Pipes, ducts, cables, fittings, footings and piles cannot be surveyed without opening ceilings, shafts or the ground. They count for the lesser of an opening-up survey (five times the plain field capture) and their value in use over the life: where opening up costs more than the data is worth, the owner does without. | Yes |
| Decisions without the data | The works on the asset planned on missing or wrong information: maintenance, renewals and alterations. Measured as the asset's replacement value × the yearly works share × the share of that spend lost to bad information × the criticality from the file's O&M property sets × the data's readiness as delivered, for each year of the life, discounted like value in use (4 %, with 2 % yearly cost increase). It counts only the share of the works that the file's elements govern: installations about 45 %, structure and envelope 30 %, finishes and components 20 %, network runs 5 %, and unidentified proxies a quarter; a structural model does not carry the exposure of the installations it does not contain. | Yes |
Kept up to date, the data keeps its usefulness; left alone, it goes stale at 6 % a year and the figure drops. The report shows both.
Re-acquisition cost is the first part plus the opening-up survey of the concealed items: what it costs to have the data again, whatever it is worth. Deprival value is the lower of re-acquisition cost and value in use. For a long useful life value in use can be larger than value at stake: a rational owner re-surveys rather than forgoing twenty years of savings.
The inputs you can set
The asset's replacement value always comes from the file: its size, measured from the model, times a replacement cost per unit. Three inputs drive the decisions part. Each has a typical figure; Pro users set their own on the Valuation page, in the rung Value at stake, and reset it to the typical figure at any time. The file remembers them, like the hourly rate.
| Asset | Size from the file | Replacement cost per unit (euro region, 2026) |
|---|---|---|
| Building | The floor area the geometry audit measures; else the top area of the slabs; else the footprint times the number of storeys; else the model's plan extent times the storeys at 60 % (marked as estimated). | €2,300 to €4,500 per m² of gross floor area (more in the UK, the US, Japan and Hong Kong, less in mainland China and Brazil). |
| Bridge, viaduct | The deck area from the slabs in the model; else the length of the model. | €2,000 to €5,000 per m² of deck (US state DOT awards 2025: about $260 to $290 per sq ft); €20,000 to €80,000 per metre when only the length is known. |
| Road | The alignment or plan length of the model. | €3,000 to €15,000 per metre, from a two-lane rural road to a six-lane urban road (FDOT cost-per-mile models, structures excluded). |
| Railway | The alignment or plan length. | €10,000 to €30,000 per metre of conventional line (new main lines benchmarking). |
| Tunnel | The alignment or plan length. | €40,000 to €150,000 per metre (bored tunnels; the cheapest fully tunnelled rail lines cost about $60M per km). |
| Marine, other civil works | The plan length. | €40,000 to €150,000 per metre for quays and marine structures; €5,000 to €40,000 per metre for other civil works. |
A tunnel with a road through it counts as a tunnel, and a viaduct declared as a road counts as a bridge: the structure sets the cost, not the carriageway. Regions other than the euro region scale the infrastructure figures with their building cost. When no size at all can be measured from a file (a model of proxies without geometry, for instance), the decisions part is left out and the report says so.
Documented band, calculated band. The bands in the table are what the sources document. The calculation does not use them in full: it takes the middle quarter of each (the 37.5th to 62.5th percentile around the geometric middle), for instance €3,000 to €3,450 per m² out of the documented €2,300 to €4,500, 2.6 % to 3.0 % out of 2 % to 4 %, and 4.9 % to 6.2 % out of 3 % to 10 %. Three bands multiply in the decisions part, and used in full they would compound into a range too wide to act on. The report shows both figures next to each input.
| Input | Typical figure | Where it comes from |
|---|---|---|
| Replacement cost per unit | The build cost per m² of floor area for a building, per m² of deck for a bridge, or per metre for other infrastructure: the typical figures in the table above, in the currency of the hourly rate. Times the size measured from the file, it gives the asset's replacement value. | 2026 new-build and infrastructure cost benchmarks, indicative for early appraisal. Set the figure from your own cost plan, asset register or quantity surveyor. |
| Works per year | 2 % to 4 % of the replacement value a year for maintenance, repair and renewal. | The stewardship rule of thumb for facilities (National Research Council, Stewardship of Federal Facilities). |
| Lost to bad information | 3 % to 10 % of that spend: repeat visits, wrong parts, rework in alterations designed on wrong information, surveys done twice. | The order of magnitude of the NIST study on the cost of inadequate interoperability in operation and maintenance, and rework studies. |
The other three parts (re-survey, survey gap, concealed items) need no size: they follow the items in the file.
Why the range is a likely range
Every assumption in the model, minutes per item, repeat and incident shares, survey cycles, build cost and the two shares, is a band. Used at its extremes, every low at once and every high at once, the bands would stack into a range so wide it says nothing. ForensicBIM uses the model's own assumptions at the middle half of their spread, and the documented cost and share bands of the decisions part at their middle quarter, so the low and the high of every figure are a likely range with a middle estimate. The extremes of practice fall outside it. An input you set yourself has no spread: the range then narrows to the spread of the other assumptions.
The only way to narrow the ranges further is calibration against timed field surveys and maintenance records, which is planned, as the rebuild-cost model was calibrated against an expert panel.
What is not included
- Lane and track closures, traffic management, scaffolding and permits for the re-survey.
- Safety incidents and their consequences: these are beyond a share of the works spend and are deliberately left out.
- Design and construction decisions: value at stake starts when the asset exists. Before handover the rung is not applicable.
- Other deliverables from the same contract: only this IFC file is measured. Several files of one asset are valued one by one; a combined valuation of a set of files is in preparation.
Files without maintainable elements (a landscape model, a file of proxies only) keep the older, typical range for value at stake, shown as such. The figures are estimates for budgeting and prioritising, not a quote or an insurance valuation.