How True Data works
True Data will provide on-demand, comparable benchmarks from a processed corpus of data sourced from thousands of real commercial construction projects spanning most sectors. It's meant to support an informed conversation between owners and the project teams who serve them about how a project compares with industry norms, both where it's standard and where it's unique. This page explains the methodology we're building and, for the organizations considering contributing project data, how we ensure that data is protected and used responsibly.
Compare against relevant cohorts, predict ranges for unknown values
Every benchmark is relative to a cohort of comparable projects. Depending on how much of a project you've described, True Data offers a blend of two things:
- Compare to cohorts. You already have a value, say a skin-to-floor ratio in the current design or $/GSF for a UniFormat category. We show where it falls within the distribution of comparable projects, so you can see whether it's typical, high, or low for work like yours.
- Predict ranges. If you don't have a value yet, we estimate the distribution implied by what you havetold us, combining statistical inference from cohort distributions with “stretch-to-fit” scaling of comparable projects' design breakdowns to your known dimensions.
Cost is normalized to the UniFormat Level 2 elemental level and where necessary made comparable by escalating for time and location. We deliberately stop there: we do not publish unit-price comparisons below Level 2, which keeps the conversation on real cost and schedule drivers and preserves the privacy of data contributors.
Program, design, sustainability, schedule, cost, and risk
A benchmark spans the dimensions that capture project outcomes, each compared against the cohort and most often shown as a box-and-whisker or violin distribution rather than a single number.
- Program. Scale and mix — units, beds, parking stalls, departments, and component and site area.
- Design efficiency. Ratios that signal how efficiently the building works — wall-to-floor, window-to-wall, and floor-area ratio (FAR).
- Cost. Normalized to UniFormat Level 2, composed to match the program of your project, and escalated for both time and location.
- Schedule. Major schedule milestones and the durations between them.
- Risk. Register entries scored by likelihood × impact, and the cost contingency they imply.
- Sustainability Embodied carbon intensity (up-front and whole-life), energy and water use intensity.
Alongside these, we capture the program requirements that drive cost and schedule — things like a LEED certification target or finish quality level. These condition the cohort and help explain why a project's numbers differ from a naive comparable, which is exactly the kind of driver the benchmark is meant to surface for discussion.
Benchmark report
Sourced from real projects, conditioned for comparability
Benchmarks are only as good as what goes into them. The corpus is built from the documents project teams already produce at their design and procurement milestones. Every contribution is conditioned and qualified before it can back a benchmark.
We source the primary artifacts a project generates at a GMP or other design-phase milestone, the same records the team relies on and uses to communicate with the owner as the project progresses:
- Program & requirements documents. The owner's program, basis of design, and requirements — unit and bed counts, department areas, finish quality, and certification targets like LEED — that explain whya project's numbers land where they do.
- Estimates Cost of Construction estimates which we normalize to UniFormat Level 2.
- Schedules Milestones and the durations between them extracted from the project schedule.
- Drawings Architectural and engineering drawings, from which the design-efficiency ratios and program quantities that other sources don't capture directly are derived.
- Specifications The written specifications that accompany the drawings — materials, systems, and quality standards — which qualify the finish and performance level behind a project's cost and design figures.
Conditioningreconciles these sources into a consistent, comparable shape — mapping costs to a common elemental taxonomy, aligning schedule milestones, and tagging the program drivers that define a project's cohort. Qualificationis the gate that decides whether a conditioned project is sound enough to count: we check that its milestone, estimate, and program are internally consistent and complete before it's allowed to influence any distribution. Projects that can't be reconciled are held back rather than allowed to distort a benchmark.
Want the specifics? See the exact data format every contribution is normalized to — each record type, field, and allowed value.
Cohorts are fuzzy, interactive, and comparable side by side
A comparable project doesn't have to match yours on every attribute. We seed one or more default cohorts from a project's known attributes, then let you adjust them in the report itself to match what's important for your project.
- Fuzzy membership. Cohorts are seeded from attributes like type, sector, region, delivery method, finish level, and floor-area range to provide relevant comparables. A cohort is not an exact match on every parameter.
- Interactive slicing. Drill into an attribute donut chart (e.g. sector or wage basis) or filter a shown distribution (e.g. total GSF) to refine a cohort or build a new one.
- Side-by-side comparison. Compare multiple cohorts at once — “union vs. open shop,” or “this metro vs. national” — to see how a single attribute moves the numbers.
This is the heart of accuracy and trustworthiness: a benchmark is only as credible as your ability to see what kind of projects back it and interpret the information. So cohort composition, attribute distributions, and approximate size are visible — but strictly as aggregates, never as a list of identifiable individual projects.
No single contributed project can be identified
Your project data is never shown to other users directly — it's only ever surfaced as aggregate distributions across cohorts of similar projects, never as identifiable individual values. We're building our privacy protections on the principles of differential privacy, the same rigorous standard used by organizations like the U.S. Census Bureau and Apple, which adds carefully calibrated noise and limits how much any single project can influence what's displayed. This approach is specifically designed to prevent reconstruction of an individual project's underlying data from any benchmark view or combination of views.