Parametric Estimating Models

Navigation: estimating-techniques | index

Parametric estimating uses statistical relationships between cost and measurable project parameters to produce an estimate faster than bottom-up methods allow. It is the primary technique at FEL-1 and FEL-2, and the primary sanity check at FEL-3 and GMP.

The estimator’s dilemma: you need an accurate estimate before you have accurate design data. Parametric models are the bridge — they use known parameters (capacity, square footage, equipment cost) to generate defensible cost ranges with explicitly stated uncertainty.

---

Three Parametric Methods — When to Use Each

Method	Best Phase	Inputs Required	Accuracy	Primary Use Case
$/SF × complexity tier	FEL-1/2	SF + facility type + automation level	±25–40%	Building envelope + fit-out ROM
Capacity scaling (six-tenths rule)	FEL-1/2	Capacity + analogous project cost	±20–35%	Process plant TIC scaling from a known project
Equipment factored (Lang/Hand)	FEL-2/3	Major equipment FOB list	±15–25%	Total installed cost when P&IDs not yet available

These methods stack — use all three and triangulate. If they diverge by more than ±20%, identify why before proceeding.

---

Method 1: $/SF Parametric Models

The simplest parametric approach. Works well for building-dominated scope (Tier 2/3 supplier plants, light industrial CPG). Breaks down when process cost dominates (paint shops, F&B with heavy utilities, pharma).

Manufacturing Complexity Tiers

Tier	Facility Profile	Benchmark $/SF (2026, all-in TIC)
Tier 1 — Shell + basic fit-out	Pre-engineered or tilt-up, minimal process	$80–$150/SF
Tier 2 — Standard manufacturing	Standard HVAC, electrical, floor, dock	$150–$250/SF
Tier 3 — Process-integrated	Automation, compressed air, specialty drainage	$200–$350/SF
Tier 4 — Process-intensive	F&B sanitary, heavy MEP, specialty finishes	$300–$600/SF
Tier 5 — Process-dominated	Pharmaceutical, paint shop, heavy chemical	$600–$1,500+/SF

Critical rule: Assign a tier BEFORE selecting a $/SF range. Never pick a number from the middle without a documented rationale.

Adjusting $/SF for Specific Drivers

Adjustments are additive on top of the base tier benchmark:

Driver	Typical Adjustment
Union labor market (Northeast, Chicago)	+10–20% on labor-heavy scopes
Remote or poor-access site	+5–15%
Brownfield premium (tie-in, phasing, active plant)	+10–30%
FF&E / process equipment installation (OFCI)	+5–15%
Cold storage / controlled environment	+$50–$150/SF (mechanical dominant)
Explosion-proof / Class I Div 1 electrical	+$15–$30/SF in affected zones

---

Method 2: Capacity Scaling — The Six-Tenths Rule

When you have a comparable reference project, you can scale cost by production capacity:

C₂ = C₁ × (Q₂ / Q₁)ⁿ

Where:

• C₁ = known cost of reference project
• Q₁, Q₂ = capacity of reference and new project
• n = scaling exponent

Scaling Exponents by Plant Type

Plant Type	Scaling Exponent (n)	Notes
Average chemical / process plant	0.60	”Six-tenths rule” default
Specialty chemical / pharma (high automation)	0.40–0.50	Less responsive to scale due to fixed infrastructure
F&B processing	0.55–0.65	Mid-range; depends on line count vs. continuous flow
Automotive assembly (line additions)	0.65–0.75	More linear — each line has fixed cost
Power plants / utilities	0.70–0.80	Strong economies of scale

Example: Reference F&B plant at 10,000 cases/day cost $20M. New plant at 25,000 cases/day:

• C₂ = $20M × (25,000 / 10,000)^0.60 = $20M × 2.5^0.60 = $20M × 1.88 = $37.6M estimate

Boundary condition: The six-tenths rule degrades outside 0.1×–10× the reference plant scale. Do not use it for very small or very large departures from the reference.

---

Method 3: Equipment Factored Estimating (Lang and Hand Factors)

The equipment factored method multiplies major equipment purchase cost (FOB) by a factor to estimate Total Installed Cost (TIC). This incorporates: installation labor, bulk materials (piping, electrical, instrumentation, insulation, painting), civil/structural, and indirect costs.

Lang Factors

Plant Type	Original Lang	Updated (AACE RP)	Accuracy
Solid processing	3.10	3.89	±25–35%
Mixed solid-fluid processing	3.63	5.04	±20–30%
Fluid processing	4.74	6.21	±15–25%

The AACE International updated factors (RP 16R-90 and Wijoseno 2023) are systematically higher than the original 1948 Lang values because they reflect modern safety, environmental, and automation requirements.

How to apply: Sum all major equipment FOB costs → multiply by applicable Lang factor → this is TIC excluding buildings.

Hand Factors — More Granular Than Lang

Hand factors differ from Lang in that they apply different multipliers to different equipment categories rather than a single factor to the whole equipment list. This improves accuracy when the equipment mix is uneven (e.g., high vessel content but low machinery content).

Hand factor components (by equipment type, multiplied against individual FOB costs):

Equipment Type	Hand Factor
Towers / vessels	4.0–5.0
Heat exchangers	3.5–4.5
Process pumps	5.0–7.0
Compressors	2.5–4.0
Fired equipment (boilers, heaters)	2.0–2.5
Instrumentation	4.5–5.5
Electrical equipment	2.5–3.5

Sum (equipment FOB × respective Hand factor) = TIC for that equipment group.

What Lang/Hand Factors DO and DO NOT Include

Included in factor:

• Foundations and structural steel for equipment
• Process piping (Div 40) and insulation
• Electrical wiring and motor control (Div 26 for process scope)
• Instrumentation (Div 40/43)
• Painting and surface preparation
• Construction labor indirects

NOT included in factor:

• Building shell and envelope (Div 03–14) — price separately
• Site work and civil (Div 31–32) — price separately
• Owner costs (permit fees, owner’s engineer, validation) — price separately
• Contingency — add separately

---

Process Complexity Index — Qualitative Adjustment

Some estimators use a Process Complexity Index (PCI) to adjust parametric benchmarks. PCI isn’t a published standard — it’s a firm-developed scoring model. A simple 1–5 scale:

Factor	1 (Low)	3 (Medium)	5 (High)
Hazardous area classification	None	Some Div 2 areas	Div 1 throughout
Process utilities required	Basic compressed air	CAS + steam or CW	CAS + steam + CW + CIP + CO₂
Equipment density	Low (open floor)	Moderate	Dense (robotics, conveyors)
Regulatory requirements	Standard IBC	USDA/FDA compliance	FDA + cGMP + PSM
Clear height / structural	Standard 24–28 ft	30–40 ft	40+ ft heavy crane

PCI adjustment: For every point above 3 on the composite score, add 8–12% to the base $/SF estimate.

---

Building and Validating a Parametric Model

A parametric model is only as good as the historical data behind it. How to build one:

1. Collect 5–10 completed analogous projects — same facility type, similar scale
2. Normalize to a common pricing date — use ENR CCI to adjust historical costs to current
3. Apply location factors — normalize to a single base geography (e.g., national average)
4. Identify the primary cost driver — SF? capacity? equipment count? equipment FOB?
5. Run regression — plot cost vs. driver; fit a line (linear or power function); note R²
6. Document boundary conditions — project size range, facility type, excluded scope
7. Establish confidence interval — ±1 standard deviation from the regression line becomes your range

Minimum for a usable model: R² ≥ 0.80, n ≥ 5 data points, homogeneous project type. Below these thresholds, document why the model is still defensible or add a wider confidence interval.

---

Calibration Check: When Parametric Diverges from Bottom-Up

At GMP stage, you have both a parametric estimate (top-down) and a bottom-up estimate. If they differ by more than 15%, investigate before signing the GMP:

Common causes of divergence	Action
Process scope change since parametric	Revise parametric for updated scope
Bottom-up missing a major trade	Check for scope gaps
Site/location factor not applied consistently	Reconcile location factor in both
Bottom-up over-detailed in building, under-detailed in process	Check process division coverage

---

Sources

• AACE International RP 16R-90, 17R-97, 18R-97 — estimate classification and factorial methods
• Wijoseno, A. (2023) “Modifying the Lang Factor” — PM World Journal
• Garrett, D.E. (1989) Chemical Engineering Economics — capacity scaling exponents
• Seider, Seader, Lewin — Product and Process Design Principles (process plant scaling)
• Internal historical project database (most reliable source for firm-specific models)