Methodology

The Assessment Economics Framework

Every number AES produces comes from a small set of transparent coefficients. This page documents them so estimates can be audited, contested, and improved.

01 · Premise

Assessments are capital projects, not line items.

A defensible assessment carries research, item development, piloting, calibration, delivery infrastructure, and a decade of maintenance. Modeling only the visible costs produces the overruns the industry is known for.

02 · Cost surfaces

Four numbers describe the economics of any assessment.

CAPEX

Σ phaseCapex × intensity × mode × strategy × salary

One-time development cost, aggregated across included phases.

OPEX (annual)

Σ phaseOpex × intensity × mode × strategy × salary

Recurring cost per year: maintenance, hosting, item bank refresh, support.

TCO (horizon)

CAPEX + OPEX × horizonYears

Total cost of ownership over the planning horizon (1–10 years).

ROI

(candidates × price − TCO) ÷ TCO

Simplified: variable cost is baked into OPEX so revenue net of TCO is the return.

03 · Phase model

Nine phases anchor every project.

Baseline costs assume a mid-sized certification project. Adjust intensity per phase to reflect scope.

Phase	Base CAPEX (USD)	Base OPEX / yr (USD)
Research & Framework	45,000	—
Assessment Design	35,000	—
Item Development	60,000	15,000
Pilot Testing	50,000	—
Psychometric Analysis	55,000	10,000
Norming & Scaling	40,000	5,000
Technology & Platform	90,000	45,000
Deployment & Training	30,000	8,000
Maintenance (annual)	—	40,000

Intensity multipliers: light 0.6 · standard 1.0 · comprehensive 1.5. Language multiplier adds +35% per language beyond the first on research, design, item, and pilot phases.

04 · Delivery mode

Paper, CBT, and CAT carry different infrastructure footprints.

Phase	Paper	CBT	CAT
Item Development (capex)	1.0×	1.0×	1.4×
Pilot Testing (capex)	1.0×	1.0×	1.5×
Psychometric (capex)	1.0×	1.0×	1.6×
Technology (capex)	0.15×	1.0×	1.8×
Technology (opex)	0.2×	1.0×	1.5×
Maintenance (opex)	0.5×	1.0×	1.4×

CAT reduces seat-time and shortens test length in operation, but requires larger calibrated banks and platform investment upfront. See the CAT planner for the trade-off model.

05 · Strategy

Build vs Buy vs Outsource trades CAPEX for OPEX.

Strategy	CAPEX multiplier	OPEX multiplier	Note
Build (in-house)	1.0×	1.0×	Highest control, highest internal load.
Buy (license)	0.25×	1.4×	Fast start; recurring license fees dominate TCO.
Outsource (vendor)	0.75×	1.1×	Vendor delivery with reduced internal capacity requirement.

06 · Uncertainty

Every estimate ships with a confidence band.

The Monte Carlo tab in the estimator resamples the model to expose sensitivity.

Distribution

triangular(low, mode, high)

Applied independently to CAPEX, OPEX, candidate volume, and inflation.

Default band

±25% CAPEX · ±20% OPEX · ±25% volume

Reflects typical variance observed in psychometric project post-mortems.

Seed

Mulberry32(seed)

Deterministic PRNG — the same seed reproduces the same run for audit.

Reported metrics

P10 / P50 / P90 · P(ROI > 0) · P(overrun > 20%)

Executive-ready percentiles rather than a single point estimate.

07 · Limitations

This is a planning model, not a procurement quote.

Coefficients are heuristic anchors drawn from public benchmarks and practitioner experience. They are useful for direction-setting and executive framing. They are not a substitute for a scoped RFP, vendor bid, or actuarial cost-model. Use AES to decide whether to invest and how — then use vendor pricing to sign the cheque.