Sizing the return on Keystone.
By buyer archetype, with transparent assumptions and a worked example. Below: an interactive calculator you can plug your own numbers into. Below that: the full economic framework with every assumption stated. Print-friendly.
ROI calculator by buyer archetype.
Adjust the inputs. Outputs recompute live. Every formula is visible below the output. Nothing is sent to a server; the calculator runs entirely in your browser.
Caveats: These are per-archetype defaults from the framework below. Every number in the outputs recomputes in real time as you adjust inputs. This calculator does not model financing cost of pilot upfront, ramp curve (loss-ratio improvements take 2-3 years to materialize), or counterparty risk. For a defensible per-partner ROI on your specific situation, bring 25-100 records to the six-week pilot.
Framing
Keystone charges per verified event, never per SaaS seat. Every buyer pays for one engine, or a combination. Return is computed against the specific bottleneck that engine unblocks. There is no generic Keystone ROI number; there is a per-engine, per-buyer framework.
Manufacturer (Registry + Modular Index)
Value driver: monetization of QA and inspection data your factory already produces. Every module that ships also produces a Keystone record that downstream buyers pay to read.
Cost: per-asset registry toll at insert (typical: sub-1% of module wholesale price). Zero fee during the six-week design-partner pilot.
Return sources: higher institutional pool pricing on your collateral (lender pays Capital Rail, uplift flows to your terms); resilience-priced insurance premium differentials on your product line (verified strong-attribute units quote sub-90bps vs 140bps for unverified); reduced warranty claim disputes (immutable per-asset record); IRA §45X provenance revenue.
Worked example. A 500-modules-per-year manufacturer. Registry toll: 500 × $50 = $25K/year to Keystone. Pool-grade uplift on financed inventory (assume 60% of units pool-financed at a 40bps rate differential on a $50M inventory book): $200K/year. Insurance premium differential (assume specialty modular carrier passes back 20% of the pricing spread): $80K/year. Net first-year return: roughly $255K gross benefit against $25K cost. Multiplier: ~10x. Sensitivity depends heavily on pool participation rate and insurance pass-through.
State HFA / program office (Disbursement Rail)
Value driver: recover a percentage of drawdown lag by moving verification from forensic audit to release-time gate. Federal and state infrastructure programs consistently lag scheduled drawdown by 6-18 months. Every month of recovery is a month of program impact accelerated.
Cost: per-release SaaS usage fee in basis points on authorized releases (typical: 20-30bps of release amount). Zero fee during pilot.
Return sources: faster drawdown (program KPI); lower forensic-audit remediation cost (IG audits terminate faster on assets with immutable state records); reduced incorrect-release exposure (releases are gated on verified installed milestone, not invoice trust).
Worked example. A state HFA disbursing $200M annually across manufactured/modular affordable-housing units. Fee: 25bps × $200M = $500K/year to Keystone. Drawdown-lag recovery of 3 months on 25% of the pipeline: $12.5M in accelerated deployment (program-KPI value, not direct cash-recovery). Forensic-audit remediation savings: assumes $500K/year in reduced remediation labor. Break-even at zero forensic savings. Positive on any acceleration.
Specialty MGA (Insurance Rail)
Value driver: reduce loss ratio by pricing off verified attributes. Currently MGAs quote modular / prefab / industrialized-construction lines at book rates because verification is unavailable. Verified strong-attribute units subsidize weak-attribute units in the same book.
Cost: per-policy SaaS subscription or a shared share of the partner MGA's commission on Keystone-priced business.
Return sources: tighter book (verified sub-90bps vs unverified 140bps on the same class); defensible underwriting with a per-quote signed record; reduced loss ratio (avg 3-8 points over 2-3 years, benchmark to Verisk residential-construction rating tables).
Worked example. A modular-lines specialty MGA writing $30M in annual premium on a 62% loss ratio. Loss-ratio improvement of 4 points over 2 years: $1.2M/year in reduced loss cost. Keystone shared commission share of 5% on written premium: $1.5M/year to Keystone. Net direct benefit modest; defensibility benefit substantial (regulator, reinsurance counterparty).
Lender / GSE / syndicator (Capital Rail)
Value driver: unlock pool-grade collateral pricing on modular / industrialized-construction pools that currently cannot achieve institutional pricing because per-asset verification is unavailable to the rating agency and to the broker-dealer placing the securitization.
Cost: per-pool analytics fee (typical: 5bps of collateral value) plus optional revenue-share on placed placements with NRSRO / broker-dealer partners.
Return sources: tighter senior tranche spread (verified-pool AAA vs unverified NR); lower gain-on-sale volatility across issuances; new asset class opened for institutional pool structuring.
Worked example. A lender pooling $200M annually of modular collateral. Fee: 5bps × $200M = $100K/year to Keystone. Senior tranche spread improvement of 25bps on 70% of the pool: $350K/year in reduced funding cost. Multiplier: ~3.5x against direct cost.
Applied AI ROI
Applied AI ROI is framed at the individual feature level, not at the engagement level. Every AI feature we ship has a measurable business metric attached at the start of the Analyze phase (revenue lift, retention delta, ops-margin change on a specific workflow). The prioritized shortlist in the Analyze deliverable ranks features by expected impact per unit build cost. See Applied AI pricing.
How to use this
Plug your own volumes into the worked examples above. Every assumption is stated and replaceable. If the framework does not fit your specific buyer archetype (unusual pool structure, non-standard payment rail, non-US regulatory posture), the six-week pilot is the fastest way to produce a defensible return estimate against your actual data.
Produce your own numbers against real data.
The six-week design-partner pilot ingests 25 to 100 of your anonymized records and produces the ROI defensible for your specific buyer archetype.