Portfolio Economics¶
Portfolio Investment Summary¶
The program encompasses 10 active projects across four sites, phased over 24 months with milestone gates at Month 6 and Month 12.
Total Investment: $22.6M¶
| Project | Phase 1 | Phase 2 | Phase 3 | Total |
|---|---|---|---|---|
| PRJ-01 Ops-Maint Integration | $540K | $684K | $612K | $1.84M |
| PRJ-03 PdM Platform | $236K* | $1,302K | $540K | $2.08M |
| PRJ-04 Quality & Yield | $756K | $1,260K | $684K | $2.70M |
| PRJ-05 Cobble & Process Risk | $648K | $1,302K | $684K | $2.63M |
| PRJ-06 Maintenance Workflow | $612K | $648K | $1,218K | $2.48M |
| PRJ-07 Logistics | $684K | $1,218K | $684K | $2.59M |
| PRJ-08 Caster Chemistry | $390K | $576K | $450K | $1.42M |
| PRJ-09 Knowledge Capture | $684K | $1,218K | $450K | $2.35M |
| PRJ-10 Process Chemistry | $312K* | $1,430K* | $648K | $2.39M |
| PRJ-11 Coke Plant Ops | $390K | $1,260K | $510K | $2.16M |
| Total | $5.25M | $10.90M | $6.48M | $22.63M |
*Charter actuals (firm or indicative). All other costs derived from phase-type template methodology.
Vooban / IE split: ~$17.5M Vooban (77%) / ~$5.1M IE (23%). IE share increases in Phase 2-3 as domain roles (metallurgists, process engineers, change management) scale with plant-floor deployment.
Investment by Horizon¶
| Horizon | Timeframe | What Launches | Investment | Cumulative |
|---|---|---|---|---|
| H1 | Months 0-6 | Phase 1 for PRJ-01, 03, 06, 07, 09, 10, 11 | $3.5M | $3.5M |
| H2 | Months 7-12 | Phase 2 for H1 projects + Phase 1 for PRJ-04, 05, 08 | $11.6M | $15.1M |
| H3 | Months 13-24 | Phase 3 for all + Phase 2 for PRJ-04, 05, 08 | $7.5M | $22.6M |
Already Underway¶
Two projects have firm or indicative pricing through executed or submitted charters:
| Project | Phase | Status | Amount |
|---|---|---|---|
| PRJ-10 Concentrator (Phase 1) | Proof of Value | Charter submitted — firm pricing | $312,000 |
| PRJ-10 Concentrator (Phase 2) | Full System Build | Charter submitted — indicative | $1.3-1.56M |
| PRJ-03 PdM (Phase 1) | Cleveland PoV | Contract in final approval | $236,000 |
| Already committed | $548K firm + $1.3-1.56M indicative |
Three-Scenario Value Model¶
All values are annual at full deployment — the steady-state return once Phase 2-3 investments are operational across target sites. Returns ramp during deployment (see Self-Funding Cascade below).
Per-Project Value Scenarios¶
| Project | Conservative | Base | Optimistic | Value Drivers |
|---|---|---|---|---|
| PRJ-01 Ops-Maint | $8M | $14M | $20M | Delay misattribution correction × production recovery, 4-5 sites |
| PRJ-03 PdM Platform | $12M | $36M | $60M | Downtime avoidance across steel + fleet, 4 sites |
| PRJ-04 Quality & Yield | $15M | $29M | $43M | Defect reduction + auto-disposition + yield improvement, 3 steel sites |
| PRJ-05 Cobble & Process Risk | $8M | $22M | $35M | Cobble prevention + BF thermal optimization, 3 sites + 6 BFs |
| PRJ-06 Maintenance Workflow | $8M | $23M | $38M | Procurement cycle + inventory reduction + copilot efficiency, 4 sites |
| PRJ-07 Logistics | $22M | $41M | $60M | Shipping velocity + coil handling + scheduling, 4 sites |
| PRJ-08 Caster Chemistry | $11M | $22M | $33M | Off-chemistry prevention + plugging avoidance, 3 steel sites |
| PRJ-09 Knowledge Capture | $4M | $12M | $20M | Training time + onboarding + risk mitigation, 5 sites |
| PRJ-10 Process Chemistry | $8M | $13M | $18M | Reagent optimization + recovery improvement, Tilden |
| PRJ-11 Coke Plant Ops | $7M | $12M | $17M | Thermal efficiency + push scheduling + blend optimization, BH + scale |
| Portfolio Total | $103M/yr | $224M/yr | $344M/yr |
What Drives Each Scenario¶
| Scenario | Capture Rate | Adoption Assumption | Cross-Site Scaling | Base Metric Assumption |
|---|---|---|---|---|
| Conservative | 30-50% | Slow adoption, resistance at some sites | 2-3 sites for most projects | Low end of stakeholder-reported metrics |
| Base | 50-70% | Normal adoption with champion engagement | 3-4 sites for most projects | Midpoint of reported metrics |
| Optimistic | 70-90% | Strong adoption, full executive backing | All target sites | High end of reported metrics, validated by best-performing CLF sites |
The conservative scenario uses only values that would survive a skeptic's challenge. The optimistic scenario uses values that are defensible but require everything to go right. We recommend planning to the base case and budgeting to the conservative case.
Return on Investment¶
| Metric | Conservative | Base | Optimistic |
|---|---|---|---|
| Total Investment | $22.6M | $22.6M | $22.6M |
| Annual Value (Full Deployment) | $103M/yr | $224M/yr | $344M/yr |
| ROI Multiple | 4.6x | 9.9x | 15.2x |
| Payback Period | 10-14 months | 7-10 months | 5-7 months |
Payback periods reflect the ramp: value builds progressively as projects move from Phase 1 to deployment. The $22.6M is not invested upfront — it flows over 24 months, and early-returning projects generate value while later projects are still in development.
The Self-Funding Cascade¶
The program does not require a $22.6M commitment upfront. Early-returning projects fund the expansion.
| Months 0-6: Prove-It Phase | Months 7-12: Scaling Phase | Months 13-24: Expansion Phase | |
|---|---|---|---|
| Investment | $3.5M | $11.6M incremental | $7.5M incremental |
| Projects active | PRJ-01, 03, 06, 07, 09, 10, 11 (all Phase 1). Active charters: PRJ-10 ($312K), PRJ-03 ($236K). | H1 projects scaling to Phase 2 (multi-site, full systems). H2 projects launching Phase 1 (PRJ-04, 05, 08). | Full cross-site deployment. Phase 3 launches (corporate dashboards, platform scale). |
| Returns beginning | PRJ-09 knowledge capture (Mo 2-3); PRJ-03 PdM PoV (Mo 3); PRJ-01 misattribution analysis (Mo 4); PRJ-06 procurement fast-track (Mo 4); PRJ-10 recovery baseline (Mo 5); PRJ-07 plate hit list (Mo 4-5). | H1 projects delivering operational value at entry sites; PRJ-10 Phase 2 concentrator optimization (high-value); PRJ-03 Phase 2 PdM expansion + Tilden fleet; PRJ-07 Phase 2 BH coil velocity (GM's #1). | Cumulative returns: $30-80M+. |
| Expected returns | $2-6M | $12-35M cumulative | $30-80M+ cumulative |
| Gate position | Gate 1: $3.5M invested, $2-6M returned. Quick wins demonstrated, data connections operational. Decision: which projects advance to Phase 2? | Gate 2: $15.1M invested, $12-35M returned. Base case: program at or near breakeven. Decision: scope and pace of Phase 3 expansion. | Gate 3: Returns exceed investment in all scenarios. Program is self-sustaining. Annual run-rate value established. |
Payback Trajectory¶
| Milestone | Cumulative Cost | Cumulative Value (Base) | Net Position |
|---|---|---|---|
| Month 3 | $1.2M | $0.5M | ($0.7M) |
| Month 6 | $3.5M | $3M | ($0.5M) |
| Month 9 | $8M | $10M | +$2M |
| Month 12 | $13M | $22M | +$9M |
| Month 15 | $18M | $40M | +$22M |
| Month 18 | $22.6M | $60M | +$37M |
| Steady state | $22.6M | $224M/yr | Ongoing |
In the base scenario, the program crosses breakeven around Month 8-9. In the conservative scenario, breakeven is around Month 12-14. Even in the worst case, the program pays for itself well within the 24-month horizon.
Sensitivity Analysis¶
What If We Are Wrong?¶
| Assumption | Impact | Portfolio Still Viable? |
|---|---|---|
| Value estimates are 30% too high | Annual returns: $72-241M instead of $103-344M | Yes — conservative payback extends to 16-18 months, well within the 24-month horizon |
| Implementation costs are 30% over | Total investment: $29.4M instead of $22.6M | Yes — still 3.5x ROI in conservative scenario. Gate structure limits exposure. |
| Only 3 of 10 projects deliver | Remove 7 lowest-value projects | Marginal — depends which 3. Top 3 (PRJ-07, PRJ-03, PRJ-04) alone deliver $49-163M vs $22.6M cost. |
| Adoption is 50% slower | Returns delayed 6-9 months per project | Yes — payback extends to Month 18-20, gate structure redirects resources early |
| Cross-site scaling fails | Only entry-site value captured | Still positive — entry-site-only values total $30-60M vs $22.6M investment |
| Data quality is worse than expected | 4-8 week remediation per affected project | Yes — delays Phase 2 entry, increases cost by $50-100K per affected project |
The Cost of Doing Nothing¶
The status quo is not free. Without intervention:
- Knowledge attrition accelerates. Every retirement without knowledge capture is a permanent loss. Burns Harbor's coke plant section managers, Middletown's finishing experts, Cleveland's utility engineers — the clock is running.
- Reactive maintenance costs compound. The 70/30 reactive-to-planned ratio at Cleveland does not improve on its own. Each year of deferred integration means more repeat failures, more misattributed delays, more lost production.
- Competitive pressure increases. Mini mills are outcompeting on fulfillment agility (Burns Harbor shipping team's explicit concern). The coil velocity gap grows as competitors invest in logistics optimization.
- The data foundation opportunity narrows. The Databricks platform rollout is proceeding. Without project-driven data integration, the platform is infrastructure without purpose. With it, the platform launches with live, high-value data flows.
- Concentrator recovery gap persists. Every year at ~70% iron recovery instead of the achievable ~75-80% means tens of millions in foregone production value at Tilden.
CLF Data Integration Points¶
Several value estimates can be tightened with corporate-provided variables. The structure below shows where CLF actuals slot in — the document is complete without them, but the ranges narrow with confirmed data.
| Variable | Used In | Current Basis | What CLF Data Would Change |
|---|---|---|---|
| Production value per hour (by site) | PRJ-01, 03, 05 throughput calculations | Stakeholder estimates ($50K-$150K/hr range from transcripts) | Narrows throughput value ranges by 30-50% |
| Marginal profit per ton (by product) | PRJ-04, 07 quality and logistics value | Public data ($100-200/ton range) | Sharpens quality and logistics ROI |
| Equipment failure costs and frequencies | PRJ-03 PdM baseline | Partial from transcripts (bag house, scrubber cycle) | Validates or adjusts asset-specific PdM value |
| Annual production tonnage by site | All throughput-based estimates | Public data (~5M tons BH, ~3M tons others) | Confirms throughput base |
| Inventory carrying cost % | PRJ-06 inventory optimization | Industry standard (20-25%) | Adjusts working capital freed calculation |
| Reagent unit costs and consumption rates | PRJ-10 concentrator | Stakeholder-confirmed ($50M/yr total) | Validates reagent optimization value |
| Customer claim and quality hold history | PRJ-04 quality value | Stakeholder estimates (100 holds/day at MDT) | Sharpens disposition automation value |
A data request has been submitted through Erico to sites. If received before the readout, updated values will be integrated into the marked slots above without restructuring the document.
What the Numbers Mean for Decision-Making¶
Chad asked on March 24: "What if the ROI analysis points somewhere else?" Here is the answer.
If you fund the full program ($22.6M over 24 months): You get a portfolio returning $103-344M/yr at full deployment, with breakeven in 8-14 months depending on scenario. The gate structure means you're never more than one phase of investment away from a decision point.
If you fund only H1 ($3.5M, first 6 months): You get 7 working proof-of-value pilots, each producing tangible results within 8-10 weeks. At Gate 1, you decide which ones earned the right to scale. Worst case: $3.5M buys you validated answers about what works and what doesn't.
If you fund only the two active charters ($548K): PRJ-10 proves whether concentrator optimization works ($8-18M/yr upside for $312K). PRJ-03 proves whether multi-asset PdM works at Cleveland ($3-12M/yr upside for $236K). Both are self-contained with go/no-go gates.
The program is designed so that every dollar invested has a defined deliverable, a defined timeline, and a defined gate before the next dollar is committed.