| To | Gate Review Board; Stage 1 functional leads; Executive Committee (information) |
| From | C. Tyrrell, NPD Program Manager & Chair, Gate Review Board |
| Date | 04 May 2026 |
| Subject | Gate 1 recycled 30 April 2026 — findings, remediation and return date |
| Reference | Issue I-01 · Governance Model §12 · Gate Decision Framework §4 |
Gate 1 was convened on 30 April 2026 to consider release of the Stage 2 development tranche of $11,640,000. The Board recycled the package. This memorandum records what was found, what was not found, why it happened, what is being done, and what it costs. It is issued within two business days of the gate as required by Governance Model §7.
1. The decision
Under the Gate Decision Framework a gate applies must-meet criteria first, then scores the should-meet criteria. The Board never reached the scoring tier. Two should-meet criteria were rated U — unscoreable, which under §5 of the Framework obliges the Chair to table a recycle recommendation. The recommendation was tabled and carried.
| Item | Record |
|---|---|
| Outcome | RECYCLE |
| Weighted score | Not scored |
| Stage 2 tranche | Withheld — $11,640,000 not released |
| Return date set | 11 June 2026 |
| Remediation window | 04 May 2026 to 05 Jun 2026 |
2. Finding 1 — Distribution commitment, rated U
What was tabled
The Year 1 premium forecast of $185,000,000 was built from internal channel modelling and historical production on the existing annuity line. No written indication from any distribution partner was included in the pack.
Why it could not be scored
The evidence standard at Framework §4 admits "a written commitment from an accountable party" and explicitly excludes "a relationship manager's confidence." What was presented was the second. The volume forecast is the single largest driver of the return calculation; a criterion that carries 20% of the weighting and drives the economics cannot be scored on an unevidenced input.
Board comment recorded: the objection was to the absence of evidence, not to the plausibility of the number.
3. Finding 2 — Financial return, rated U
What was tabled
Product IRR of 14.6% against an 11.0% hurdle.
Why it could not be scored
The calculation used a capital charge that had been superseded by Corporate Actuarial's assumption refresh (DEP-07). The output was of an admissible kind — a model output with an identified version — but its input was not current, which Framework §4 places outside the standard.
This was not a presentational point. Recomputation on the current charge reduced the projected return to 13.4%, a fall of 120 basis points. The case still clears the hurdle, but by 240 basis points rather than 360.
4. What the Board did not find
This is recorded explicitly because a recycle is routinely misread as a soft rejection, and the distinction changes how the program should behave for the next six weeks.
- The Board did not find the product concept unsound.
- The Board did not find the volume forecast implausible — it found it unevidenced.
- The Board did not reject the pricing basis. All five must-meet criteria were on track to certify.
- The Board did not instruct any change to product design, scope or launch date.
Accordingly the remediation at §6 is confined to evidence and recalculation. No design work was undertaken during the loop, and no scope was changed. Treating a recycle as an invitation to redesign is how a six-week loop becomes a six-month one.
5. Root cause
Both findings have the same cause, and it is a sequencing failure rather than a diligence failure. The Stage 1 plan scheduled channel validation and the capital-charge refresh to complete after the business case was drafted, on the assumption that both would confirm what had been modelled. The pack was therefore assembled to the gate date rather than to the availability of its own evidence.
Two consequences followed directly. Channel commitments were still in negotiation when the pack closed, so modelling stood in for evidence. And the capital charge in the model was the one current when the model was built, not the one current when the gate convened — an eleven-week gap during which the refresh landed.
6. Remediation plan
| # | Action | Owner | Due | Status |
|---|---|---|---|---|
| 1 | Secure written Year 1 indications from the three largest IMO partners | R. Castellanos | 05 Jun 2026 | Complete |
| 2 | Recompute product economics on the current capital charge; restate the return and all sensitivities | S. Ravichandran | 05 Jun 2026 | Complete |
| 3 | Re-run the downside case against the restated return | B. Trombley | 05 Jun 2026 | Complete |
| 4 | Add a standing dissent section to the gate package template | C. Tyrrell | 05 Jun 2026 | Complete |
| 5 | Re-table the package at the second convening | C. Tyrrell | 11 Jun 2026 | Complete |
Actions 1 and 2 are direct remediation of the two findings. Action 4 is not — it addresses the root cause at §5 by giving unresolved analysis somewhere to sit inside the pack rather than being resolved by omission on the way to the gate.
7. Cost of the loop
The loop ran 6 weeks and cost $232,000, met by a contingency draw authorised by the Board at the second convening under Governance Model §15.
| Retained core team - pricing, product, distribution analytics (6 weeks) | $148,000 |
| External capital-charge refresh and recomputation support | $34,000 |
| Channel commitment negotiation and legal review of indications | $28,000 |
| Extended program management and gate re-preparation | $22,000 |
| Total | $232,000 |
Stage 1 consequently closed at $2,412,000 against a tranche of $2,180,000. The overrun is exactly the cost of this loop. It was funded from reserve rather than absorbed into the Stage 2 release, so that the cost of the recycle remains visible instead of disappearing into development spend.
8. Schedule consequence
Launch on 06 March 2028 is set by the distribution cycle and the filing window behind it, and moving it was not considered a reasonable response to a self-inflicted six-week loop. The consequence is that the compression is now carried inside Stage 2 rather than shown as a delay, and that is a materially less comfortable position than the plan the Board approved at Gate 0.
Two live risks are elevated by it. R-04 — platform configuration capacity — now has less float against Cordelane's quarterly configuration windows (DEP-03). R-08 — illustration engine readiness inside the filing window — loses the same 6 weeks. Neither was re-scored at the second convening; both should be examined at Gate 2 specifically in light of this compression rather than in general.
9. Process changes adopted
Three changes were adopted before the second convening. They are recorded here because the value of a recycle is largely in what it changes about the next gate.
| Change | Where it now lives |
|---|---|
| The U (unscoreable) rating formalised on the scoring scale, having been applied ad hoc at this gate | Gate Decision Framework v1.1, §3 |
| Evidence standard tightened — an explicit table of what does and does not count, including currency of model inputs | Gate Decision Framework v1.1, §4 |
| Standing dissent section made mandatory in every gate package | Governance Model §10; applied at the second convening, where it carried two entries |
A fourth change was considered and rejected: requiring evidence to be complete before a gate date is set. It would have removed the failure mode at §5, but it also lets any incomplete workstream postpone a gate indefinitely, which is a worse governance problem than the one it solves. The Chair's existing authority to decline to convene (§5 above) is the right control; it simply was not used.
10. Conditions for the return
The Board specified what the re-tabled package must contain. All were satisfied at the second convening on 11 June 2026, at which the package scored 3.20 and carried GO WITH CONDITIONS 4–0–1.
- Written Year 1 indications from no fewer than three distribution partners, with the aggregate stated against the Year 1 target and any shortfall disclosed rather than closed by assumption.
- Return recomputed on the capital charge current at the date of the second convening, with the change from the first package shown explicitly.
- A downside case run against the restated return.
- A dissent section, containing any material analysis that does not support the recommendation.
The first of these produced $168,000,000 of written indications against the $185,000,000 target, leaving $17,000,000 disclosed as uncommitted. The fourth produced two entries, one of them from the Distribution function against its own forecast.
11. Document control
| Version | Date | Change |
|---|---|---|
| 1.0 | 04 May 2026 | Issued. Findings, root cause, remediation plan and return date. |
| 1.1 | 11 Jun 2026 | Closed out following the second convening. Remediation statuses set to complete; cost of loop finalised at $232,000; schedule consequence (§8) added once the compressed Stage 2 window was fixed. |
Related: Stage-Gate Governance Model · Gate Decision Framework · Gate 1 Business Case Package · RAIDD Log (I-01, R-04, R-08, DEP-03, DEP-07).