# Value Proposition Optimization Package ## B2B Expense Management SaaS — Website Hero --- ## 1. Mandate, Audience, and Definition of Done **Outcome.** This optimization exists to settle which single headline (and optional supporting subheadline) replaces the current website-hero line, "Expense management made easy," for the primary conversion asset on the marketing site. The decision in scope is the website hero only — no ad headline, pitch-deck, or app-onboarding copy is authorized by this document. **Stakeholders.** Candidate drafting is owned by growth/brand marketing. Internal rubric scoring (Section 4) is performed by a marketing + product-marketing review panel informed by existing sales and customer-success input where available. Approval to run the pilot test and to publish the winning candidate to the live website hero (Section 12) sits with the marketing decision owner named at sign-off, with a mandatory legal sign-off on any candidate carrying a flagged superiority claim. **Requirements table** | ID | Requirement | Source | |---|---|---| | R1 | Deliver a website-hero headline (and optional short subheadline) to replace "Expense management made easy" | Supplied | | R2 | Must land specifically with Finance Directors at 100–500 person companies currently running expenses through spreadsheets and email approvals | Supplied | | R3 | Must differentiate against how the segment currently frames Expensify and Ramp, the two alternatives it compares this product to | Supplied | | R4 | Must never use the word "free" anywhere in candidate copy, in any form | Supplied (hard constraint) | | R5 | Must not merely re-word the current line's generic-benefit framing without first accounting for why that line was reported to be ignored | Gap | | R6 | Internal scoring must not be treated as, or presented as, market validation; broad publication is gated on a real test, not on rubric totals | Gap | **Explicit exclusions.** No claim of a free tier, trial, or "no cost" framing in any form (R4). No unqualified superiority language — "only," "best," "fastest," "cheapest" — without a stated, checkable basis and legal sign-off (Principle D). No attempt in this hero to simultaneously address the Controller/AP or IT/Security personas (see Section 5). **Solution shape.** Candidate drafting and the rubric-scoring rationale (Section 4) are judgment calls made by the marketing panel. Whether a candidate passes the hard constraint (R4), whether a claim requires legal review (Section 10), and whether the pilot test (Section 9) actually passed are checkable, not matters of internal taste. **Completion criteria** - Every rubric score in Section 4 traces to a cited reason and an assumption ID — no unjustified numbers. - The `<VALIDATE: c4-faster-claim>` item is either resolved with a checkable basis or the candidate carrying it remains permanently excluded from publication. - A small-pilot test (Section 9) has actually run and produced a real, recorded result for the top candidates against the current baseline — not a projected or assumed result. - Both decision-critical Section 7 items are either resolved or reported by name as open blockers with a stated next step. - Every action in Section 12 carries an explicit status — none is left implicitly approved. A scored shortlist alone is **not** completion. Completion requires the test in Section 9 to have actually run and the evidence in Section 13 to exist. --- ## 2. Optimizer Instructions ``` Optimize against real customer reactions, not your own preference for clever phrasing. Review any current value-prop language and past test results first. Preserve any candidate that already scores well instead of discarding it for novelty's sake. Resolve every <VALIDATE> item through real customer feedback, a real test, or current competitive research — not internal guesswork. Score every candidate against the rubric with a stated reason per score; a number with no justification is not a score. Run the test defined in Section 9, read the actual results, and revise or discard candidates accordingly. Stop defending a candidate that underperforms in testing; report a concrete blocker when progress needs budget, legal sign-off, or data outside your authority. Do not present an internally-preferred candidate as validated, invented customer reactions, fabricated test results, or a superiority claim with no checkable basis. Continue scoring and refining candidates internally while a decision-critical validation is pending. Publish the winning candidate broadly only within approved authority. Report progress using actual test data, and clearly separate internally- scored, pilot-tested, and fully validated status for each candidate. Operational defaults (how many candidates to draft, which internal reviewers score them) may be chosen and documented autonomously. Do not autonomously invent test results or claim a test was run when it was not. ``` In particular: do not assume that finance directors at 100–500 person companies rank audit-trail control and approval trust above raw speed or ease-of-use when choosing between this platform, Expensify, and Ramp — that ranking is `<VALIDATE: top-priority-ranking>`, a required real-customer validation, not a scoring default to build the whole differentiation strategy on. --- ## 3. Customer and Competitive Assumption Ledger **Claims about the customer's priority ranking** | ID | Claim | Source | Confidence | Consequence if wrong | |---|---|---|---|---| | A1 | The segment currently manages expenses through spreadsheets and manual email approval chains | Supplied | High | Any candidate built around "replacing spreadsheets/email" (C2, C3, C4) loses its central relevance hook | | A7 | Finance directors at this company size rank audit-trail control and approval trust above raw speed or ease when evaluating expense software | Unverified — `<VALIDATE: top-priority-ranking>` | Low (working hypothesis) | Invalidates the "control/trust" differentiation thesis behind C1 and the Segment Relevance scores for C1 and C3 | | A8 | The primary daily pain driving a switch off spreadsheets/email is approval-chasing delay, not reconciliation error rate or audit/compliance risk | Unverified — `<VALIDATE: primary-pain-driver>` | Low | If the real driver is compliance risk or error rate instead, C2's "retire the inbox chase" framing solves the wrong problem | **Claims about the competitive alternatives** | ID | Claim | Source | Confidence | Consequence if wrong | |---|---|---|---|---| | A3 | The segment most often compares this product to Expensify and Ramp | Supplied | High | Differentiation framing in Section 4 targets the wrong competitive set entirely | | A5 | Ramp's current public hero positioning centers on "Time is money. Save both," paired with self-reported speed and savings figures (e.g., faster book close, cost savings), as of September 2026 public sourcing | Supplied (current public sourcing) | Medium–High (self-reported marketing claims, not independently audited) | Deliberately avoiding a head-on speed claim to differentiate from Ramp could be targeting a positioning Ramp has already moved past by the time this ships | | A6 | Expensify's consistent public brand line is "the easiest way to do your expenses, travel, and corporate cards," repeated across its investor and marketing materials as of September 2026 | Supplied (current public sourcing) | High | Avoiding "ease" language to differentiate from Expensify may be unnecessary, or may be miscalibrated if Expensify has since repositioned | **Claims about what has and hasn't worked before** | ID | Claim | Source | Confidence | Consequence if wrong | |---|---|---|---|---| | A2 | The current hero line, "Expense management made easy," tested poorly and gets ignored | Supplied | High | Used only as the negative baseline candidate in Section 9's test, not as a basis for new claims | | A9 | The current line underperformed because "made easy" is generic and closely echoes Expensify's own "easiest" claim (A6), rather than for an unrelated reason (layout, CTA, page speed, etc.) | Unverified — `<VALIDATE: root-cause-current-failure>` (routine) | Low–Medium (plausible inference, not confirmed) | If the real failure mode was unrelated to wording, none of the new candidates fix the actual problem | **Hard constraint** | ID | Claim | Source | Confidence | Consequence if wrong | |---|---|---|---|---| | A4 | The product has no free tier; the word "free" cannot appear anywhere in candidate copy | Supplied (hard constraint) | N/A — a rule, not an empirical claim | N/A — enforced regardless of any other finding | The single assumption whose failure would most invalidate this rubric is **A7**. Every candidate's Segment Relevance and Differentiation score in Section 4 assumes finance directors weight control/audit-trust highly relative to speed or ease; if that ranking is wrong, the entire shortlist needs to be re-scored against whatever the real top priority turns out to be. This is the first thing Section 9's test must probe. --- ## 4. Candidate Value Propositions and Scoring Rubric ``` Rubric — score each candidate 1–5 per criterion: 1. Clarity — a 5 is understood in under 3 seconds by someone skimming, with no jargon and no ambiguous referent. 2. Differentiation vs. Expensify/Ramp — a 5 stakes out ground neither competitor's current public positioning (A5, A6) already occupies. 3. Credibility — a 5 contains no explicit or implicit claim lacking a supplied or checkable basis; nothing requires legal substantiation before it could run as written. 4. Segment Relevance — a 5 speaks directly to the named pain or priority of a finance director at a 100–500 person company still running this process on spreadsheets/email (A1), not a generic buyer. 5. Constraint Compliance — a 5 contains no banned word (A4) and no claim excluded by Section 1. ``` **Candidate 1 — Control/trust framing** ``` Expense management your finance team can finally trust. Automated approvals, real-time GL coding, and a clean audit trail — built for finance directors, not expense-report chasers. ``` **Candidate 2 — Pain-point framing** ``` Retire the spreadsheet. Retire the inbox chase. One system for expense approvals, coding, and reporting — no more version-control emails or buried approval requests. ``` **Candidate 3 — Persona-specific framing** ``` Built for finance directors who outgrew spreadsheets. Approval workflows, GL coding, and audit-ready expense records in one platform — made for teams still running this process by hand. ``` **Candidate 4 — Efficiency/speed framing** ``` Close expenses faster, without losing control. Automated approvals and audit-ready coding replace the spreadsheets and email chains slowing down your close. ``` **Scoring table** | Candidate | Clarity | Differentiation | Credibility | Segment Relevance | Constraint Compliance | Total | Rationale | |---|---|---|---|---|---|---|---| | C1 | 4 | 4 | 5 | 4 | 5 | 22 | Targets the A7 hypothesis (control/trust) directly, differentiating from Ramp's speed framing (A5) and Expensify's ease framing (A6); makes no unverified claim | | C2 | 5 | 3 | 5 | 5 | 5 | 23 | Names the exact stated pain (A1) with no ambiguity; differentiation is moderate since "eliminate manual work" is a theme both competitors also use, even if worded differently | | C3 | 4 | 4 | 4 | 5 | 5 | 22 | Strong Segment Relevance via explicit persona naming (A1); "outgrew spreadsheets" carries a mild unstated assumption about company growth stage, slightly softening Credibility | | C4 | 4 | 3 | 2 | 4 | 4 | 17 | "Faster" is an unverified superiority claim with no supplied or checkable basis — flagged `<VALIDATE: c4-faster-claim>` per Principle D; also sits closest to Ramp's own speed positioning (A5), weakening Differentiation | **Proposed for testing:** C1 and C2. Both score highly on Credibility and Constraint Compliance, and together they let the pilot test directly probe the unresolved A7 assumption (control/trust framing in C1) against the pain-point framing (C2) rather than testing two candidates that make the same bet. C3 is a viable alternate held back from this pilot round (Principle E) — it overlaps thematically with C2's spreadsheet-replacement framing and scored lower on Differentiation, so testing it alongside C2 in a small pilot would mostly split the same signal rather than add a distinct data point. C4 is **not** proposed for testing. Its Credibility score is provisional and gated on `<VALIDATE: c4-faster-claim>` (Section 7); it cannot be declared a testing candidate until that basis exists. **Evidence rule.** This scoring is an internal narrowing step, not a market verdict. No candidate above is authorized for public use on the strength of this rubric alone — public use is gated by the Section 12 approval register and confirmed by the real test in Section 9. --- ## 5. Segment and Use-Case Contracts **Contract: Primary buyer — Finance Director** - Segment description: Finance Director at a 100–500 person company, currently running expense approvals through spreadsheets and email, actively evaluating this product against Expensify and Ramp. - Core priority addressed: The unresolved A7 hypothesis (control/audit trust) and the supplied pain point A1 (manual spreadsheet/email workflow). - Candidates targeting it: C1, C2 (proposed for testing); C3 (held alternate). - Channels where this framing applies: Website hero only, per the Section 1 mandate. - Owner: Marketing decision owner named at Section 12 sign-off. **Use-cases explicitly NOT addressed** - **Controller/AP Manager** (the frequent day-to-day system user, often an influencer rather than the buyer): not addressed here because Section 1 scopes this decision to the finance-director-facing hero; a single 5–12 word headline attempting to speak to both the buyer and the day-to-day user risks diluting Segment Relevance (Principle A) for the primary audience this decision is meant to serve. A persona-specific asset for this role is a separate, future scope decision, not an oversight here. - **IT/Security stakeholder** (evaluates data handling and integration risk): not addressed because a marketing hero headline is not the vehicle for security/trust claims; those require a vetted, separately-owned asset and fall outside Section 1's requirements. **Internal vs. external claim separation.** The assumption IDs, confidence ratings, and rubric rationale in Sections 3–4 are internal scoring rationale only and must never reach a customer-facing surface. Only the finalized headline/subheadline text — after it clears Section 12 approval — is authorized to appear publicly. --- ## 6. Selection Rationale and Usage Map C2 ("Retire the spreadsheet. Retire the inbox chase.") scored highest overall (23/25), driven by maximal Clarity and Segment Relevance from naming the supplied pain point (A1) with no ambiguity. C1 (22/25) is the closest alternative and is proposed alongside it specifically because it tests the opposite bet — that control/trust (A7), not pain-relief, is what actually moves this buyer. C3, also at 22/25, would have to outperform on Differentiation or Credibility to overtake either — currently it does not, since its "outgrew spreadsheets" framing carries a softer, unstated growth-stage assumption that neither C1 nor C2 needs. **Proposed lead candidate: C2.** It best satisfies Requirement R2 (landing with the stated segment) by directly naming their current process (A1) rather than a generic benefit — explicitly pending Section 9 confirmation against C1 and the current baseline. **Usage map** | Asset | Use-case contract served | Owner | |---|---|---| | Website hero headline + subheadline | Primary buyer — Finance Director (Section 5) | Marketing decision owner | No other asset is in scope. Flowing the winning candidate to an ad headline, pitch-deck slide, or sales opener would require a separate optimization against those decision contexts' own length/tone constraints — Section 1 does not authorize that here. **Directions deliberately not pursued.** A head-on speed/efficiency claim (demonstrated in C4) was not pursued as a primary direction because it lacks a checkable basis (`<VALIDATE: c4-faster-claim>`) and sits closest to Ramp's existing public positioning (A5), the opposite of differentiation. An "ease" reframe closer to Expensify's own line (A6) was also not pursued as a primary direction, since it risks repeating whatever actually caused the current line's reported failure (A9). No claim beyond what is scored in Section 4 is authorized for this hero. --- ## 7. Validation and Research Requirements **`top-priority-ranking`** (decision-critical) - What must be established: Whether finance directors at 100–500 person companies weight audit-trail control/trust above speed or ease when choosing expense software. - How to establish it: Run a real customer test (Section 9). - Dependent tasks/actions: Blocks declaring either C1 or C2 the confirmed winner; blocks Section 12 Action AR2 (publish to hero). - Status until resolved: Both C1 and C2 remain "internally scored, pilot-proposed" only — neither is the winner until this resolves. **`primary-pain-driver`** (decision-critical) - What must be established: Whether approval-chasing delay (vs. reconciliation error rate or compliance risk) is the primary driver behind a switch off spreadsheets/email. - How to establish it: Run a real customer test (Section 9), ideally the same instrument used for `top-priority-ranking`. - Dependent tasks/actions: Blocks confirming C2 specifically as the winner; blocks AR2. - Status until resolved: C2 remains "internally scored, pilot-proposed" only. **`root-cause-current-failure`** (routine) - What must be established: Why "Expense management made easy" actually underperformed (wording vs. an unrelated page factor). - How to establish it: Check current internal data first — past A/B test notes, heatmap or scroll data, sales/CS anecdotes about the current hero — before commissioning new research. - Dependent tasks/actions: Informs whether the wording-focused hypothesis behind this entire optimization is the correct lever at all. **`c4-faster-claim`** (decision-critical) - What must be established: A checkable basis (e.g., a measured reduction in close-time or approval-cycle time across a real cohort) for any "faster" claim before it can appear in public copy. - How to establish it: Run a real customer test — a benchmark comparison is a decision-critical claim, not a routine competitive check. - Dependent tasks/actions: Blocks C4 from re-entering the candidate pool for testing or publication. - Status until resolved: C4 remains excluded from Section 9's test and Section 12's publication actions. **`current-competitor-positioning`** (routine) - What must be established: That A5 and A6 still accurately describe Expensify's and Ramp's public positioning at the time of publication. - How to establish it: Check current public competitive/market data before finalizing — this was checked via public sourcing as of September 2026 for this draft, but competitor homepages change and must be re-checked immediately before Section 12's publish action executes. - Dependent tasks/actions: Does not block internal scoring; blocks only the final publish action if it has gone stale. **Existing-data-first note.** Before commissioning any new customer research for the items above, check past message or A/B test results, sales call notes, and support/CS feedback already held internally — any customer-priority claim used to resolve a `<VALIDATE>` item must be sourced from real data, never invented for plausibility, and any competitor claim must be re-verified from current public material rather than recalled from memory. --- ## 8. Optimization Sequence and Requirement Traceability 1. Draft candidates (Section 4) — satisfies R1, R2, R3. *Complete.* 2. Internal rubric scoring (Section 4) — satisfies R6 (internal-only status maintained). *Complete.* 3. Deterministic legal/brand check against the hard constraint (A4/R4) and Section 1 exclusions, applied to every candidate including the held alternates C3 and C4. Depends on: 1. 4. Small-pilot test (Section 9) comparing C1, C2, and the current baseline. Depends on: 3. Addresses: A7, A8. 5. Route `root-cause-current-failure` and `current-competitor-positioning` checks through existing internal data and current public sourcing, in parallel with 4. Depends on: none (can run immediately). 6. Confirm the winning candidate only once `top-priority-ranking` and `primary-pain-driver` resolve with an actual pilot result. Depends on: 4. 7. Report any unresolved decision-critical item as a named blocker, with a concrete next step, before any candidate is rolled out beyond the pilot. Depends on: 6. **Traceability table** | Requirement/Assumption ID | Candidate/Behavior | Owner | Acceptance Check | |---|---|---|---| | R1 | C1, C2 delivered as hero headline + subheadline pairs | Marketing copywriter | Both candidates exist in the required format (Section 4) | | R2 | C1, C2 both name the stated segment/pain | Marketing panel | Segment Relevance ≥ 4 in Section 4 scoring, with cited rationale | | R3 | C1, C2 avoid A5/A6 competitor framing | Marketing panel | Differentiation score justified against A5/A6 in Section 4 | | R4 | All candidates, including held C3/C4 | Legal/brand reviewer | Deterministic scan confirms no instance of "free" in any candidate | | A7 | C1 vs. C2 head-to-head | Whoever runs the pilot test | Small-pilot-test result recorded (Section 9) | | A8 | C2 specifically | Whoever runs the pilot test | Small-pilot-test result recorded (Section 9) | | c4-faster-claim | C4 (held, not in pilot) | Legal reviewer | Benchmark data exists and is signed off before C4 re-enters candidate pool | --- ## 9. Test Design and Readiness Checks Any audience size or sample below is a small-scale allocation for an initial read, not a statistically representative measurement of broad-market performance. Every "expected result" is a hypothesis, not a finding, until the test actually runs and produces a recorded outcome. Because `top-priority-ranking` and `primary-pain-driver` need a stand-in before a broad-rollout decision can be made, the small-pilot test below doubles as that stand-in — passing it establishes only that one framing beat the others at small scale on the stated metric; it does **not** establish the underlying priority ranking for the whole segment, which remains unverified until an expanded test or full rollout confirms it. **Readiness table** | Assumption/Requirement | Test level | Audience/method | Expected result | Pass condition | Prerequisites | |---|---|---|---|---|---| | R1–R3 (candidate quality) | internal-review | Marketing + product-marketing panel scoring (Section 4) | Rubric totals with cited rationale | Every score has a stated reason | None — complete | | A7 / A8 (top-priority, pain-driver) | small-pilot-test | Head-to-head message test (e.g., a message-testing panel or landing-page A/B) of C1 vs. C2 vs. current baseline, small sample of finance directors at 100–500 person companies | One candidate shows a clearly higher engagement/click-through signal than the current baseline | The leading candidate beats the current baseline by a pre-agreed margin at small scale | R4 legal/brand scan complete (task 3, Section 8) | | R4 (hard constraint) | internal-review | Deterministic text scan of every candidate, including held C3/C4 | Zero instances of "free" | Zero instances found | None | | c4-faster-claim | small-pilot-test (held, not yet scheduled) | If revisited: a real before/after benchmark on close-time or approval-cycle time across a pilot customer cohort | NOT RUN | A measured, reproducible reduction, reviewed by legal | Legal review scheduled first | | Legal review of any superiority claim | internal-review | Legal counsel review of C4 specifically | NOT RUN until C4 is reconsidered | Counsel sign-off on record | c4-faster-claim benchmark exists | | Full-market performance | full-rollout | All website visitors to the hero page | NOT RUN until Section 12 approvals granted | Sustained lift over baseline across a full traffic period | Small-pilot-test PASS + all Section 12 actions approved | | Fallback: all candidates underperform baseline | small-pilot-test outcome check | Same instrument as A7/A8 test | NOT RUN | If neither C1 nor C2 beats the current baseline, retain the current line and re-open drafting with fresh input from `root-cause-current-failure` and `primary-pain-driver` findings | Small-pilot-test data collected | A small-pilot-test PASS proves only that one candidate outperformed the others at small scale under test conditions. It does not prove the winning framing holds up across the full range of the segment's buying situations — that is what full-rollout data, gated by Section 12, is for. --- ## 10. Legal, Compliance, and Brand Boundaries - **HIGH — Publishing while decision-critical items are unresolved.** Publishing any candidate to the live hero broadly while `top-priority-ranking` or `primary-pain-driver` remain unresolved is not permitted. Backstop: Section 12's Action AR2 is held at status BLOCKED BY VALIDATION until both resolve. - **MEDIUM — Unverified superiority claim.** C4's "faster" claim has no supplied or checkable basis and creates legal exposure if published. Backstop: mandatory legal-review gate (Section 9) before any use; C4 is excluded from the pilot and from publication until that gate is cleared. - **Never appear in a public candidate:** the word "free" in any form (A4, hard constraint); any unqualified superiority word — "only," "best," "fastest," "cheapest" — without a stated, checkable basis and legal sign-off. - **Brand-contradiction check.** Before shipping, the winning candidate must be checked against the existing approved brand-messaging library by the brand owner; that library is not supplied in this document and this check is a named, unassigned task rather than an assumed pass. - **Regulated-claim screen.** None of the four candidates make a health or financial-advice claim. The audit-trail language in C1 and C3 makes an implicit product-accuracy claim (that the product actually produces an audit-ready record) — this requires a product/legal accuracy check separate from message testing, confirming the feature claim is true today, not just that the wording tested well. - **Genuinely unassessable until Section 7 resolves:** whether the chosen framing (control/trust vs. pain-relief) will outperform the current baseline under real market conditions at scale. That risk cannot be reduced further by more internal review — only by the pending decision-critical validations. --- ## 11. Rollout, Verification, and Correction Plan **Runnable now (internal drafting and scoring):** Candidate drafting (Section 4), rubric scoring, the deterministic R4 legal/brand scan, and the existing-data-first checks for `root-cause-current-failure` and `current-competitor-positioning`. **Blocked on testing:** The small-pilot test (Section 9) must run and produce a recorded result before any candidate is confirmed as the winner. **Approved pilot test execution:** Gated on Section 12 Action AR1's approval; status recorded there, not assumed here. **Expanded-test checks:** Before any broad rollout, an expanded test at higher traffic/sample would need to confirm the small-pilot result holds; this step is NOT scheduled until the small-pilot-test passes. **Approved full rollout requires ALL of:** - The small-pilot test (Section 9) passing at its stated level, with a recorded result. - Every action in Section 12 marked complete, not merely listed. - Both decision-critical Section 7 items (`top-priority-ranking`, `primary-pain-driver`) resolved by real test data — not internally stubbed or assumed. **Post-rollout verification (bounded, low-cost, real check):** Within the first usage window after any hero update, confirm the new headline is actually rendering correctly on the live site (not just approved in a design file), and confirm no excluded claim from Section 1 — the word "free," or an unreviewed "faster"/"best" claim — was accidentally shipped alongside it. **Correction plan.** Before assuming an update "reaches everywhere," confirm which channels currently carry the old line "Expense management made easy" (the live website hero is confirmed in scope here; any other channel carrying that line, such as a cached landing page or paid ad using the same copy, is not confirmed and must be checked separately). Updating this document does not undo assets already published elsewhere. If an unauthorized claim is found live, containment means flagging it and re-issuing the corrected asset — never assuming the fix propagated automatically. --- ## 12. Publication and Test Approval Register | Action ID | Purpose | Candidate/Channel/Scope | Expected Effect | Trigger Condition | Status | Approved by / Date | |---|---|---|---|---|---|---| | AR1 | Run the small-pilot test | C1 vs. C2 vs. current baseline; website-hero message test | A recorded engagement signal for each candidate | R4 legal/brand scan complete | READY FOR HUMAN REVIEW | ____ / ____ | | AR2 | Publish the winning candidate to the live website hero | Winning candidate (C1 or C2); website hero, sole channel per Section 1 | Live hero copy updated | AR1 passes AND `top-priority-ranking` AND `primary-pain-driver` resolved | BLOCKED BY VALIDATION (`top-priority-ranking`, `primary-pain-driver`) | ____ / ____ | | AR3 | Expand the test to a larger traffic allocation | Winning candidate vs. baseline; website hero | A broader-scale confirmation of the small-pilot result | AR1 complete with a passing result | Blocked on AR1 | ____ / ____ | | AR4 | Retire the current hero line, "Expense management made easy" | Website hero | Old line no longer live | AR2 complete | BLOCKED BY VALIDATION (same items as AR2) | ____ / ____ | | AR5 | Legal review of any future superiority claim (applies if C4 is ever reconsidered) | C4, held out of current scope | Counsel sign-off or rejection on record | `c4-faster-claim` benchmark data exists | READY FOR HUMAN REVIEW | ____ / ____ | Approving one action above never implies approval of another. Listing AR4 here documents that retiring the current line is a foreseeable next step — it does not authorize executing it now. --- ## 13. Completion Evidence and Handoff A final report on this optimization must contain: - The confirmed winning candidate (C1 or C2, or neither if the fallback in Section 9 triggers) and exactly where it now appears (the specific website-hero asset/URL) — only once actually shipped, not the design-file version. - Every Section 9 check itemized as PASS / FAIL / NOT RUN / BLOCKED, each with one line of real evidence (e.g., the actual small-pilot-test result, not a projected one). - Internally-scored status (Section 4 rubric) and test-confirmed status (Section 9) reported separately, with an explicit note that a high internal score is never presented as evidence of real-world performance. - The actual pilot sample size and result, only if the test genuinely ran; otherwise stated plainly as "NOT YET TESTED." - The status of every unresolved Section 7 item, named individually: `top-priority-ranking`, `primary-pain-driver`, `root-cause-current-failure`, `c4-faster-claim`, `current-competitor-positioning`. - Any residual limitation still true at handoff (e.g., the small-pilot-test result, even if a PASS, does not establish full-market performance). - Exact next steps for any outstanding validation or Section 12 approval item, including who owns it. Fabricating a test result, a customer reaction, or an approval that was not actually given is prohibited without exception. A high-scoring candidate on paper, from Section 4 alone, is not evidence that it converts in market.