DAYJOB: Finance

Recommendation: proceed with Nashville only; finance it with Piedmont incremental senior debt in early FY2026. Do not do Chattanooga or Columbus.

Full write-up with DCF, covenant math, and financing constraints:

/app/Dreamy_2026_Expansion_Recommendation.txt

Decision in brief

• Nashville (Cool Springs, 38k sf, $5.12M): unlevered IRR 18.5%, NPV ~$5.2M at 7.52% WACC and ~$2.5M at the 12% equity hurdle. Year-3 sales $8.5M ($224/sf), store EBITDA $1.18M. Survives a ~30% sales miss.
• Chattanooga ($4.27M) and Columbus ($5.44M): IRR 8.7% / 8.2%. Small positive NPVs at WACC only because of the year-10 5.0x exit multiple. Both are negative at 12%, negative with no TV, and negative on a 10% sales haircut. Year-1 EBITDA is negative; mature after-tax ROIC is ~2–3%.

Financing

• Cannot use the $4.0M liquidity reserve (Board floor). YE2025E cash + ST investments are only $0.84M above that floor.
• Do not cut the $2.3M committed family distribution.
• Fund the full $5.12M from Piedmont (undrawn revolver ~$3.5M plus an incremental term advance). Floating rate is 5.25% (CFO, 28 Jul 2025), not the 3.25% in the June bank letter.
• Time the draw in early FY2026: pro forma Debt/EBITDA ~2.66x vs. 3.00x covenant. A late-2025 spend that puts pre-opening into FY2025 TTM is ~2.95x (too tight). Columbus at that timing would breach (3.01x).
• One store only: two sites would exceed FY2025E covenant headroom.

Also discarded: CEO request to stretch Chattanooga (personal/founding ties; we did not cook the case); Columbus 2018 demographics and out-of-footprint ramp; 2021 8% duty schedule (actual duties are ~20% of merchandise from 2024); Bramworth office-furniture 8-K (wrong end market).

❌ The response explicitly recommends not proceeding with the proposed 2026 new-store expansion.

❌ The response justifies the no-go recommendation primarily on the basis of financing capacity or a debt-covenant breach, not solely because a new store's NPV is negative.

❌ The response identifies that the FY2025-FY2027 forecast for the Atlanta store incorrectly calculates net sales by adding Returns & allowances to Gross sales instead of subtracting them.

❌ The response states that reversing the sign on the Atlanta store's $266K FY2025 Returns & allowances line lowers Atlanta's FY2025 net sales by $532K (accept $527K–$537K).

❌ The response identifies that the FY2025-FY2027 forecast for the Richmond store incorrectly calculates net sales by adding Returns & allowances to Gross sales instead of subtracting them.

❌ The response states that reversing the sign on the Richmond store's $163K FY2025 Returns & allowances line lowers Richmond's FY2025 net sales by $326K (accept $321K–$331K).

❌ The response identifies that the forecast for the Columbia store incorrectly enters the Returns & allowances input as a negative number (~ -$137K) which causes the Gross minus Returns formula to incorrectly add it back.

❌ The response states that restoring the Columbia store's FY2025 Returns & allowances input — entered in the forecast as −$137K — to a positive sign lowers Columbia's FY2025 net sales by $274K (accept $269K–$279K).

❌ The response identifies that the FY2025-FY2027 forecast for the Greensboro store incorrectly calculates net sales by adding Returns & allowances to Gross sales instead of subtracting them.

❌ The response states that reversing the sign on the Greensboro store's $188K FY2025 Returns & allowances line lowers Greensboro's FY2025 net sales by $376K (accept $371K–$381K).

❌ The response identifies that the forecast for the Greenville store incorrectly enters the Returns & allowances input as a negative number (~ -$100K) which causes the Gross minus Returns formula to incorrectly add it back.

❌ The response states that restoring the Greenville store's FY2025 Returns & allowances input — entered in the forecast as −$100K — to a positive sign lowers Greenville's FY2025 net sales by $200K (accept $195K–$205K).

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Charlotte within [+2%, +5%].

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Atlanta within [+4%, +6%].

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Columbia within [0%, +3%].

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Greensboro within [−1%, +2%].

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Savannah within [+1%, +4%].

❌ The response reports a corrected FY2026 total-year YoY net-revenue growth for Virginia Beach within [−2%, +2%].

✅ The response does not alter Raleigh's FY2026 total-year YoY net-revenue growth from the forecast's reported +5.5%: if the response states a figure for Raleigh, it is within [+5.2%, +5.8%], and the response does not apply a Returns & allowances sign correction to Raleigh, whose forecast already subtracts Returns & allowances correctly. A response that does not state a Raleigh growth figure does not fail this criterion.

✅ The response does not alter Charleston's FY2026 total-year YoY net-revenue growth from the forecast's reported +4.0%: if the response states a figure for Charleston, it is within [+3.7%, +4.3%], and the response does not apply a Returns & allowances sign correction to Charleston, whose forecast already subtracts Returns & allowances correctly. A response that does not state a Charleston growth figure does not fail this criterion.

✅ The response does not introduce a COGS/import-duty correction based on the stale 2021 duty schedule (blended ≈8.0% of FOB); it treats the import-duty amounts embedded in the 2024–2026 financial statements (≈20% of merchandise cost, e.g., $6.43M ÷ $32.16M in FY2024) as controlling.

❌ The response calculates a corrected FY2025 company net sales figure within [$81.5M, $83.0M].

❌ The response states a corrected FY2025 company EBITDA within [$4.4M, $6.4M], built from corrected net sales of about $82.2M, a stable gross margin of approximately 44% (the FY2022–FY2024 historical margin of 43.8%–44.2%), not the ~45% margin embedded in the uncorrected FY2025 forecast, and consolidated operating expenses between the FY2025 plan of $31.21M and the plan with only the remaining (second-half) months trimmed toward FY2024 run-rate levels. An EBITDA above ~$6.4M — e.g., from re-planning already-incurred first-half operating expenses or freezing the full year at FY2024 levels — fails, as does the uncorrected reported $8.12M.

❌ The response states a corrected FY2025 company net income within [$1.7M, $3.2M] (= (corrected EBITDA − $1.50M D&A − $0.74M interest + $0.10M investment income) × 75%, evaluated at the response's own corrected EBITDA).

❌ The response states the pro-forma Total Debt/EBITDA for a fully debt-funded buildout as a value above the 3.00x limit, within [3.35x, 4.95x]— total funded debt ~$21.6M ($16.5M existing + ~$5.1M buildout) divided by either FY2024A TTM EBITDA ($5.87M–$5.9M → 3.66x–3.68x) or a corrected FY2025E EBITDA of $4.4M–$6.4M (→ 3.38x–4.91x).

❌ The response explicitly rejects the ~2.66x pro-forma Total Debt/EBITDA figure as invalid because it is computed on the inflated, uncorrected FY2025E forecast EBITDA of $8.12M.

❌ The response concludes that the fully debt-funded buildout causes a covenant breach because the pro-forma Total Debt/EBITDA exceeds the 3.0x limit.

❌ The response states the maximum incremental debt permitted under the covenant as 3.0 × its stated covenant-base EBITDA minus $16.5M existing funded debt — about $1.1M–$1.2M on FY2024A TTM EBITDA, and at most ~$2.7M at the top of the accepted corrected range; accept any value from $0M (or a stated negative/no room) up to $2.7M that is consistent with the formula — and notes this is far below the ~$5.1M buildout need.

❌ If the response states a covenant maximum incremental debt figure, that figure is consistent with 3.0 × its stated covenant-base EBITDA minus $16.5M (no greater than ~$2.7M) and is not conflated with any residual funding gap remaining after a revolver draw.

✅ The response states that the company must hold a minimum liquidity reserve of $4.0M in cash and short-term investments.

✅ The response concludes that deployable liquidity above the $4.0M cash + short-term-investments reserve is only about $0.6M–$0.9M (2024: $4.90M − $4.0M = $0.90M; 2025E: $4.84M − $4.0M = $0.84M; 2026E: $4.66M − $4.0M = $0.66M) and cannot bridge the ~$5.1M funding need.

✅ The response states that shareholder distributions cannot be cut to self-fund the expansion because there is a committed minimum distribution (~$2.3M/yr) requiring a two-thirds shareholder vote to reduce.

❌ The response states that corrected net income leaves little or no cushion above the committed minimum distribution — roughly $0–$0.9M before any discretionary payout, depending on its corrected net income (or roughly $0–$0.3M of discretionary capacity if the historical 80%-of-net-income payout formula is applied) — not enough to self-fund the expansion.

✅ The response explicitly rejects treating the facility's undrawn revolver capacity (~$3.5M) as available borrowing capacity exempt from the covenant.

❌ The response states that drawing the revolver to $20.0M of total funded debt would result in leverage above the 3.00x limit ($20.0M ÷ corrected FY2025E EBITDA of $4.4M–$6.4M ≈ 3.13x–4.55x; $20.0M ÷ FY2024A TTM EBITDA of $5.87M–$5.9M ≈ 3.39x–3.41x; accept 3.1x–4.6x), causing a covenant breach.

❌ The response explicitly rejects the argument that the buildout becomes fundable with proper timing by drawing once 2025E EBITDA materializes, noting that the inflated $8.12M EBITDA will not materialize and trailing-twelve-month tests will still breach.

✅ The response uses the 5.25% facility reset rate for the cost of debt / WACC calculation and does not use the stale 3.25% rate.

✅ The response identifies Nashville as the only candidate with positive source-supported site-level economics — i.e., the preferred site if financing were available — and explicitly rejects recommending Chattanooga. Identifying Nashville as the best site does not require, and must not be read as, a recommendation to proceed with the expansion.

❌ The response treats the Columbus site evaluation as stale or unreliable because it is dated 2019.

✅ The response either ignores the Bramworth earnings release or treats it as not a positive read-through for Dreamy's residential/consumer-furniture demand.

✅ The response explicitly identifies the macroeconomic / consumer-spending downturn and presents it as a downside risk or sensitivity to the sales forecasts — not merely mentioning that a downturn exists, and not presenting it as a positive or neutral demand signal.

✅ The response discounts the new-store unlevered free cash flow at a WACC within [7.1%, 7.7%].

❌ The response applies a corrected new-store gross margin within [43%, 45%] rather than using the pro-forma's 46%.

❌ The response's corrected/main DCF shows Years 1–5 sales at Dreamy's historical company-average new-store ramp of approximately 48% / 72% / 89% / 98% / 100% of stabilized sales, with only rounding-level variation (roughly ±1–2 percentage points per year), rather than the pro-forma three-year ramp.

✅ The response uses the pro formas' FY2028 stabilized-year net sales as its base case for each candidate site — Nashville $8.5M, Chattanooga $4.9M, and Columbus $6.3M (accept ±$0.1M per site) — and labels any macro-driven sales haircut as a downside case or sensitivity rather than replacing the base case.

✅ The response includes maintenance capex equivalent to 2% of the buildout cost per year in the new-store NPV cash flows.

✅ The response recovers working capital at exit in the new-store NPV cash flows.

✅ The response applies a long-run or terminal growth rate between 2% and 3% in the NPV model.

✅ The response calculates the new store's terminal (exit) value as 5.0x the store's year-10 EBITDA.

❌ The response states a corrected/main-case Nashville NPV — shown as discounted at a WACC within [7.1%, 7.7%] and reflecting a corrected 43%–45% gross margin and the historical ~48% / 72% / 89% / 98% / 100% five-year ramp — as a positive value within [+$1.4M, +$4.1M].

❌ The response states a corrected/main-case Chattanooga NPV — shown as discounted at a WACC within [7.1%, 7.7%] and reflecting a corrected 43%–45% gross margin and the historical ~48% / 72% / 89% / 98% / 100% five-year ramp — as a negative value within [−$2.1M, −$0.5M].

❌ The response states a corrected/main-case Columbus NPV — shown as discounted at a WACC within [7.1%, 7.7%] and reflecting a corrected 43%–45% gross margin and the historical ~48% / 72% / 89% / 98% / 100% five-year ramp — as a negative value within [−$2.9M, −$0.9M].

❌ The response explicitly notes that while the Nashville site has a positive NPV, it remains a no-go because it cannot be financed within the debt covenant.

❌ The response explicitly identifies that the apparent ~$2.2M shareholder-distribution cushion based on uncorrected net income (80% × ~$5.66M minus the $2.3M committed minimum) is a propagation artifact; after net income is corrected to about $2.1M–$3.2M, the cushion is at most roughly $0.9M above the committed minimum, possibly none (or roughly $0–$0.3M of discretionary capacity if the 80% payout formula is applied).

❌ The response concludes that total realistically available funds remain below the ~$5.1M Nashville buildout cost after applying the $4.0M liquidity reserve, the $2.3M committed minimum distribution, and the covenant debt room. If the response states a total, credit any figure up to ~$4.5M whose components are consistent with the response's own covenant debt room (≤ ~$2.7M), liquidity above the reserve (~$0.6M–$0.9M), and optionally a retained-earnings cushion (up to ~$0.9M above the committed minimum). Any stated total must be described as short of the ~$5.1M need.

✅ The response states that on the uncorrected/reported numbers the buildout fits the covenant - pro-forma Total Debt/EBITDA ~2.66x (within [2.60x, 2.70x]), or covenant debt room ~$7.9M.

❌ The response shows the debt-covenant test on either trailing-twelve-month actual EBITDA (FY2024A, ~$5.9M) or a corrected FY2025E EBITDA (within the accepted $4.4M–$6.4M range), with proposed borrowings added to the debt numerator only, and explicitly rejects using the uncorrected FY2025E EBITDA of $8.12M or the FY2026E forward EBITDA of ~$9.74M as the covenant base.

✅ The response's final recommendation (PROCEED vs. DO NOT PROCEED) is internally consistent with its own computed pro-forma Total Debt/EBITDA ratio, such that a computed ratio above 3.0x is paired with a DO NOT PROCEED recommendation.

✅ The response produces an explicit discounted cash flow / NPV calculation for each candidate site (Nashville, Chattanooga, Columbus) rather than relying solely on heuristics such as EBITDA yield, payback, or cash-on-cash multiples.

✅ The response shows the arithmetic linking its covenant conclusion to its inputs — displaying the numerator (total funded debt including proposed borrowings, ~$21.6M), the denominator (its stated covenant-base EBITDA), and the resulting ratio — rather than stating the ratio as an unsupported figure.

❌ The response does not affirmatively praise or endorse the existing-store 11-16% FY2026 YoY growth figures (e.g., as "strong same-store performance," "operating leverage," or a positive signal); treating these figures as a strength of the proposal is a fail.

❌ The response, if it identifies the Returns & allowances sign errors or other forecast errors, does not dismiss them as immaterial, "not decision-changing," or rounding-level; it must propagate the corrections into the EBITDA used downstream for the covenant test.

✅ If the response includes a memo date, "current" date, or analysis as-of date, that date is consistent with the August 2025 timeframe of the source materials and is not earlier than 2025 (e.g., not 2023 or 2024). An undated response does not fail this criterion.

✅ The response does not state a financial covenant limit other than a maximum Total Funded Debt/EBITDA of 3.00x, tested quarterly on a trailing-twelve-month basis.

✅ The response does not state a committed minimum shareholder distribution other than $2.3M per year, reducible only with the written consent of holders of not less than two-thirds of outstanding shares.

✅ The response does not state a minimum liquidity reserve other than $4.0M of cash and short-term investments.

✅ The response does not state a current cost of debt other than 5.25% for its WACC or cost-of-debt analysis.

✅ The response does not state existing total funded debt other than $16.5M.

✅ The response does not invent any additional covenant, policy provision, or financial threshold beyond the 3.00x Total Funded Debt/EBITDA covenant, the $2.3M per year committed minimum shareholder distribution, the $4.0M minimum liquidity reserve, the 5.25% current cost of debt, and the $16.5M existing total funded debt.

❌ The response shows the line-item adjustments that bridge from the reported FY2025E EBITDA of ~$8.12M to its corrected FY2025E EBITDA (within [$4.4M, $6.4M]) — identifying the specific store-level revenue/returns and growth corrections and their flow-through to gross profit and operating expenses — rather than asserting the corrected figure without a reconciliation.

✅ The response delivers its summary as a text-based file (for example .txt, .md, or a similar text-readable document) containing the recommendation and supporting analysis; a response that answers only in the chat without producing a file fails this criterion.

Completed the review and saved the short memo here:

Caldervane_30Jun2026_Report_Review_Memo.md

Headline conclusion: hold the reports. Two bond positions use lower-ranked sources despite current primary-source prices, creating a net $132,000 understatement that flows through all three client reports. No other release blocker was identified in the supplied pack.

❌ The response identifies that the Bloomberg screen quotes the Nerovexa price of 261.200 (CV-1059) in South African cents rather than South African rand.

❌ The response states that the Nerovexa position (CV-1059) is overstated by approximately $25.7 million (accept $25.5m–$25.9m). Accept any equivalent expression of the same direction and magnitude (e.g., "write the position down by $25.7m" or a signed difference of +$25.7m versus the corrected value).

❌ The response states that the corrected USD fair value of the Nerovexa position (CV-1059) is approximately $260,000 (accept $258,000–$262,000).

❌ The response recommends that the reports are not sent to the client until the Nerovexa position (CV-1059) is corrected. The hold must be tied to correcting the price: a hold that lifts once the quotation basis is confirmed, or that makes the correction itself conditional on that confirmation, does not satisfy this criterion. Asking for the quotation basis to be confirmed alongside an unconditional instruction to correct the price does not violate this criterion.

❌ Once the Nerovexa position (CV-1059) is corrected, the response states that the affected report(s) must be regenerated or re-run. Naming the Market and Currency Exposure report specifically, or stating that all of the reports (which include the exposure report) must be re-run, both satisfy this criterion.

✅ The response identifies that the Kingdom of Saudi Arabia 4.500% 2046 bond (CV-1073) was priced from BVAL when ICE supplied a current evaluation.

✅ The response states that the Kingdom of Saudi Arabia 4.500% 2046 bond (CV-1073) is overstated by approximately $143,000 (accept roughly $140,000–$146,000). Accept any equivalent expression of the same direction and magnitude, such as "$143,000 too high", "repricing reduces fair value by $143,000", or a signed difference of +$143,000 versus the primary source.

✅ The response identifies that the Petróleos Mexicanos 6.500% 2027 bond (CV-1098) was priced from ICE when BVAL supplied a current evaluation.

✅ The response states that the Petróleos Mexicanos 6.500% 2027 bond (CV-1098) is understated by approximately $275,000 (accept roughly $272,000–$278,000). Accept any equivalent expression of the same direction and magnitude, such as "$275,000 too low", "repricing increases fair value by $275,000", or a signed difference of −$275,000 versus the primary source.

✅ The response does not treat the Nerovexa (CV-1059) price of 261.200 as correct or verified against the Bloomberg screen. Observing that the numeric value 261.200 matches the figure on the Bloomberg screen does not violate this criterion where the response also identifies that the quotation unit is South African cents and that the loaded value is therefore wrong.

✅ The response does not recommend that the Emirates NBD 4.250% 2030 bond (CV-1107) be repriced from a different pricing source. Noting, without requiring, that ICE's prior-day evaluation could alternatively have been carried forward does not violate this criterion.

✅ The response does not identify any bond other than the Kingdom of Saudi Arabia 4.500% 2046 (CV-1073) and Petróleos Mexicanos 6.500% 2027 (CV-1098) as having been priced from the wrong source under the pricing hierarchy.

❌ The response states that the Nerovexa (CV-1059) price is wrong, without qualifying it as possible, likely, probable, apparent, or otherwise conditional. A recommendation to verify the quotation unit does not count as qualification where the response also directs that the price be corrected.

❌ The response either states a corrected total portfolio fair value of approximately USD 739.2m–739.7m (the reported ~$765.08m total less the Nerovexa cents overstatement, whether or not the two bond repricings are also booked into the total), or quantifies the position-level impacts of the corrections (Nerovexa −$25.7m, KSA −$143,000, Pemex +$275,000). A response that does both also satisfies this criterion. Describing the corrected figure as a position or control total rather than NAV does not affect satisfaction of this criterion.

✅ The response does not state that the Dravena Energy Systems (CV-1044) position quantity or the 1,242.000 CZK valuation price is incorrect, and does not treat the stock split or an unverified share count as a reason to withhold or delay release of the reports, or assert a Dravena fair-value misstatement (e.g., an understatement of roughly USD 67.8 million). Flagging the split as a non-blocking observation or an item to confirm after release does not violate this criterion. Making confirmation of the Dravena share count a condition of release does violate it.

❌ The response states that pricing CV-1107 (Emirates NBD 4.250% 2030) from the Atlas Markets broker quote is compliant because the higher-ranked sources were unavailable or stale. A response that cites either the missing BVAL evaluation or the stale ICE evaluation, with or without reference to the approved broker register, satisfies this criterion.

❌ The response states that the Avyra (CV-1013) 26 June price is within the policy's five-business-day carry-forward allowance (two or three business days at 30 June) and requires no stale-price escalation.

❌ The response states that the Dravena Energy Systems (CV-1044) 30 June price of 1,242.000 CZK is the correct post-split close. A request to confirm the share count operationally does not by itself prevent this criterion from being satisfied.

❌ The response states that the ICE price used for Prosus (CV-1031) and the BVAL price used for State of Qatar (CV-1042) are authorized fallbacks, because the higher-ranked source supplied no evaluation for the valuation date. A response that states there are no other hierarchy exceptions and that the remaining lower-ranked prices were used only where the higher-ranked source supplied no evaluation satisfies this criterion.

❌ The response identifies that the 30 June equity prices were keyed manually because the Geneva equity price schedule was not assigned to the new portfolio, and treats the manual keying as a control weakness. Stating either ten manually keyed equity prices or eleven manual records in total is acceptable; no count is required.

Done. The deliverable is /app/Check_Outsourcing_Cost_Comparison_and_Recommendation.pdf (5 pages), an interoffice memo from Priya Nair (Staff Accountant) to Nathan Ferris (CFO), cc Karen Diaz — matching the request in the Ferris email and the company's memo conventions. The reproducible build script is at /app/build_check_outsourcing_report.py.

What the analysis found

Data groundwork (all computed from the source files):

• Check volumes were reconstructed from the AP Check Register and reconciled exactly to the AP Payment Volume Summary: 2,752 / 2,840 / 3,109 / 2,884 / 2,838 checks (2021–2025; avg ≈2,885/yr ≈ 240/mo), ~1 run/week (258–261 runs/yr), 184 voids excluded per the Q4 2025 AP memo, and 22 duplicate register lines removed.
• In-house cost built from the 2025 supplies catalog (check stock $0.10 + envelope $0.10 + $1.00 postage = $1.20/check) plus loaded labor (AP Clerk $67,200 + 30% burden ÷ 2,080 hrs = $42.00/hr × 0.5 hr/run ≈ $1.93/check).
• Outsourced cost priced from the actual Services Agreement — including the $0.75/check volume tier, which Cascadia qualifies for in every month (min month = 174 checks) — rather than the $0.90 base rate the bank's proposal used.

Headline results (2021–2025 at actual volumes):

• 5-year total — In-house: $44,545 — Outsourced¹: $37,894 — Savings: $6,651 (≈$1,330/yr, ~15%)
• 2025 — In-house: $8,887 — Outsourced¹: $6,994 (+ fee) — Savings: $1,893

¹includes $2,500 implementation fee amortized over 5 years.

Recommendation: adopt Bluegrass Managed Check Disbursement beginning in 2026. The recurring saving is ~$1,893/yr at 2025 volume ($3.13 → $2.46 per check); the $2,500 implementation fee pays back in ~16 months; the saving holds in every year and every scenario tested (±25–50% volume, tier loss); break-even would require volume to more than double (~5,843 checks/yr). The memo also reconciles the result to the bank's $1,356/yr proposal (ours is better due to the volume tier), flags the labor-realization caveat (the ~130 freed hours/yr must be redeployed — outsourcing turns uneconomic only if >10 of the 30 setup minutes per run are retained), and lays out next steps (Q4 2026 enrollment, pricing confirmation, Positive Pay, continued ACH migration).

✅ The submitted PDF contains a 2021-2025 annual in-house versus outsourced check-processing cost comparison and a recommendation on adoption in 2026.

❌ The response recognizes that a large share of register rows are electronic disbursements (not printed checks) and excludes them from true check volumes. Electronic items are identifiable because their Check/Ref No is not a check number from the issued sequence — they carry long transaction/trace numbers rather than a 5-digit number in the 10001–20093 check range — and/or their payee is a payroll, tax, utility, insurance, or auto-pay vendor (e.g., US Treasury, ADP Payroll, KY EFTPS).

✅ The response excludes voided/cancelled check entries (e.g., Status = Void) from both the check-volume counts and all cost calculations, counting only cleared/issued checks.

❌ The response calculates check volumes independently and does not use the AP Payment Volume Summary's gross check counts (e.g., 2,838 for 2025) as the true check volumes.

❌ The response states the 2021 true (non-electronic) check volume as 1,852.

❌ The response states the 2022 true (non-electronic) check volume as 2,340.

❌ The response states the 2023 true (non-electronic) check volume as 2,809.

❌ The response states the 2024 true (non-electronic) check volume as 1,984.

❌ The response states the 2025 true (non-electronic) check volume as 138.

❌ The response uses 183 check run days for 2021.

❌ The response uses 224 check run days for 2022.

❌ The response uses 259 check run days for 2023.

❌ The response uses 214 check run days for 2024.

❌ The response uses 73 check run days for 2025.

✅ The response uses the AP Clerk base salary of $67,200 to calculate in-house labor costs.

✅ The response applies a 30% burden to the base salary for in-house labor costs.

✅ The response divides the loaded salary by 2,080 hours (or a 40-hour work week) to derive the hourly rate.

✅ The response states the fully loaded AP Clerk hourly rate as $42.00.

✅ The response applies a time assumption of 0.5 hours per check run for in-house costs.

✅ The response calculates the in-house labor charge as $21.00 per run day.

✅ The response uses a cost of $0.10 per check for check stock, based on $250/2,500.

✅ The response uses a cost of $0.10 per check for envelopes in the in-house calculation, based on $50/500.

❌ The response derives the toner cost as $0.05 per check by dividing the $374.99 cartridge cost by its 7,500-page yield (rounded up from 0.049).

✅ The response uses a cost of $1.00 per check for postage in the in-house calculation.

❌ The response uses $1.25 per mailed check as the total in-house non-labor variable cost, whether stated as one subtotal or shown as $0.25 for check stock, envelope, and toner plus $1.00 for postage.

❌ The response calculates the 2021 in-house total cost as approximately $6,158 (acceptable range $6,100 to $6,200).

❌ The response calculates the 2022 in-house total cost as approximately $7,629 (acceptable range $7,600 to $7,700).

❌ The response calculates the 2023 in-house total cost as approximately $8,950 (acceptable range $8,900 to $9,000).

❌ The response calculates the 2024 in-house total cost as approximately $6,974 (acceptable range $6,900 to $7,000).

❌ The response calculates the 2025 in-house total cost as approximately $1,706 (acceptable range $1,650 to $1,750).

✅ The response uses an outsourced variable cost of $1.83/check in monthly cycles that qualify for the $0.75 tier of per-check processing (100+ checks/month - every month of 2021–2024).

❌ The response uses an outsourced variable cost of $1.98/check in monthly cycles that don't qualify for the $0.75 tier and are charged 0.90 per-check processing (fewer than 100 checks/month - every month of 2025).

✅ The response states the outsourced envelope cost as $0.08 per check.

✅ The response states the outsourced postage cost as $1.00 per check.

✅ The response applies a fixed platform/service fee of $150 per month (or $1,800 per year).

✅ The response explicitly treats the $150/month platform fee as a fixed annual cost, not as a variable per-check cost.

✅ The response includes a one-time setup fee of $2,500 for the outsourced option.

✅ The response applies the $2,500 setup fee only once, not as an annual recurring cost.

❌ The response calculates the 2021 outsourced operating cost (excluding setup fee) as approximately $5,189 (acceptable range $5,150 to $5,250).

❌ The response calculates the 2022 outsourced operating cost (excluding setup fee) as approximately $6,082 (acceptable range $6,050 to $6,150).

❌ The response calculates the 2023 outsourced operating cost (excluding setup fee) as approximately $6,940 (acceptable range $6,900 to $7,000).

❌ The response calculates the 2024 outsourced operating cost (excluding setup fee) as approximately $5,431 (acceptable range $5,400 to $5,500).

❌ The response calculates the 2025 outsourced operating cost (excluding setup fee) as approximately $3,000 (acceptable range $2,950 to $3,050).

✅ The response either excludes the $0.80/check Positive Pay fee from the business case cost comparison or presents it as a separately disclosed equal add-on to both options outside the annual cost totals, on the basis that the fee applies regardless of whether checks are produced in-house or outsourced; the annual in-house and outsourced totals graded elsewhere in this rubric exclude this fee.

❌ The response either excludes the $0.30/check Check Clearing fee from the business case cost comparison or presents it as a separately disclosed equal add-on to both options outside the annual cost totals, on the basis that the fee applies regardless of whether checks are produced in-house or outsourced; the annual in-house and outsourced totals graded elsewhere in this rubric exclude this fee.

❌ The response concludes that outsourcing is cheaper than in-house processing on an operating basis for the years 2021 through 2024, but more expensive in 2025.

❌ The response calculates that at 2025 volumes, outsourcing loses approximately $1,294 annually on operating costs compared to in-house processing (accept a stated annual difference of $1,200 to $1,400).

❌ The response states that the $2,500 setup fee is unrecoverable (or its payback fails), given the 2025 check volumes.

❌ The response provides a final recommendation advising against implementing the check outsourcing project.

❌ The response's recommendation is based on the most recent 2025 run rate (instead of the 2021-2024 results), as outsourcing is more expensive to operate (a declining number of checks in 2025 and the one-time setup fee cannot be recovered).

✅ The response shows the per-unit rates underlying each year's in-house and outsourced operating cost totals.

✅ The response shows the annual check volumes underlying each year's in-house and outsourced operating cost totals.

✅ The response shows the run-day counts underlying each year's in-house and outsourced operating cost totals.

✅ The response shows the fixed fees underlying each year's in-house and outsourced operating cost totals.

✅ The response contains no internal contradiction where the same year is assigned materially different true-check volumes, run-day counts, or total costs across different sections (differences attributable to rounding/display precision do not count)

✅ Any added break-even, sensitivity, or other analytical figure is reproducible from the assumptions and arithmetic shown in the response and does not contradict its stated rates, volumes, monthly minimum charges, or annual totals.

❌ The response applies the $250 minimum monthly service charge because 2025's per‑check plus platform fees fall below $250 every month, 2025 outsourced operating cost equals the minimum 12 × $250 = $3,000 - not the ~$2,073 a plain per‑check calculation gives.

✅ The response identifies and removes 22 total duplicate check entries, so true check volumes are not overstated.

✅ The response bases check volumes on the AP Check Register.

✅ The response bases pricing on the Services Agreement.

❌ The response does not rely on the Bluegrass "Preliminary Savings Estimate" proposal — e.g., it does not adopt the proposal's assumed ~2,800 checks/year or its "worth pursuing" conclusion, recognizing it as an illustrative marketing estimate that ignores the 2025 volume decline, the $250 minimum, and volume-tier pricing.