Master Outline & Study Guide

FP&A Financial Model — Skechers U.S.A. (SKX) (Excel · 3-Statement · Bear/Base/Bull · Sensitivity · KPI Dashboard)

Purpose of this document: A complete, self-contained reference for studying this portfolio project. Use it to prepare for interviews — to explain what was built, how it works, why each decision was made, and to answer probing technical questions. When studying with Claude, this is the file Claude reads to teach and quiz you.

One sentence to anchor everything: This is a full three-statement FP&A model for Skechers U.S.A. (NYSE: SKX) — five years of historical actuals reconstructed from SEC 10-Ks plus a three-year forecast — built around a segment-level revenue engine, switchable Bear / Base / Bull scenarios driven by a CHOOSE/MATCH selector, a genuine (not forced) three-statement tie, live two-variable sensitivity tables, and a self-checking KPI dashboard.

Table of Contents

1. The 30-Second Pitch
2. Why This Project Exists (Business Context)
3. The Tech Stack & Skills
4. The Data Foundation — SEC 10-Ks
5. The Workbook Architecture
6. Tab 1 — README (cover page)
7. Tab 2 — Historical Financials (the anchor)
8. Tab 3 — Assumptions (the scenario engine)
9. Tab 4 — Supporting Schedules (PP&E, Debt, Shares)
10. Tab 5 — Income Statement Forecast
11. Tab 6 — Balance Sheet Forecast
12. Tab 7 — Cash Flow Forecast
13. Tab 8 — Variance Analysis
14. Tab 9 — Scenario Analysis & Sensitivity Tables
15. Tab 10 — KPI Dashboard
16. The Three-Statement Tie (deep dive)
17. Excel & Modeling Techniques Demonstrated
18. Key Results & Numbers to Memorize
19. Limitations & Honest Caveats
20. Design Decisions & Trade-offs (the “Why”)
21. Interview Q&A
22. How to Walk Through This Project Live
23. Glossary

1. The 30-Second Pitch

This is a fully-integrated three-statement financial model for Skechers U.S.A., Inc. (NYSE: SKX) — income statement, balance sheet, and cash flow statement that fully articulate, so a change to any driver flows correctly through all three. It covers FY2020–FY2024 actuals reconstructed from SEC 10-K filings plus a FY2025–FY2027 forecast driven by a transparent, switchable Bear / Base / Bull assumption set.

The analytical thesis is channel mix shift: Direct-to-Consumer (DTC) carries roughly 66% gross margin versus 40–43% in Wholesale, and grew from 38.5% of revenue in FY2020 to 43.1% in FY2024. The model forecasts each segment separately and continues the mix shift (DTC 43.6% → 44.5% across FY2025–27), making the margin-expansion story explicit and auditable rather than buried in a single blended revenue line.

The Base case lands inside Skechers’ February 2025 forward guidance — FY2025 revenue ~$9.76B, gross margin 53.3%, operating margin 10.4%, diluted EPS $4.40, and free cash flow $343.8M. The headline engineering result is integrity: the balance sheet ties to zero and CF ending cash ties to BS cash exactly — across all three scenarios in every forecast year — verified by audit.

One-line version: “I built a full three-statement Excel model for Skechers from public 10-Ks — a segment-level revenue engine, Bear/Base/Bull scenarios driven by CHOOSE/MATCH, working capital from DSO/DIO/DPO, live two-variable sensitivity grids, and a KPI dashboard with self-checking model-integrity gates that all read PASS.”

Deliverable: excel/SKX_Financial_Analysis_Project_Final.xlsx

2. Why This Project Exists (Business Context)

The simulated scenario. The project plays the role of an FP&A analyst building a defensible forecasting model for Skechers U.S.A., Inc. — a publicly-traded global footwear company with two reportable segments (Wholesale and Direct-to-Consumer). The model needs to support standard FP&A workflows: scenario planning, sensitivity analysis, variance reporting, and executive-ready KPI reporting.

The business questions it answers. (1) How does Skechers’ financial picture look on a forward three-year horizon? (2) How sensitive is the answer to revenue growth, gross margin, OpEx ratio, and CapEx? (3) What does the spread between a pessimistic, base, and optimistic view actually look like in revenue, EPS, and free cash flow? (4) How did FY2024 actuals land versus guidance, and what drove the variances? (5) Does the model itself remain internally consistent under every scenario — and how is that proved, not asserted?

Why it’s a strong portfolio project. It demonstrates the full FP&A modeling toolkit in one artifact: reading and reconstructing financials from SEC 10-K filings, segment-level revenue and margin modeling, working-capital forecasting via days ratios, PP&E / debt / shares rollforward schedules, switchable scenarios wired through CHOOSE/MATCH, two-variable sensitivity analysis, variance analysis, named-range architecture, and a dashboard layer that reports live model-integrity results — not just charts. Each piece is something a real corporate-finance team would actually build.

The core principle behind the whole build: transparency and integrity. Every driver is editable; every formula is traceable; the three statements articulate correctly; and the model proves its own integrity via four live PASS/FAIL checks on the dashboard. The model is not “trust me, it ties” — it’s “look, the dashboard tells you it ties.”

3. The Tech Stack & Skills

Everything is in one Excel workbook, 10 tabs, designed to run from a single scenario dropdown. The skills, and where each shows up:

The mental model. Think of the workbook as a factory:

• Inputs: the Assumptions tab (raw drivers, three scenarios side-by-side).
• The selector switch: Scenario dropdown on Assumptions!B5, fed into ScenarioIndex on E5 via MATCH, then used in every CHOOSE formula in the “Active” block.
• The mid-stream processors: the Supporting Schedules tab turns assumptions into PP&E, D&A, interest expense, debt balances, and share counts.
• The three engines: the IS / BS / CF Forecast tabs build the financials.
• The QA station: the Variance Analysis + Scenario Analysis + dashboard integrity checks all confirm the output is internally consistent.
• The shipping bay: the KPI Dashboard presents the result.

Why “switchable” is the recurring theme. A static model that only shows one scenario is a snapshot, not a tool. Flipping Assumptions!B5 from Base to Bear or Bull re-drives every forecast statement, every schedule, every dashboard KPI, every scenario recap value, and every integrity check — together, in one click. That switchability is what separates an FP&A model from a one-time analysis.

4. The Data Foundation — SEC 10-Ks

What it is. All historical data comes from Skechers’ SEC EDGAR 10-K filings for FY2020–FY2024, verified against the XBRL instance documents (the structured-data filings that hold the exact numbers). The Base-case calibration is anchored to Skechers’ February 2025 forward guidance.

What’s reconstructed in the Historical Financials tab:

Key historical facts to know cold:

• Revenue trajectory: $4.61B (FY2020) → $8.97B (FY2024) — nearly doubled in 4 years.
• Wholesale revenue: $2.835B → $5.1005B; DTC revenue: $1.7784B → $3.8689B.
• DTC mix: 38.5% (FY2020) → 43.1% (FY2024) — the central thesis.
• Wholesale GM: ~39.2% → 43.3% (general expansion).
• DTC GM: 61.5% → 66.2% (stable around the mid-60s).
• Blended GM: 47.8% → 53.2% (mix shift + WH GM improvement together).
• FY2021 tax line is anomalous: a $346.8M one-time tax benefit from an intra-entity IP transfer drove a negative effective tax rate of −43.2% and a one-year EPS spike to $4.73. This is documented in A117 of the Historical Financials tab and called out again in the README caveats. Always mention it if asked why FY2021 looks weird.

The Cash Flow Statement unit note. The Historical CF is reported in $M, while the rest of the workbook is in $000s. The forecast CF is in $000s. The conversion (×1000) shows up explicitly in the formulas linking back, e.g., =-B88*1000/B8 for CapEx % of revenue.

5. The Workbook Architecture

The deliverable SKX_Financial_Analysis_Project_Final.xlsx has 10 sheets. Knowing the sheet roles and the flow between them is knowing the project’s structure.

The data flow through the sheets (read this as the model’s “pipeline”):

Historical Financials (10-K data, FY2020–24)
            │
            ├──► Assumptions  ───[CHOOSE/MATCH on B5]──►  Active block (cols L–N)
            │                                                  │
            │                                                  ▼
            │                                       Supporting Schedules
            │                                       (PP&E, Debt, Shares)
            │                                                  │
            ▼                                                  ▼
       Income Statement Forecast  ────►  Balance Sheet Forecast
              │                                  │
              ▼                                  ▼
        Cash Flow Forecast  ──[ending cash]──►  BS Cash (closes the loop)
              │
              ├──► Variance Analysis (compares actuals vs. FY24 budget)
              ├──► Scenario Analysis (snapshots + sensitivity grids)
              └──► KPI Dashboard (year selector + integrity checks)

The two control cells. Two yellow-fill cells govern what the whole workbook displays:

• Assumptions!B5 → the scenario dropdown (Bear / Base / Bull). Drives ScenarioIndex on E5 via =MATCH(B5,{"Bear","Base","Bull"},0). Every formula in the “Active” column block uses =CHOOSE($E$5, …) to pick the right scenario’s value.
• KPI Dashboard!B5 → the forecast year dropdown (FY2025 / FY2026 / FY2027). Drives Dash_YearIdx on B6 via =MATCH(B5,{"FY2025","FY2026","FY2027"},0). The scorecard KPIs use =CHOOSE(Dash_YearIdx, …) against three staging columns (X / Y / Z).

Named-range architecture. The workbook defines roughly 150 named ranges. Every cross-sheet reference is via a name, not a coordinate — e.g., the Income Statement uses Wh_Growth_25 instead of Assumptions!L13. Why this matters: names survive row inserts, and they make formulas readable (=B8*(1+Wh_Growth_25) is obviously “prior-year sales times next-year growth”). The forecast outputs are also exposed as names (IS_NetEarnings_25, BS_BalanceCheck_25, CF_TieCheck_25, etc.) so the dashboard’s PASS/FAIL formulas read the live model without hard cell addresses.

6. Tab 1 — README (cover page)

The README is the model’s front door and works without opening any other tab. It contains seven sections:

1. What this model is — the one-paragraph thesis (channel mix, FY2025 Base outputs).
2. How to read it — model map — a 10-row table mapping each tab to a one-line role.
3. Color legend — the modeling convention used in cells:
   • Blue text = hardcoded input (the user can change).
   • Black text = formula / calculation.
   • Green text = link to another sheet in the workbook.
   • Yellow fill = the two selector cells (Assumptions!B5 and KPI Dashboard!B5).
4. Data sources — SEC EDGAR 10-Ks + February 2025 forward guidance.
5. Key design notes — the three big architectural decisions: (a) no circular reference (interest on beginning-of-period debt), (b) genuine three-statement tie (cash built as plug first, then CF built independently, then swapped in), (c) CHOOSE/MATCH-driven scenario & year switching.

6. Live integrity status — a 3-row × 4-col table that reads the live model via named ranges:

   Check	FY2025	FY2026	FY2027
   Balance Sheet (Assets = Liabilities + Equity)	PASS	PASS	PASS
   Cash Flow tie (CF ending cash = BS cash)	PASS	PASS	PASS

   Formula example: =IF(ABS(BS_BalanceCheck_25)<1,"PASS","FAIL"). The <1 tolerance is $1K rounding (the model is in $000s, so $1K is effectively zero).

7. Limitations & caveats — public data only, FY2021 tax-benefit anomaly, September-2025 3G Capital take-private, illustration-only purpose.

Why this README earns its keep. Most Excel models open to a wall of numbers. This one opens to “here is what this is, here is how to read it, here is proof it works.” It’s the single most important onboarding artifact in the file.

7. Tab 2 — Historical Financials (the anchor)

This is the foundation tab — five years of actuals that every forecast value eventually compares against. 120 rows by 7 columns, organized into five sections:

(A) Consolidated Income Statement (rows 7–34). Sales, COGS (with computed Gross Profit and Gross Margin %), Selling vs. G&A separated, EBIT and Operating Margin %, Other Income/Expense, EBT, tax expense and Effective Tax Rate %, Net Earnings, NCI, Net Earnings Attributable to SKX, Net Margin %, Diluted EPS, Diluted Shares (M), plus derived %-of-revenue lines for Selling and G&A and Revenue Growth % YoY.

(B) Segment Revenue & Gross Margins (rows 36–45). Wholesale revenue + growth + GM, DTC revenue + growth + GM, and DTC Mix % of Total Revenue. This is where the channel thesis is most visible.

(C) Consolidated Balance Sheet ($000s, Dec 31) (rows 47–75). Current assets (Cash, ST Investments, AR, Inventory, Other CA → Total CA), non-current assets (PP&E, ROU Assets, Other NCA → Total Assets), current liabilities (AP, Accrued, Current Leases, Current Debt → Total CL), long-term (LT Debt, LT Leases, Other LT → Total Liabilities), equity (Retained Earnings, Total Equity incl. NCI) → Total L+E, and a Balance Check row that subtracts and shows 0 in every historical year. Note: Other Current Assets and Other LT Liabilities are derived as residuals to make the balance sheet balance to the as-reported totals.

(D) Consolidated Cash Flow Statement ($M) (rows 77–101). Operating section (net income, D&A, SBC, other operating adjustments, AR/Inv/AP changes, other WC), Investing (CapEx + other), Financing (debt issued/repaid, buyback, other), FX adjustment, Net Change in Cash, plus a derived Free Cash Flow = OCF + CapEx line. Units: $M here, unlike the rest of the workbook.

(E) Derived Metrics & Working-Capital Analysis (rows 103–114). DSO, DIO, DPO, CCC, CapEx % of revenue, D&A as % of prior PP&E, total debt, NCI as % of net earnings, SBC as % of revenue. These are the cells that calibrate the forecast assumptions — the Base-case DSO 40, DIO 165, DPO 108 are anchored to what actually happened in FY2024.

The notes block at the bottom documents the FY2021 tax-benefit anomaly, the $M vs. $000s convention, the residual derivation for Other CA / Other LT Liabilities, and the XBRL verification.

Why this tab is worth studying in itself. Reading the entire historical block in order is the cleanest way to learn the SKX business — and it makes the forecast tabs make sense, because every forecast number is a continuation or normalization of a historical trend visible here.

8. Tab 3 — Assumptions (the scenario engine)

This is the single tab the user actually interacts with to drive the model. Roughly 50 rows by 14 columns, organized into eight assumption groups.

The control row at the top:

• B5 = the scenario dropdown (“Bear” / “Base” / “Bull”). Default: Base.
• E5 = =MATCH(B5,{"Bear","Base","Bull"},0) → returns 1, 2, or 3.

The column structure. For each driver, three columns per forecast year for the three scenarios — B:D for FY2025 (Bear/Base/Bull), E:G for FY2026, H:J for FY2027 — and then the “ACTIVE (live scenario)” block in columns L:N for FY2025, FY2026, FY2027 respectively. Each “Active” cell is a CHOOSE formula:

L13 = =CHOOSE($E$5, B13, C13, D13)        ← FY2025 active value (Wholesale growth)
M13 = =CHOOSE($E$5, E13, F13, G13)        ← FY2026 active value
N13 = =CHOOSE($E$5, H13, I13, J13)        ← FY2027 active value

Every formula downstream in the model reads from the Active block via named ranges (Wh_Growth_25, DTC_Growth_25, Wh_GM_25, DSO, CapEx_25, Tax_25, …). Flipping B5 re-drives E5, which re-drives every CHOOSE, which re-drives every named range, which re-drives every forecast cell. Effectively a one-cell switchboard.

The eight assumption groups, with FY2025 Base values:

Calibration logic. Notice what varies across scenarios versus what doesn’t. Working capital ratios (DSO/DIO/DPO/Accrued/OCA), tax rate, NCI, Other Income, interest rate, and SBC are held flat — those are structural features of the business, not bets on the future. Revenue growth, gross margin, OpEx ratio, CapEx, and buyback intensity vary because those are where the real uncertainty and management discretion live.

9. Tab 4 — Supporting Schedules (PP&E, Debt, Shares)

This tab transforms a handful of high-level assumptions into the balance-sheet-grade detail the forecast IS/BS/CF need. Roughly 28 rows by 7 columns, three rollforwards.

(A) PP&E / D&A Rollforward (rows 6–11). Standard “BASE” rollforward — Beginning + Additions − Depreciation = Ending:

Beginning PP&E         (from prior period: FY24 actual for FY25, then prior FY ending)
(+) CapEx              = CapEx_25 / CapEx_26 / CapEx_27 (from Assumptions)
(-) D&A                = -Beginning PP&E × DA_Rate (14% of prior PP&E)
(=) Ending PP&E        = sum

A separate “D&A (positive, for CF add-back)” row exposes D&A with a positive sign for the indirect-method cash flow.

(B) Debt Schedule (rows 13–20). Same BASE structure:

Beginning Debt         (FY24 total debt $454,919; then prior ending)
(+) New Borrowings     = New_Borrow (0 in all scenarios)
(-) Scheduled Repay    = -Repay ($50,000 each year)
(=) Ending Debt        = beginning + borrowings − repayments
Interest Expense       = Beginning Debt × 5%   ← the no-circular-reference trick
Current Portion        = Repay ($50,000)
Long-term Portion      = Ending Debt − Current Portion

The interest-expense calculation is the single most important modeling decision in the file. Interest is computed on the beginning-of-period debt balance, not the average or the ending balance. This is the industry-standard simplification that breaks the circular loop:

Interest expense → Net income → Cash flow → Cash balance → Debt balance → Interest expense

Computing on beginning debt freezes the input to interest before any of those downstream effects can feed back. The workbook contains zero circular references and does not need iterative calculation enabled.

(C) Shares Outstanding Schedule (rows 22–26). Same BASE structure, in millions:

Beginning Diluted Shares   (Shares_Begin = 153.8M for FY25, then prior ending)
(-) Shares Repurchased     = -Repurchase / Repurchase_Price / 1000  (e.g., $350K÷$65÷1000 = 5.38M)
(+) SBC Dilution           = (Revenue × SBC_Pct) / Repurchase_Price / 1000
(=) Ending Diluted Shares  = sum

The SBC dilution proxy is revenue × 1% / $65 / 1000 — i.e., the dollar value of stock- based compensation is converted to “new shares” at the assumed market price.

Why the schedule names matter. Every output from this tab is exposed as a named range — PPE_End_25, DA_25, Debt_End_25, Int_Exp_25, Debt_Curr_25, Debt_LT_25, Shares_End_25 — so the IS/BS/CF tabs can just reference the names directly.

10. Tab 5 — Income Statement Forecast

The IS forecast is structured as a segment build → consolidated IS → tax → EPS flow. 42 rows by 7 columns.

(A) Segment build (rows 7–15). The heart of the channel-mix thesis:

Wholesale Revenue (FY25) = FY24 Wholesale × (1 + Wh_Growth_25)   ← Active scenario growth
DTC Revenue       (FY25) = FY24 DTC       × (1 + DTC_Growth_25)
Total Revenue            = Wholesale + DTC
DTC Mix %                = DTC / Total

Wholesale COGS = Wholesale Revenue × (1 − Wh_GM_25)
DTC COGS       = DTC Revenue       × (1 − DTC_GM_25)
Total COGS     = WH COGS + DTC COGS

(B) Consolidated Income Statement (rows 17–27). Sales = total revenue from the segment build; Cost of Sales = total COGS; Gross Profit = difference; Gross Margin % = ratio. Selling Expenses = Sales × Selling_Pct; G&A = Sales × GA_Pct; Total OpEx = Selling + G&A; Earnings from Operations = Gross Profit − OpEx; Operating Margin % = ratio.

(C) Below-the-line and bottom-line (rows 29–39). Interest Expense pulled from schedules (=-Int_Exp_25, shown negative as an expense); Other Income from assumptions; EBT = EBIT + Interest + Other; Tax = EBT × Tax_25; Net Earnings = EBT − Tax; NCI = Net Earnings × NCI_Pct; Net Earnings Attributable to SKX = Net Earnings − NCI; Net Margin %; Diluted EPS = Net Earnings (SKX) / Shares_End / 1000 (the /1000 converts $000s ÷ Millions of shares to $/share); Diluted Shares from the schedule.

(D) Blended GM check (row 42). A consistency check:

=GrossMargin% − ((WH Revenue × WH GM + DTC Revenue × DTC GM) / Total Revenue)

Should equal zero. If the segment-level build ties to the consolidated statement, the blended GM% must equal the revenue-weighted segment GM%. All three forecast years show 0. This is the “did I do the segment math right” gate.

FY2025 Base outputs the model produces:

11. Tab 6 — Balance Sheet Forecast

The BS forecast follows the same line ordering as the historical balance sheet so they’re trivially comparable. 38 rows by 7 columns.

Where each line comes from (the data lineage is the lesson here):

Working-capital math, step by step. AR = Sales × DSO / 365 means: “if I collect in 40 days, my receivables balance at year-end is 40/365 of a year’s worth of sales.” Inventory = COGS × DIO / 365 (“if I hold 165 days of inventory…”). AP = COGS × DPO / 365 (“if I pay vendors in 108 days…”). These three formulas are the bridge from the income statement to the balance sheet — every dollar of revenue or COGS flowing into IS forecasts generates a proportionate movement in working capital on the BS.

Why the balance check being zero is the headline integrity test. Cash was originally built as the plug (any imbalance flows to cash), then the cash flow statement was built independently and its ending cash swapped in. Now cash is not a free variable — it’s a calculated value from CF. So the only way Assets = L+E is if every other line is right: working capital, PP&E, debt, retained earnings, and the equity rollforward all have to articulate. The zero in row 38 is the model proving itself.

12. Tab 7 — Cash Flow Forecast

Indirect-method cash flow, derived from the forecast IS and the change in BS lines — no new assumptions. 35 rows by 7 columns.

The structure:

OPERATING ACTIVITIES
  Net Earnings (incl. NCI)             ← from IS, IS_NetEarnings_25
  (+) D&A                              ← from Supporting Schedules, DA_25 (non-cash add-back)
  (+) SBC                              ← Revenue × SBC_Pct (non-cash add-back)
  (-) Increase in AR                   ← = Prior AR − Current AR  (asset increase = use of cash)
  (-) Increase in Inventory            ← = Prior Inv − Current Inv
  (+) Increase in AP                   ← = Current AP − Prior AP  (liability increase = source)
  (+) Increase in Accrued              ← same pattern
  (-) Increase in Other CA             ← same pattern
  = Net Cash from Operating

INVESTING ACTIVITIES
  CapEx                                ← -CapEx_25 (outflow)
  = Net Cash from Investing

FINANCING ACTIVITIES
  Debt Issued                          ← New_Borrow (= 0)
  Debt Repaid                          ← -Repay (= -50,000)
  Share Repurchases                    ← -Repurchase (= -350,000 Base)
  = Net Cash from Financing

NET CHANGE IN CASH = OCF + ICF + FCF
Beginning Cash                         ← Prior period ending cash
Ending Cash    = Beginning + Net Change   ← This is what BS Cash links to

Memo: Free Cash Flow = OCF + CapEx     ← Negative CapEx, so CF + CapEx = CF − |CapEx|

The indirect-method working-capital convention is worth knowing cold. “Δ” in this context is signed for cash flow, not for the balance sheet:

• Asset increase = use of cash → AR rises means you sold but didn’t collect → cash flow line is negative, computed as prior − current (so a bigger AR gives a negative number).
• Liability increase = source of cash → AP rises means you bought but didn’t pay → cash flow line is positive, computed as current − prior.

FY2025 Base CF outputs:

The CF tie check (row 35): = Ending Cash − BS_Cash_25. Zero by construction after the BS cash plug was swapped to point at this cell. The genuine integrity gate is now the BS balance check, not this row — but the row is left in the model as a visible “linkage holds” confirmation.

13. Tab 8 — Variance Analysis

A standard FP&A workflow: how did FY2024 actuals compare to a budget calibrated to February 2024 forward guidance? 22 rows by 6 columns.

Format. For each line: Budget column (hardcoded), Actual column (linked to the Historical Financials tab), $ Variance = Actual − Budget, % Variance = $ Variance / Budget, plus a one-sentence Commentary explaining the driver.

The lines covered. Wholesale Revenue, DTC Revenue, Total Revenue, Gross Profit, Gross Margin %, Selling, G&A, Earnings from Operations, Operating Margin %, Net Earnings (SKX), Diluted EPS.

Headline variances (FY2024 actual vs. February 2024 budget):

The story. Revenue beat plan by ~3% (both segments), GM beat by ~65 bps (channel mix

• freight). Operating income narrowly missed (−$9M) because G&A came in 5% above plan on U.S./China distribution-center expansion and DTC store buildout. Below the line, net earnings and EPS still beat by ~11% — the revenue + GM upside translating through, plus buyback-driven share count reduction lifting EPS.

The teachable framing. This is the “what’s the right way to talk about a beat that isn’t really a beat” exercise. Revenue beat, gross margin beat, but operating income missed — and the EPS beat is partly a share count phenomenon, not pure operating performance. An analyst reading just the EPS line gets the wrong story; an analyst reading the whole variance stack gets the right one.

14. Tab 9 — Scenario Analysis & Sensitivity Tables

Two distinct things on one tab:

14.1 The Bear/Base/Bull recap table (rows 6–16)

A snapshot table comparing FY2025 outputs across all three scenarios:

The spread shows operating leverage at work. A roughly ±4% swing in revenue (9.41B → 10.17B) drives roughly a ±40% swing in EPS ($2.48 → $6.18), exactly as a fixed-cost-bearing earnings stream should behave. That’s the lesson: small changes at the top become big changes at the bottom because fixed costs don’t scale linearly.

Why Bear/Bull values are pre-captured. Because the model displays one scenario at a time (whichever is selected on Assumptions!B5), the Bear and Bull columns in this table are captured snapshots, not live formulas — they’re blue inputs you can overwrite by re-running the model in those scenarios. The Base column ties to the live model.

14.2 The self-contained FY2025 sensitivity mini-model (rows 21–87)

The bottom of the Scenario tab carries a separate, self-contained FY2025 mini-model — roughly 30 cells that recompute Net Earnings (SKX) and FCF from scratch with their own input “knobs.” That mini-model drives three two-variable sensitivity grids:

Every cell in every grid is a formula — for instance Table 1’s general form:

=(($B$33*(1+$B64))*C$63 − ($B$33*(1+$B64))*$B$40 − $B$34 + $B$28) × (1−$B$32) × (1−$B$29)

Reads as: (Prior Revenue × (1 + Row growth)) × Column GM (gross profit) minus that same revenue times the OpEx ratio (EBIT) minus Interest, plus Other Income, then taxed and NCI-adjusted.

The engineering decision worth knowing. Excel’s built-in Data Tables can drive sensitivity, but they require the input cells to be direct, hardcoded input cells. The model’s actual revenue-growth and gross-margin “inputs” are CHOOSE-driven formulas in the Active block — Data Tables can’t substitute into formula cells. So the equivalent functionality was engineered manually: build a self-contained FY2025 mini-model whose inputs are direct cells (B38, B39, B40, B41), then write each grid cell as a formula that swaps in the row and column values directly. The result is live, formula-filled, fully auditable sensitivity grids that recalculate on any change — better than a static screenshot, and able to coexist with the CHOOSE-driven main model.

The mini-model anchor cells (the “base constants” you should know exist): prior-year WH revenue (B23 = 5,100,500), prior DTC revenue (B24 = 3,868,900), prior PP&E (B25 = 1,834,930), beginning debt (B26 = 333,096), interest rate (5%), Other Income (-$26,000), NCI (12%), SBC (1%), D&A rate (14% × prior PP&E), tax rate (22.5%).

15. Tab 10 — KPI Dashboard

The executive deliverable. Roughly 51 rows by 26 columns, where columns R:Z are a hidden-able data-staging area that pulls every series the dashboard needs from the other tabs via cross-sheet links.

The control cell. B5 is a forecast-year dropdown (FY2025 / FY2026 / FY2027) with Dash_YearIdx on B6 = MATCH(B5,{"FY2025","FY2026","FY2027"},0). Every KPI in the scorecard uses =CHOOSE(Dash_YearIdx, ...) to pick the matching year from the staging columns.

The six-KPI scorecard (one big cell per KPI, with a smaller “YoY change vs. prior year” cell underneath each):

1. Total Revenue (A8 / A9)
2. Gross Margin % (D8 / D9)
3. Operating Margin % (G8 / G9)
4. Net Earnings (SKX) (A12 / A13)
5. Diluted EPS (D12 / D13)
6. Free Cash Flow (G12 / G13)

Example formula: A8 = =CHOOSE(Dash_YearIdx, $X$8, $Y$8, $Z$8) where X/Y/Z are the staging columns for FY2025E / FY2026E / FY2027E. The YoY delta below is just the selected-year value minus the prior-year value.

The data-staging block (cols R:Z, hidden-friendly). Three rows of context per metric: row label (col R), then values for FY2020 / FY2021 / FY2022 / FY2023 / FY2024 (actual, cols S:W) / FY2025E / FY2026E / FY2027E (forecast, cols X:Z). The staging block carries Wholesale revenue, DTC revenue, Total Revenue, Gross Margin %, Operating Margin %, Net Margin %, DSO, DIO, DPO, Net Earnings (SKX), Diluted EPS, and Free Cash Flow. The five charts (Revenue by Segment, Margin Trend, Working-Capital Days, FCF Bridge, Scenario Comparison) all read from this staging area.

The FY2025 scenario comparison block (rows 19–22). Pulls the Bear / Base / Bull Net Earnings and FCF figures directly from the Scenario Analysis tab so the dashboard visualizes the spread without re-deriving it.

The Model Integrity Checks table (rows 45–51). The most important block on the dashboard — four PASS/FAIL gates reading the live model:

All four pass across every forecast year. Formula example: =IF(ABS(BS_BalanceCheck_25)<1,"PASS","FAIL"). The <1 tolerance is $1K rounding.

The dashboard is a model, not a screenshot. Because every KPI, chart, and integrity flag is a live formula reading the model via named ranges, flipping the scenario on Assumptions!B5 re-drives the entire dashboard. That’s the wow factor — and it’s also the honesty factor, because the user can see the integrity checks update as the scenario changes.

16. The Three-Statement Tie (deep dive)

If you only memorize one section of this model for an interview, memorize this one.

What “three-statement tie” actually means. The IS, BS, and CF aren’t three independent reports — they’re three views of the same underlying activity. For the model to be internally consistent:

1. The cash balance on the BS at year-end must equal the ending cash on the CF.
2. The change in retained earnings between years on the BS must equal Net Earnings (SKX) on the IS (less any dividends — Skechers pays none in this period).
3. Every change in a BS line must be reflected somewhere on the CF (working-capital changes, PP&E movements, debt rollforward, equity transactions).

If any of those break, you have a “model that doesn’t tie.”

How this model achieves a genuine tie (not a forced one). Standard junior-analyst practice is to make cash a balancing plug — whatever number is needed to make Assets = L+E. That works, but it hides errors: a mistake anywhere in the model just gets absorbed into cash, and you’d never know.

This model does it the harder, better way:

1. Cash starts as a plug. The balance sheet was first built with cash as a temporary balancing variable, so every other line — working capital, PP&E, debt, retained earnings, equity rollforward — could be verified independently as correct.
2. The cash flow statement is built separately. The CF starts from Net Earnings, adds non-cash items (D&A, SBC), subtracts working-capital builds (which it derives from BS differences, using the indirect-method sign convention), nets out CapEx, debt repayments, and buybacks — and produces an ending cash balance entirely on its own.
3. The CF ending cash is swapped into the BS cash line. Now BS Cash = CF Ending Cash by link, not by plug.
4. The balance check is now a real test. With cash no longer free, the only way Assets = L+E is if every other line is right. The fact that B38 = 0 in every forecast year is the model proving its arithmetic.

The four integrity-check formulas on the dashboard are the runtime evidence:

• BS Balance Check (Assets − L+E < $1K)
• CF Ending Cash ties to BS Cash (0 by linkage, but visible)
• IS Net Earnings = CF starting line (verifies the income-statement number is the same one flowing into the cash flow)
• Retained Earnings consistent (logically implied by the balance check passing)

The other half: no circular reference. Computing interest on beginning-of-period debt freezes interest before it has a chance to feed back through NI → cash → debt → interest. The workbook has zero circular references, does not require iterative calculation enabled, and recalculates cleanly on every change in well under a second.

17. Excel & Modeling Techniques Demonstrated

A consolidated checklist — useful for “what did you use?” questions.

Excel mechanics

• CHOOSE + MATCH for scenario and year switching.
• Named ranges (~150 of them) for cross-tab clarity and resilience to row inserts.
• Cross-sheet references with absolute and relative anchors.
• Cell-level color coding (blue / black / green / yellow).
• IF(ABS(x)<1,"PASS","FAIL") pattern for tolerance-based integrity checks.
• Conditional formatting on the Variance tab (direction-aware $/% variance display).
• Charts (line, bar, combo) wired to a staging area, not directly to the model.

Financial modeling techniques

• Segment-level revenue and gross-margin build (Wholesale + DTC) with a blended-GM consistency check.
• BASE rollforwards (Beginning + Additions − Subtractions = Ending) for PP&E, debt, and shares.
• Interest on beginning-of-period debt to break the circular loop.
• Working capital from days ratios (DSO/DIO/DPO converted to dollar balances via Sales/COGS × days / 365).
• Indirect-method cash flow with the correct sign convention (asset ↑ = use, liability ↑ = source).
• Equity rollforward combining net earnings, NCI, buybacks (treasury method), and SBC (paid-in capital).
• CapEx-driven PP&E with D&A as a fixed % of prior-year PP&E.
• Two-variable sensitivity built as a self-contained mini-model + formula-filled grids (engineered alternative to Excel Data Tables when inputs are formula cells).
• Switchable Bear/Base/Bull with three full driver blocks and a CHOOSE-driven Active block.

FP&A workflow demonstrations

• Variance analysis with budget calibration to public guidance, $/% variances, and written commentary.
• Scenario recap for executive comparison.
• KPI dashboard with year selector and YoY deltas.
• Live model-integrity checks reported on the dashboard.

Data work

• Reconstruction of 5 years of full financials from SEC 10-K filings, including segment detail and derived working-capital metrics.
• XBRL-verified balance-sheet numbers; residuals used to make “Other CA” and “Other LT Liabilities” reconcile to as-reported totals.

18. Key Results & Numbers to Memorize

Interviewers will want concrete numbers. Memorize the headline figures.

FY2024 actuals (the launching pad)

FY2025 Base (the headline forecast year)

Scenario spread (FY2025)

Three-year Base trajectory

The four-takeaway story the model tells

1. The DTC mix shift continues to expand blended gross margin. DTC at 43.6% of revenue in FY2025E (vs. 43.1% in FY2024) lifts blended GM by ~10 bps; mix shift compounds out to 53.5% blended GM by FY2027.
2. The Base case lines up with company guidance. $9.76B revenue and $4.40 EPS land inside Skechers’ February 2025 forward guidance — the Base case isn’t a wishlist; it’s calibrated.
3. Operating leverage is real. A roughly ±4% revenue swing across scenarios produces a ±40% EPS swing — buyback intensity (Bear $150K, Base $350K, Bull $500K in repurchases) amplifies it further.
4. Cash generation strengthens through the period. FCF rises from $344M (FY25) → $553M (FY26) → $722M (FY27) Base, as CapEx tapers from $650K → $550K → $500K and operating cash flow scales with the business.

19. Limitations & Honest Caveats

Volunteer these — knowing your model’s limits signals maturity.

1. Public data only. Built entirely from SEC 10-K filings and the company’s February 2025 forward guidance. No non-public diligence, no management interviews, no proprietary data.
2. FY2021 tax-benefit anomaly. FY2021 carries a $346.8M one-time tax benefit from an intra-entity IP transfer, which produces a negative effective tax rate of −43% and an artificially-high FY2021 EPS of $4.73. This is called out in the Historical tab (A117) and the README caveats. Comparing FY2021 numbers head-to-head with surrounding years without adjusting for this is misleading; the model uses the reported numbers as-is and documents the anomaly.
3. The September 2025 take-private. 3G Capital acquired Skechers in September 2025 for $9.4B / $63 per share. The model is built on public data through FY2024 plus February-2025 guidance — the acquisition does not invalidate the modeling exercise, but it does mean the forecast can’t be checked against subsequent public filings. This is documented in the README and the index.md page.
4. Some lines are held flat. Right-of-Use Assets, LT Lease Liabilities, Other Non-Current Assets, Other LT Liabilities, Current Lease Liabilities, Short-term Investments, and Other Equity (the non-RE portion’s lease and other-account starting balance) are held at FY2024 values across all forecast years. A more thorough model would forecast lease activity and ST investment yield explicitly; this model deliberately holds them flat to keep the forecast focused on the operating drivers that matter for the thesis.
5. Working capital is structural, not seasonal. DSO/DIO/DPO are held flat in all scenarios because they’re business-model features, not bets. A more sophisticated model would test sensitivity to working-capital intensity (especially DIO, where apparel/footwear inventory cycles can swing meaningfully).
6. Beginning-of-period interest is a simplification. It’s the industry-standard way to avoid circular references, but it slightly under-states interest in years when debt rises and over-states it when debt falls. With $50K/year in repayments on a $455K starting balance, the effect is immaterial here — but worth flagging.
7. No statistical or stochastic component. This is a deterministic FP&A model. No Monte Carlo, no regression, no probabilistic distribution over scenarios. The Bear/Base/Bull spread is judgment, not statistics.
8. Forecast is for analytical illustration; not investment advice. Documented in the README.

20. Design Decisions & Trade-offs (the “Why”)

Interviewers reward “why” answers. The deliberate choices and their rationale:

Why model Wholesale and DTC separately rather than one blended revenue line? Because the thesis is the channel mix. Wholesale GM is ~40–43%; DTC GM is ~66%. Every point of mix shift is worth ~25 bps of blended margin. Modeling them separately makes the margin-expansion story visible and auditable. Blended growth would have hidden it.

Why compute interest on beginning-of-period debt instead of average? Industry-standard simplification to break the interest → NI → cash → debt → interest circular loop. Models without this either need iterative calculation enabled (slow, opaque) or end up with circular-reference warnings the user has to ignore. With $50K repayments on $455K starting balance, the precision loss is immaterial; the architectural benefit (no circular references, fast recalc, clean dependency graph) is substantial.

Why build cash as a plug first, then swap in the CF ending cash? So the balance check is a real integrity test. If cash were a permanent plug, any error anywhere in the model would silently flow into cash and the balance would always “balance.” By making cash a calculated value from CF, the balance check becomes the gate — it can only equal zero if every other line is correct. The model proves itself.

Why expose ~150 named ranges? Three reasons. (1) Readability: =B8*(1+Wh_Growth_25) is obviously next-year wholesale revenue; ='Assumptions'!L13 is not. (2) Resilience: named ranges survive row inserts that would break coordinate references. (3) Cross-sheet PASS/FAIL formulas: the dashboard can read IS_NetEarnings_25 from any tab without depending on the IS layout.

Why three Bear/Base/Bull driver blocks side-by-side, plus an Active block? So all three scenarios are simultaneously visible (comparison without flipping the selector) and the Active block isolates the single set of values the model actually uses right now. Switching scenarios is one CHOOSE away. If the three blocks weren’t visible side-by-side, comparing the assumptions across scenarios would mean toggling the dropdown three times.

Why a self-contained FY2025 mini-model for the sensitivity tables? Because the main model’s revenue-growth and gross-margin “inputs” are CHOOSE-driven formula cells, and Excel’s built-in Data Tables can only substitute into direct hardcoded input cells. The engineering alternative — a self-contained FY2025 mini-model whose inputs are direct cells, with each grid cell as a formula that swaps the row/column values directly — preserves Bear/Base/Bull switchability and gives fully-live sensitivity grids. The trade-off is that the sensitivity is on a simplified FY2025 mini-model (no WC dynamics, no buyback, no NCI on FCF), not the full forecast — which is acceptable for a directional sensitivity view.

Why three two-variable grids instead of one? Each highlights a different sensitivity surface. Grid 1 (Growth × GM) shows the profit sensitivity of revenue size vs. margin quality. Grid 2 (Growth × OpEx %) shows the cost sensitivity — same revenue size, what does the cost structure do to FCF. Grid 3 (Growth × CapEx) shows the investment sensitivity — what does reinvestment intensity cost in near-term free cash flow.

Why anchor the Base case to February 2025 forward guidance? Because the alternative is “I made up the numbers and they look reasonable to me.” Tying the Base case to management’s public guidance makes the model defensible — the implicit benchmark is the same one analysts and investors were working from.

Why a year selector on the dashboard, not three side-by-side? Single-screen, executive-friendly layout. Three columns of KPIs would split attention; one column of “current year” with a YoY delta line below each KPI gives a focused scorecard. The user can flip the year to see the trajectory.

Why does the dashboard show four integrity checks instead of one? Defense in depth. Each check tests a different part of the linkage — BS balance (overall arithmetic), CF tie (the cash linkage specifically), IS-to-CF (the income statement flows correctly into operating CF), and retained-earnings consistency. Showing all four gives the reader confidence the model is consistent in multiple distinct senses, not just by accident on one metric.

21. Interview Q&A

Practice these out loud. Answers are written the way you’d actually speak.

Q1. Give me the overview of this model. “It’s a full three-statement FP&A model for Skechers, built from public SEC 10-K filings. I’ve got five years of historical actuals — 2020 through 2024 — and a three-year forecast through 2027. The forecast is driven by a switchable Bear/Base/Bull set of assumptions that I flip with one dropdown via CHOOSE/MATCH. The thesis is the DTC mix shift — Skechers’ direct channel grew from about 38.5% of revenue in 2020 to 43.1% in 2024, and DTC carries roughly 66% gross margin versus 40–43% in Wholesale, so each point of mix shift expands blended margin. I model the two segments separately to make that thesis visible. The base case lands inside the company’s February 2025 forward guidance, the balance sheet ties to zero across all three scenarios in every forecast year, and the dashboard has four live PASS/FAIL integrity checks that all read PASS.”

Q2. Walk me through the three-statement tie. How do you know it’s a real tie and not a forced plug? “I built it the harder way deliberately. The balance sheet was first built with cash as a temporary plug — every other line, working capital, PP&E, debt, retained earnings, the equity rollforward, all had to be independently correct. Then I built the cash flow statement separately, indirect method, starting from net earnings, with the working capital changes derived from the BS deltas. Then I swapped the CF ending cash into the BS cash line. Now cash is no longer a free variable. The only way Assets equals L plus E is if every other line is right. The balance check on the BS reads zero across all three forecast years in all three scenarios — and that’s a real arithmetic gate, not a forced zero. The dashboard reports it as a live PASS.”

Q3. How do you handle the circular reference between interest expense and cash? “I compute interest on beginning-of-period debt. So FY25 interest is FY24 ending debt times 5%, not FY25 average debt and not FY25 ending debt. That breaks the loop — interest no longer depends on FY25 net income, cash, or debt movement, so there’s no circular dependency. It’s the industry-standard simplification, the precision loss is negligible at our debt levels — we’re talking $50K of repayments on $455K starting balance — and the architectural benefit is that the workbook has zero circular references and doesn’t need iterative calculation enabled.”

Q4. How does the scenario switching actually work? “There’s a dropdown on the Assumptions tab in cell B5 — Bear, Base, or Bull. That feeds a MATCH formula in E5 that returns 1, 2, or 3, called ScenarioIndex. Every assumption on the Assumptions tab has its three scenario values side-by-side, and then an Active column where each cell is a CHOOSE($E$5, Bear, Base, Bull) formula. The IS, BS, CF, schedules, and dashboard all reference the Active column through named ranges. So flipping the dropdown re-drives every forecast cell in the workbook in one click. The Bear/Base/Bull recap table on the Scenario Analysis tab also pre-captures the snapshot values, because the model can only display one scenario at a time.”

Q5. Why model Wholesale and DTC separately? “Because the channel mix shift is the story. DTC gross margin is roughly 66%, Wholesale is 40–43%. Every point of mix shift toward DTC expands blended gross margin. Skechers went from 38.5% DTC in 2020 to 43.1% in 2024, and that mix shift drove most of the blended margin expansion over those years. If I modeled one blended revenue line, the thesis disappears into an averaged number. By forecasting each segment separately — with its own growth rate and its own gross margin — I make the mix-shift story explicit and auditable. There’s also a blended-GM consistency check on the IS that confirms the segment math ties to the consolidated number.”

Q6. How does the working capital forecast work? “I drive accounts receivable, inventory, and accounts payable from days ratios. AR equals sales times DSO over 365 — so 40 days of receivables on $9.76B of sales gives me about $1.07B in AR. Inventory equals COGS times DIO over 365 — 165 days. AP equals COGS times DPO over 365 — 108 days. These aren’t hardcoded values; they’re formulas that respond to the revenue and COGS forecasts. The change in net working capital from year to year automatically flows through to the indirect-method cash flow statement as operating-cash-flow adjustments.”

Q7. Why a sensitivity grid built by hand instead of Excel’s Data Tables? “Because my actual model inputs — the revenue-growth and gross-margin Active cells — are CHOOSE formulas, not direct hardcoded cells. Excel Data Tables only substitute into direct input cells, so they don’t work here. The alternative I built is a self-contained FY2025 mini-model on the Scenario tab, with its own four direct input cells — growth, GM, OpEx %, CapEx — and three two-variable grids where each cell is a formula that swaps the row and column values directly into a mini income statement. Each cell recomputes live; the whole grid updates if I change any of the base constants. The trade-off is that the sensitivity is on a simplified single-year mini-model — no working capital movement, no NCI on FCF — but for a directional sensitivity view that’s fine, and the main forecast tabs still drive everything else.”

Q8. What’s the Bear/Base/Bull spread tell you about Skechers’ operating leverage? “That it’s significant. The revenue spread across scenarios is about ±4% — Bear $9.4B, Base $9.76B, Bull $10.17B. The EPS spread is about ±40% — Bear $2.48, Base $4.40, Bull $6.18. So a small revenue swing produces a much bigger EPS swing. That’s classic operating leverage — fixed cost base that doesn’t scale linearly with revenue, so gross profit changes flow disproportionately to the bottom line. The buyback intensity amplifies it too: Bear has $150K of buybacks, Base $350K, Bull $500K, so Bull also has a lower share count to divide earnings into.”

Q9. Why anchor the Base case to forward guidance? “Two reasons. One, defensibility — saying ‘I made these numbers up and they feel reasonable’ is not a good answer in a real meeting. Saying ‘the Base case is calibrated to management’s February 2025 forward guidance’ is. Two, it forces the model to be honest. If my Base case had been $11B in revenue, I’d be implicitly betting against management — and unless I have a non-public reason to do that, I shouldn’t. Bear and Bull are still my judgment calls, but they bracket a defensible Base.”

Q10. The dashboard has four PASS/FAIL checks. Why four and not one? “Defense in depth. Each one tests a different aspect of the model integrity. The first is the overall balance check — Assets minus Liabilities plus Equity equals zero. The second is the cash flow tie specifically — CF ending cash equals BS cash. The third is that the income statement’s net earnings number is the same number flowing into the CF as the indirect-method starting line. The fourth is retained earnings consistency. They’re not all independent — if the balance check passes, retained earnings is consistent by implication — but showing all four gives the reader confidence the model is consistent in multiple distinct senses. It’s also a UX thing: a single red FAIL would be ambiguous; four green PASSes is unambiguous.”

Q11. What’s the FY2021 weirdness with the tax line? “FY2021 has a $346.8M one-time tax benefit from an intra-entity IP transfer — Skechers moved intellectual property between legal entities, which generated a one-time tax benefit under U.S. tax law. So FY2021 effective tax rate is negative — minus 43% — and net earnings and EPS spike that year. EPS was $4.73 in FY2021 versus $0.64 in FY2020. I documented it in the historical tab notes and in the README caveats, because if you just look at the FY2020-21 EPS jump without that context, the trend looks completely wrong. It’s the most important caveat in the historical data.”

Q12. What’s a ‘named range’ and why did you use so many? “A named range is a label you attach to a cell or range of cells, so you can reference it by name in formulas — Wh_Growth_25 instead of Assumptions!L13. I used about 150 of them. Three reasons. One, readability: =B8*(1+Wh_Growth_25) is obviously next-year wholesale revenue. Two, resilience: if I insert a row on the Assumptions tab, every coordinate reference would silently shift; named ranges follow the cell automatically. Three, the dashboard’s integrity check formulas can read the live model — like IS_NetEarnings_25 — without depending on the IS tab’s row layout.”

Q13. If you kept building this, what would you improve? “A few things. I’d model leases more explicitly instead of holding them flat — adding new ROU assets and lease liabilities based on store-count and DTC-buildout assumptions. I’d build sensitivity to working-capital intensity, because for an apparel/footwear company DIO is a real lever — a 10-day swing in inventory days is meaningful on $5B-plus of COGS. I’d add a discounted cash flow valuation tab as a logical extension of the FCF forecast. And I’d build a stochastic layer — Monte Carlo over the key drivers — to produce a probability distribution of FY25 EPS rather than three discrete scenarios. Long term, this is a planning model; with a DCF and probabilistic layer it would also be a valuation model.”

Q14. Why Excel and not Python for this? “Deliberate. FP&A teams run on Excel. Even teams that build big things in Python ship the deliverable to leadership in Excel because that’s where the audience lives. This model is meant to read like real FP&A work — a senior person can open it, click through the Assumptions tab, change the dropdown, and watch the dashboard re-drive. That’s the right artifact for the role. I also use Python to build the file via openpyxl as part of the development workflow, but the deliverable is Excel.”

Q15. How do you know your forecast numbers are reasonable, not just internally consistent? “Three calibration anchors. First, the Base-case FY2025 revenue is $9.76B and EPS is $4.40 — both inside Skechers’ February 2025 forward guidance, so the Base lines up with what management itself was telling the market. Second, the working capital drivers — DSO 40, DIO 165, DPO 108 — are calibrated to FY2024 actuals on the historical tab, where I computed them from the as-reported balance sheet and income statement. Third, the operating-margin trajectory of 10.4% to 10.6% across the forecast is consistent with the historical trend of 10.1% in FY2024 — modest, plausible expansion driven by mix shift, not a step-change. Internal consistency proves the model arithmetic; calibration to actuals and guidance proves the inputs are reasonable.”

22. How to Walk Through This Project Live

If asked to screen-share the workbook, use this order. About 8–10 minutes.

1. Open the README first. It’s the front door. Point out the model map, the color legend, and especially the live integrity status table with all PASSes — “this is the headline: the model balances and ties across every forecast year.”
2. State the thesis — “Skechers has two segments, Wholesale and DTC. DTC is roughly twice the gross margin of Wholesale and has been growing as a share of revenue. This model forecasts each segment separately so the mix-shift margin story is visible.”
3. Open the Historical Financials tab. Walk the segment block briefly: DTC mix going from 38.5% to 43.1%, DTC gross margin around 66% vs. Wholesale at 40–43%. Mention the FY2021 tax anomaly if asked about EPS volatility.
4. Open the Assumptions tab. Show the dropdown on B5. Flip it once Bear → Base → Bull and let the watcher see the Active column on the right change. Point out that the working-capital ratios and tax rate are held flat — those are structural — while growth, margin, OpEx, and CapEx vary.
5. Show the Income Statement Forecast. Point out the segment build at the top — WH revenue × (1 + Wh_Growth), DTC revenue × (1 + DTC_Growth), the COGS lines using the segment GMs, and the consolidated build below. Point out the blended-GM check at the bottom (= 0).
6. Open Supporting Schedules. Spend a moment on the interest-expense row — “this is on beginning-of-period debt, which is the standard simplification that breaks the circular reference. No iterative calculation needed, no circular warning.”
7. Open the Cash Flow Forecast. Show the indirect-method structure. Show the ending cash row. Show the tie check below — “this row equals zero because CF ending cash is what the BS cash line points at.”
8. Open the Balance Sheet Forecast. Point at the cash line — “this links to CF ending cash.” Then point at the balance check at row 38 — zero across all three forecast years. “That’s a real test: cash is no longer a plug, so the only way that’s zero is if every other line is right.”
9. Open the Scenario Analysis tab. Show the recap table — Bear/Base/Bull side by side. Then scroll to the sensitivity grids. “Each grid cell is a formula, not a static value. I built a self-contained FY2025 mini-model on this tab so the sensitivity stays live even though the main model’s inputs are CHOOSE formulas that Excel’s built-in Data Tables can’t drive.”
10. End on the KPI Dashboard. Flip the year selector once (FY2025 → FY2026 → FY2027) to show the scorecard re-drive. Show the four-row integrity table — “all PASS, every year, every scenario.” If time permits, flip the scenario back on Assumptions!B5 and watch the dashboard re-compute in real time.

Pacing tip: spend the most time on Assumptions + the three statement forecast tabs + the integrity story. The dashboard is the wow factor to close on.

23. Glossary

• 3-statement model — an income statement, balance sheet, and cash flow statement that fully articulate, so a change to any driver flows correctly through all three.
• Active block — the column block (L:N) on the Assumptions tab whose cells are CHOOSE formulas selecting one of the three scenario values. The whole model reads from this block.
• Articulate / articulation — when the three financial statements link to each other correctly: net income flows from IS to CF and to retained earnings on BS; cash on BS equals CF ending cash; WC changes on BS appear as adjustments on CF.
• BASE rollforward — the standard schedule structure: Beginning + Additions − Subtractions = Ending. Used here for PP&E, debt, and shares.
• Bear / Base / Bull — three scenarios capturing a pessimistic, central, and optimistic view of the forecast drivers.
• Beginning-of-period interest — computing interest expense on the debt balance at the start of the year, not the average or ending. Breaks the interest → NI → cash → debt circular loop.
• Blended gross margin — total gross profit ÷ total revenue, across all segments.
• CapEx — capital expenditures; cash outflow to acquire long-lived assets (PP&E).
• Cash conversion cycle (CCC) — DSO + DIO − DPO; days between paying for inputs and receiving cash from sales.
• CHOOSE — Excel function that returns one of a list of values based on an index. Used with MATCH for the scenario switch.
• D&A — depreciation and amortization; non-cash expense recognizing PP&E and intangible-asset consumption.
• DSO / DIO / DPO — Days Sales Outstanding (AR ÷ Sales × 365), Days Inventory Outstanding (Inv ÷ COGS × 365), Days Payable Outstanding (AP ÷ COGS × 365).
• DTC (Direct-to-Consumer) — Skechers’ direct channel: e-commerce and company-owned retail stores. Higher gross margin than Wholesale.
• EBIT / EBT / Net Earnings — Earnings Before Interest and Tax / Before Tax / After Tax (and before/after NCI).
• EBITDA — Earnings Before Interest, Tax, Depreciation, and Amortization. Not a primary line in this model but implicitly = EBIT + D&A.
• EPS (Diluted) — Net Earnings (SKX) ÷ Diluted Shares Outstanding.
• Free Cash Flow (FCF) — Operating Cash Flow + CapEx (CapEx is negative, so this is OCF − |CapEx|). Cash left after maintenance investment.
• Forecast period / projection period — FY2025–FY2027 in this model.
• G&A — General & Administrative expenses; one of two operating expense buckets.
• Indirect method — the most common way to build the operating section of the cash flow statement: start from net income, add back non-cash items (D&A, SBC), adjust for working-capital changes.
• Integrity check — a formula that verifies the model is internally consistent (e.g., Assets = L+E, CF ending cash = BS cash).
• MATCH — Excel function that returns the position of a value in an array. Used here to convert the scenario name to an index (Bear → 1, Base → 2, Bull → 3).
• Named range — a label attached to a cell or range; can be used in formulas in place of the coordinate reference.
• NCI (Non-Controlling Interest) — the portion of a subsidiary’s earnings/equity owned by minority shareholders, not the parent. Held at 12% of net earnings in this model.
• OCF / ICF / FCF (CF sections) — Operating, Investing, and Financing Cash Flow.
• OpEx — Operating Expenses; here = Selling + G&A.
• Operating Leverage — the degree to which a small change in revenue produces a larger change in operating income (and EPS), driven by a fixed-cost base.
• PASS/FAIL gate — an IF(ABS(x)<tolerance,"PASS","FAIL") formula on a check line.
• Plug — a balancing variable. In an early draft of the model, cash was the plug; it was replaced by a calculated value from CF.
• PP&E — Property, Plant, and Equipment. Long-lived productive assets, net of accumulated depreciation.
• Retained Earnings (RE) — the cumulative portion of past net earnings retained in the business rather than paid out. Rolls forward as Prior RE + Net Earnings − Dividends (Skechers pays no dividend in this period).
• ROU (Right-of-Use) Assets — the operating-lease asset under ASC 842 lease accounting.
• SBC (Stock-Based Compensation) — non-cash expense for employee stock awards. Added back on CF; credited to paid-in capital on BS.
• Segment / Reportable Segment — operating divisions a company discloses separately. Skechers reports Wholesale and DTC.
• Sensitivity table / Data Table — a two-input grid showing how an output changes as two drivers vary. Here, built as a self-contained mini-model with formula-filled grids.
• Tax benefit / one-time item — a non-recurring item (e.g., FY2021’s $346.8M tax benefit from an intra-entity IP transfer) that distorts a single year’s effective tax rate.
• Three-statement tie — the property that the IS, BS, and CF all articulate consistently — a change anywhere flows correctly to everywhere.
• Wholesale — Skechers’ indirect channel: sales to retailers, distributors, and third-party platforms. Lower gross margin than DTC.
• Working capital (WC) — Current Assets − Current Liabilities (a balance sheet concept); the operating-cash investment required to run the business.
• XBRL — eXtensible Business Reporting Language; the structured-data format SEC filings use, which makes filings machine-readable.

This study guide documents the model as built. The authoritative reference is the workbook excel/SKX_Financial_Analysis_Project_Final.xlsx — its Assumptions dropdown, the Active block, the IS/BS/CF forecast tabs, the Scenario Analysis sensitivity grids, and the KPI Dashboard’s live integrity checks. When this guide and the workbook disagree, the workbook wins.