27 The Lumen Capstone Ties The Course Together.
The course capstone: build the linked three-statement model for Lumen Tutoring Co. end to end, produce an investor return, and defend a decision against the same case every module used.
27 The Lumen Capstone Ties The Course Together
Every module since 22 has reserved its hardest deliverable for this page. Module 22 reserved the linked summary table, 23 the combined-case sensitivity, 24 the real attribution, 25 the joint downside, and 26 the full simulation. This is where those reservations come due. The capstone is one build: the linked Lumen model, from drivers to decision, checked against the same case the course has used throughout.
1. What the capstone asks
Take Lumen Tutoring Co. from the case primer to a defensible answer. Starting inputs are given below — operating drivers, FY24A actuals, working-capital conventions, capex, the financing stack and its covenants, and the deal terms. From those inputs, produce three things:
- A reconciled model. The three-statement forecast FY25E–FY27E, built the way module 03 and module 21 require: the balance sheet ties every period, the cash roll ties to the cash flow statement, and the debt schedule rolls without plugs.
- An investor return. The equity story under the course deal terms — entry at 5.5x EV/EBITDA on the FY24A basis, exit underwritten at 6.5x in FY29P, five-year hold — measured the way module 06 and module 17 measure it: MOIC and IRR on the sponsor check, with the decomposition behind the number.
- A justified decision. One sentence, supported by the summary table and the downside views — not a verdict about whether Lumen is a good company, but whether the deal clears its own stated floors: 20% sponsor IRR and 1.0x MOIC.
2. Starting inputs
Everything below is in the shared case file. If your numbers and these inputs disagree, the inputs win — the point of the capstone is that every module worked the same company.
Operating drivers
Lumen runs 24 sites at FY24A end, adding 6 per year, with 6,800 active students growing to 11,800 by FY27E. Sessions price off a $78 blended hourly rate at 6.2 sessions per student per month, tutor utilization ramps to 62%, and the revenue mix runs 34% online / 66% in-person with 18% annual student attrition. Sites reach steady state — $2.4M revenue at a 24% EBITDA margin — on a 24-month ramp curve.
Historicals and forecast anchors (USD thousands)
FY24A closes at revenue 48,200, EBITDA 8,200 (17.0% margin), net income 4,050, capex 1,900, FCF 5,400. The published path then runs revenue 58,800 / 71,600 / 86,400 and EBITDA 11,000 / 14,100 / 17,600 across FY25E–FY27E — margin expands 70bps a year as the site base matures into steady state.
Working capital, capex, and tax
Subscriptions keep working capital short: 15-day DSO, 30-day DPO, and a 60-day deferred-revenue balance — packages billed in advance are the swing item, not receivables. Inventory is zero. Capex steps 2,400 / 2,800 / 3,200 with the site build schedule; book tax runs at a flat 25% statutory rate.
Financing and covenants
The capital structure at FY24A: a $15.0M senior term loan at 8.5% amortizing $800K per year, a $2.0M revolver drawn against an $8.0M facility (SOFR + 350, borrowing base 85% of eligible AR plus 50% of unbilled work), and $28.0M of equity invested across the founder and two PE rounds. The covenant set: senior leverage ≤ 3.5x, fixed-charge coverage ≥ 1.25x, minimum liquidity $3.0M — the tests module 05 taught you to wire as check rows, not prose.
Deal terms
Entry: 5.5x EV/EBITDA on the FY24A basis of 8,200 — an implied EV of 45,100, less 13,000 of net debt assumed, is a 32,100 equity purchase price, with the sponsor writing a 10,000 check. Exit: FY29P at 6.5x on the extended forecast, with net debt at exit of 8,915. The base underwrite the course has carried is a 22.4% sponsor IRR; the decision floors are 20% IRR and 1.0x MOIC.
3. What the model must reconcile
The capstone grade is the check strip, not the prose. Before any return is quoted, the model must show:
- Balance sheet ties. Assets equal liabilities plus equity in every period — FY24A opening and FY25E–FY27E. The deferred-revenue plug carries the subscription float; it does not become a balancing line.
- Cash roll ties. Ending cash on the balance sheet equals the cash flow statement's ending cash: 7,863 in FY25E, 12,451 in FY26E — including the ~50%-of-net-income distribution policy the case assumes.
- Debt schedule rolls. The term loan steps down 800 a year (15,000 → 14,200 → 13,400 → 12,600) with interest on the right balance convention, and the revolver holds flat at 2,000 unless liquidity forces a draw.
- Covenant rows compute. Senior leverage, fixed-charge coverage, and minimum liquidity render every period against the actual covenant definitions — FY25E senior leverage comes out near 1.4x against the 3.5x ceiling, and liquidity (cash plus undrawn revolver availability) sits far above the 3,000 floor in the base case.
- No plugs. Any cell you typed to force a tie is a finding, not a fix — the module 17 checklist's order applies here: identities first, assumptions second.
4. The decision packet
The deliverable is one page, and it uses the output surfaces the course built:
- The summary table (module 22): the rows a decision-maker reads, tied to the model — revenue, EBITDA, net income, FCF, net debt, covenant headroom.
- The sensitivity table (module 23): IRR across entry price and exit multiple, centered on the 5.5x / 6.5x underwrite.
- The decomposition (module 24): how much of the return is growth, how much is deleveraging, how much is multiple arbitrage — the 1.0x entry-to-exit delta is the honest number to defend.
- The tornado and simulation (25, 26): the ranked drivers and the probability the deal misses its 20% floor when ramp, pricing, and exit multiple miss together.
The decision sentence comes last and is one line: does this deal clear its floors with headroom, and what has to be true for that to hold.
Before you build, sanity-check the entry math by hand: 8,200 × 5.5x = 45,100 EV; 45,100 − 13,000 net debt = 32,100 equity; sponsor check 10,000. If a 22.4% IRR holds over five years, what does the sponsor's 10,000 come back as? Work it before opening the model — the capstone's first check is whether you can read the deal before it is dressed in cells.
Check your work
Sponsor proceeds ≈ 10,000 × 1.2245 ≈ 27,500 — roughly a 2.7x MOIC. Both floors clear on the back of an envelope: 22.4% > 20%, and 2.7x > 1.0x by a mile. The point of the exercise: three numbers anyone can recompute in a minute already tell you the deal clears — the model's job is to prove the path, not to discover the direction.
5. Answer key — the reconciled path
Compare your build against the published case figures. These are the numbers the component lessons reconciled to; a capstone model that produces different ones is either modeling a different company or carrying an error the check strip should have caught.
| Line | FY24A | FY25E | FY26E | FY27E |
|---|---|---|---|---|
| Revenue | 48,200 | 58,800 | 71,600 | 86,400 |
| EBITDA | 8,200 | 11,000 | 14,100 | 17,600 |
| EBITDA margin | 17.0% | 18.7% | 19.7% | 20.4% |
| Net income | 4,050 | 5,813 | 7,838 | 10,163 |
| Capex | 1,900 | 2,400 | 2,800 | 3,200 |
| Free cash flow | 5,400 | 7,100 | 9,300 | 11,700 |
| Senior term loan (EOY) | 15,000 | 14,200 | 13,400 | 12,600 |
| Revolver drawn (EOY) | 2,000 | 2,000 | 2,000 | 2,000 |
| Ending cash | 4,500 | 7,863 | 12,451 | — |
| Sites / active students (EOY) | 24 / 6,800 | 30 / 8,200 | 36 / 9,900 | 42 / 11,800 |
Return check. Entry equity 32,100 on a 45,100 EV; base sponsor IRR 22.4% over the five-year hold; exit at 6.5x on FY29P EBITDA less 8,915 net debt. Against the floors — 20% IRR and 1.0x MOIC — the base clears with ~2.4pts of IRR headroom and roughly 2.7x on the sponsor check.
Decision check. The justified answer in the base case: the deal clears both floors, but the headroom lives in the extension years — the FY28P/FY29P EBITDA the exit multiple multiplies is a deal-team assumption, not a model output, and the 1.0x multiple arbitrage is a third of the thesis. A reviewer who defends those two lines has earned the yes; a reviewer who cannot name what drives them has only read the model back to you.
Run the changed-assumption test on the exit basis, because it is where the capstone grade actually lives. Base case: FY29P EBITDA extended from the published ramp — margin steady near FY27E's 20.4%, sites fully ramped — underwritten at 6.5x for the 22.4% sponsor IRR. Altered case: exit EBITDA held flat at FY27E (17,600 × 6.5x = 114,400 EV, less 8,915 net debt = 105,485 total equity proceeds against the 32,100 entry) — no extension growth at all. Held constant: entry terms, debt schedule, distribution policy, hold length. The reviewer's read: total equity proceeds still run ~3.3x the entry equity, so the deal clears its floors even with the extension years removed — but if a student's version kills the deal under this test, the underwrite was the extension, not the business, and the summary table should have shown that.
Closing
This is the end of the build. Everything before it was a part; the capstone is the whole. The model reconciles or it does not, the return clears its floors or it does not, and the decision sentence either survives the downside views or it does not. When all three hold, the course has done its job — you built the file you would hand to an IC, and you can defend every line on it.