Work one real month-end close — messy variances, a stakeholder inbox, numbers that reconcile to the cent. Then submit the commentary you'd actually send the CFO and have it scored against a published rubric.Work one real month-end close, then have the commentary you'd send the CFO scored against a published rubric.
You're on a phone — and the case needs a spreadsheet. Send it to your inbox and it'll be waiting when you're at a laptop.
✅ On its way — check your inbox in a minute.
Something went wrong — email hello@fpacareerlab.com and we'll send it by hand.
Free, no account. Both cases, in one email.
Free. No account. Nothing to unlock. The full case takes 60–90 minutes.
An EBITDA bridge, a CFO commentary and a forecast, on a real messy month. The specific example you've never had when an interviewer asks for one.
Scored on five published dimensions, with every driver you found, every one you missed, and exactly where a stronger answer would have gone.
Generated from what you actually wrote. The follow-ups your commentary invites — before someone senior asks them for real.
Two numbers you're given. Five you have to find. Revenue came in on plan, so nothing on the P&L explains the gap. The answer is in the GL, the weekly cut and an inbox where one person is confidently wrong.
Before the full case, here is the whole idea in one question. Same month, same company, four colleagues reading the same numbers. Every statement below contains arithmetic that checks out — but one of them draws a conclusion the data does not support.
| Line | Oct | Nov |
|---|---|---|
| Revenue | €900k | €900k |
| Units shipped | 12,000 | 14,400 |
| Revenue / unit | €75.00 | €62.50 |
| Gross profit | €270k | €243k |
| Gross margin | 30.0% | 27.0% |
Everything here ties. 900,000 ÷ 14,400 = 62.50. 243 ÷ 900 = 27.0%. Check any of it.
Pick one. You'll get the answer straight away — there's no sign-up and nothing is being collected.
A is the one that costs you. Not because the arithmetic is wrong — revenue really was flat — but because "flat" is the disguise, not the result. The company shipped 20% more units to stand still, at €12.50 less per unit, and gave up €27k of gross profit doing it. B, C and D each state something the table proves. A states something the table proves and then adds a word the table cannot support: steady.
That gap — between a number being correct and a conclusion being earned — is the entire skill. It is what a CFO probes in the follow-up question, it is what no course teaches, and it is what gets scored here. A confident, well-written commentary built on A would read beautifully and be wrong.
That was 60 seconds and five rows. Case 01 is the same skill across a full month — eight drivers, a stakeholder inbox, and a scored review at the end.
Start Case 01 — free →Finance careers stall in a specific, boring way: you can operate the tools, but you've never been handed a messy month and asked what happened. There's no safe place to practise that — so people buy another course, learn more tools, and stay stuck. Every case here is built from real finance workflows by a finance professional, not by a content team.
Every case is a different company in a different industry with its own way of going wrong — logistics, distribution, SaaS, hospitality. You come back and work another close, the way a musician plays another piece. One case teaches you a story. Twenty build judgment.
Every figure is validated before publication. The bridges reconcile exactly, the GL ties to the P&L, weekly revenue sums to the month. Rebuild any of it from the raw data and it ties. You can't practise judgment on numbers you can't trust.
Fluent, confident, professional-sounding analysis that names no cause scores badly here, on purpose. Claims you can't prove from the data are penalised even when they happen to be right. That's the habit a CFO review punishes and no course corrects.
You're the FP&A analyst at a B2B logistics company. November is closed. Commercially everyone is relaxed — revenue came in on plan. The CFO wants to know what's really going on underneath that flat revenue line, and the answer is not in the P&L.
| Line | Budget | Actual | Var |
|---|---|---|---|
| Revenue | €4,200k | €4,215k | +€15k |
| Gross profit | €1,176k | €1,140k | −€36k |
| Gross margin | 28.00% | 27.05% | −0.95pp |
| Warehouse payroll | €380k | €402k | +€22k |
| Maintenance & repairs | €150k | €170k | +€20k |
| Admin & IT | €210k | €228k | +€18k |
| Sales & office | €180k | €180k | +€0k |
| EBITDA | €256k | €160k | −€96k |
| On-time delivery | 96.5% | 93.8% | −2.7pp |
Analyse the variances, identify what actually drove them, and update the December outlook. Work in your own spreadsheet — the full data pack is below: GL account detail with owners, revenue by week × channel, client-level contract revenue, and a four-month trend.
Reading this on a phone? Send the pack to your inbox instead.
✅ Sent to your inbox — and here it is right now:
Something went wrong — try again, or email hello@fpacareerlab.com.
Some of it matters. Some of it doesn't. One of them is confidently wrong.
As discussed on the ops call: the main conveyor at warehouse 2 was down for 4 days in the second week. We switched to manual handling — a lot of overtime for the crew — and brought in an external contractor for the repair, roughly €20k. A number of outbound shipments were missed or delayed during those days. Line should be stable now, vendor did a full recalibration.
Commercially November was a solid month — revenue basically on plan. Meridian Retail even asked us to ship part of their December peak volumes early, which we handled without issues; great service story for the account review. Nothing to flag from my side.
To confirm: the TMS implementation remains fully within the total budget approved by the steering committee. Roughly €18k of consultant invoices originally scheduled for December landed in November — procurement agreed to the earlier billing. December project spend will be correspondingly lower. No overruns expected on the project as a whole.
Heads up for the close: Meridian Retail asked us to ship roughly €100k of their December peak volumes in the last two weeks of November — their DC wanted stock early. We invoiced on shipment as per contract. Their December call-offs will be correspondingly lighter.
Diesel is up around 9% since late October — you'll see it in direct transport costs. For contract clients the fuel surcharge clauses kick in automatically, so most of it comes back on the revenue side. Spot freight has no such protection, but that's a smaller share of volume.
You'll get a graded review against the public rubric — dimension scores, the drivers you caught, the ones you missed, and where a stronger answer would have gone. Within 24 hours.
Free. No account. Four text boxes.
Tell us where. Where the pack was confusing, where the case was too easy or too cruel, or a number you think doesn't tie. If you find a break, we publish the correction. Three sentences is plenty and no email is needed.
✅ Got it — thank you. This genuinely changes what gets built next.
Didn't send — email hello@fpacareerlab.com instead and it reaches the same place.
Below is a genuine review of a genuine submission on this exact case — not a mockup. Seven drivers found, one discipline gap flagged, three ways the answer could have been stronger, and the three questions a CFO would have asked next.
Diesel as a margin driver. "Contract customers have surcharge protection, but spot freight does not."
Surcharge recovery through revenue — and why that recovery still didn't protect margin.
The Meridian pull-forward as a one-off that reverses. "…approximately €85k of favorable variance in W3 and W4, when part of Meridian Retail's December peak volume was shipped early."
€45k of warehousing revenue lost to the outage, concentrated in W2.
€22k of overtime and €20k of emergency repair traced to the same conveyor failure.
TMS spend classified as timing, not overrun — the approved-decision reading, which is the correct one.
And the connection that earns the marks: "This is one connected operational event, not three unrelated variances."
Discipline gap. The inbox states Meridian's pulled-forward volume was "roughly €100k". The bridge consistently uses €85k, computed from the W3+W4 variance, and never reconciles the ~€15k gap — which then propagates into the December exposure figure.
Correctly not penalised: one driver was unprovable from this pack — the client-level detail simply isn't there, and the commercial-summary email is a red herring. The analyst was right not to assert it, and the review says so rather than marking it missed.
Credited hedges, quoted by name. "I would avoid assigning the entire €36k margin miss to fuel, because the pack does not provide a direct transport-cost breakdown." — naming your evidence gap earns marks here. Guessing confidently loses them.
Reconcile the €100k the account manager stated against the €85k computed from the weekly table, and explain the gap rather than silently substituting one number for the other.
Use the four-month trend (payroll 381 → 380 → 377 → 402) to evidence that November's spikes are anomalous against a stable run rate, instead of asserting they're one-off.
Explicitly request client-level revenue detail to test whether the early-month softness is one account eroding or broad-based — rather than leaving it as "underlying weakness".
"Can you reconcile the €100k pull-forward the Key Account Manager cited against the €85k you computed from W3–W4 contract-logistics variance — where does the ~€15k gap sit?"
"You flag underlying weakness in W1–W2 contract logistics before the pull-forward — do we have client-level revenue detail to tell us whether that's one account eroding or broad softness?"
"Within the €51k increase in direct operating costs, what portion is diesel specifically versus the conveyor disruption?"
The same case, graded again — this time a polished, confident, perfectly formatted write-up that named no cause. Communication scored 7/10. Everything that measures analysis collapsed. Fluency is not analysis, and the grader is built not to buy it. (A deliberately constructed control, not anyone's real work, and run on a smaller model than the 88.)
An EBITDA bridge that doesn't reconcile caps the total at 60/100 regardless of how good everything else is. And investigating immaterial lines costs marks — knowing what not to analyse is part of the job.
Case 01 is free and always will be. The library grows from there — each case a different industry, and a different way for a month to go wrong.
Road freight & warehousing. Revenue on budget, EBITDA −€96k, and an inbox that half-explains it.
Start it now →Revenue beat budget by €180k and EBITDA still fell €113k. Discounting, channel mix and rebate accruals, all fighting each other.
Start it now →The month the net revenue retention story stops matching the cash. Deferred revenue, churn timing and a hiring plan that ran ahead of plan.
The full pack — workbook, five data files, stakeholder inbox — and a full scored review against the same public rubric. No payment, no card, nothing to unlock.
What we want instead is your written feedback: where the pack was unclear, where the case was too easy or too cruel, and where the grading got you wrong. That's worth more to us right now than €19 is. It won't be free forever — while the case is new, this is the trade.
✅ You're on the list — you'll hear from us the day it opens.
Case 01 and Case 02 are both open and both free — start either one above, no email required to read them. Case 03 is the month a SaaS company's retention story stops matching its cash. It gets a date when we're confident in it and not before; we'd rather leave it undated than invent one. Unsubscribe anytime.
Short answers, and no marketing in any of them. If something here reads like a dodge, tell us and we'll rewrite it.
Yes. No account, no card, nothing gated. We would rather you work a full case and judge the standard yourself than take our word for it.
Finance people who can operate the tools but have never been handed a messy month and asked what happened. Students and graduates targeting FP&A, accountants moving across, analysts in adjacent seats, and anyone who has been asked "tell me about a forecast you built" and had nothing specific to say.
A written review against the five published rubric dimensions: your score on each, the drivers you correctly identified, the ones you missed, and where a stronger answer would have gone. Typically within 24 hours.
A spreadsheet of some kind. Excel, Google Sheets and Numbers all work, and the pack is a standard .xlsx. The case involves real spreadsheet work, so it is not practical on a phone — use the email option and pick it up at a laptop.
Check them. Every figure in every pack is validated before publication, and the bridges reconcile exactly from the raw data. Nothing here is illustrative or rounded to make a point. If you find a break, email it to us and we'll publish the correction.
Reviews are produced against the published rubric by a grading system calibrated on submissions from a working FP&A professional — including one run where it flagged a genuine error in our own reference answer. Every review is checked before it is sent.
No. It gives you worked examples you can discuss specifically in an interview, and an honest read on where your analysis is currently weak. Anyone promising more than that is selling you something.
Realistic month-end closes where every number reconciles, and a grader that scores judgment rather than fluency.