To quantify an operational improvement plan in a private equity interview, work in a fixed order: establish entry EBITDA, measure working capital in days, size the one-time cash release, size each recurring earnings lever separately, build an EBITDA bridge that excludes the cash release, then carry the result through net debt to an equity value, a money multiple and an IRR. The order matters more than the arithmetic. Candidates rarely fail this question because they cannot multiply; they fail it because they mix a cash effect into an earnings bridge, or because they cannot defend a single assumption when pushed.

This walkthrough uses the figures from the Operational Improvement case study, a €400m-revenue industrial heating systems business being carved out of a listed group at a 3.0% EBITDA margin. If you want the conceptual grounding on what each lever does before you calculate anything, read what operational improvement in private equity actually means first.

Step 1: Establish Entry EBITDA and the Margin Gap

Start with the number the deal is priced off. From the vendor due diligence pack: revenue €400.0m, materials and bought-in components €180.0m, direct labour €70.0m, other production and logistics €50.0m, SG&A including allocated group overhead €88.0m.

EBITDA = €400.0m − €180.0m − €70.0m − €50.0m − €88.0m = €12.0m, a 3.0% EBITDA margin.

Benchmarking the Margin Gap

Say the margin out loud and compare it to peers. Listed industrial peers in this space earn 9% to 12%, so there is a six to nine percentage point gap. That single sentence establishes the shape of your entire answer: the asset is cheap because it underperforms, and the return has to come from closing the gap. If you are unsure why EBITDA rather than net income is the reference metric here, what EBITDA is and why interviewers ask about it explains why capital structure and tax jurisdiction are deliberately stripped out.

Also flag what is inside SG&A. In a carve-out, allocated group overhead is a cost the standalone business will not carry in the same form — some of it disappears, some of it has to be rebuilt, and the difference is a stand-alone cost adjustment that belongs in the plan. This is the same class of adjustment covered in EBITDA normalisation and non-recurring items.

Step 2: Convert Working Capital Into Days

Balance sheet stocks are not comparable across companies; days are. Convert each line against its natural driver, using a 365-day year and a cost of goods sold base of €300.0m (materials plus direct labour plus other production costs).

  • DSO = €80.0m / €400.0m × 365 = 73.0 days
  • DIO = €65.0m / €300.0m × 365 = 79.1 days
  • DPO = €40.0m / €300.0m × 365 = 48.7 days
  • Net working capital = €80.0m + €65.0m − €40.0m = €105.0m, or 26.3% of revenue

Choosing the Right Driver

Note the driver choice: receivables run against revenue because that is what customers owe you against, while inventory and payables run against cost of goods sold because that is what they are carried at. Getting this wrong is a small error with a large downstream effect. The mechanics are drilled in how to calculate working capital in an interview using DSO, DIO and DPO.

Then read the pattern aloud, because it is diagnostic. Collecting in 73 days, holding 79 days of stock and paying suppliers in 49 days is the signature of a division that was never measured on cash: the parent group managed cash centrally and judged the division on revenue and margin. That observation tells the interviewer you know why the opportunity exists, not just that it exists.

Step 3: Size the One-Time Cash Release

Apply benchmarked target days to unchanged drivers. Commercial due diligence set targets of 55.0 days DSO, 60.0 days DIO and 65.0 days DPO.

Target balance = driver × target days / 365.

  • Target receivables = €400.0m × 55 / 365 = €60.3m, releasing €80.0m − €60.3m = €19.7m
  • Target inventory = €300.0m × 60 / 365 = €49.3m, releasing €65.0m − €49.3m = €15.7m
  • Target payables = €300.0m × 65 / 365 = €53.4m, releasing €53.4m − €40.0m = €13.4m

Total one-time release = €48.8m. New net working capital = €60.3m + €49.3m − €53.4m = €56.2m, or 14.1% of revenue.

What to Say After the Number

Two things to say immediately after producing this number. First, €48.8m against a €60.0m purchase price means the balance sheet very nearly funds the acquisition — which is exactly why working-capital-heavy carve-outs attract turnaround sponsors. Second, this is cash, it is one-time, and it will not appear in your EBITDA bridge. Saying so before the interviewer asks is worth more than saying it correctly when challenged.

Add the caveat on payables. Extending suppliers from 48.7 to 65.0 days on a business that has just changed owner can trigger tighter credit terms or a price response, which lands later as a procurement cost. A candidate who models the payables release at full value with no acknowledgement of that dynamic is modelling a spreadsheet, not a business.

Step 4: Size the Procurement Lever

Procurement savings = total spend × addressable share × savings rate.

With materials and bought-in components of €180.0m, an addressable share of 70% and a negotiated savings rate of 6.0%:

Addressable spend = €180.0m × 0.70 = €126.0m. Savings = €126.0m × 0.060 = €7.6m.

Defend both inputs before you are asked. The addressable share comes from a spend cube built in due diligence, splitting spend by supplier and category and carving out anything sole-sourced, single-qualified for regulatory reasons, or under a contract that does not expire inside the plan period. Landing at 70% implies roughly €54m of locked spend, which is normal for an industrial manufacturer with qualified components. The savings rate is benchmarked from prior carve-outs and from a pilot tender on one category; 6.0% is defensible for a business that has never tendered independently, because the parent's central purchasing function bundled it with much larger units and never optimised for it.

The Addressable-Spend Trap

The trap here is applying the savings rate to total spend rather than addressable spend. On these figures that would produce €10.8m instead of €7.6m — a 42% overstatement of the lever, delivered with complete confidence. Interviewers watch for it specifically.

Step 5: Size the Headcount Lever

Headcount savings = positions removed × fully loaded cost per employee.

With 180 positions removed from indirect and administrative functions at a fully loaded cost of €72,000: 180 × €72,000 = €13.0m.

Fully Loaded Cost and the Severance Bill

Three points to make alongside the number. The 180 positions are 7.5% of the 2,400 total headcount and sit entirely in functions duplicated by the carve-out, so production output and therefore revenue are unaffected. Fully loaded cost includes employer social contributions, pension provision and facilities, not just gross salary — using salary alone would understate the saving by roughly a quarter in a German industrial setting. And the saving has to be bought: a €15.0m restructuring provision covering severance, works council process and one-time implementation cost is funded in cash at closing, roughly twelve months before full run-rate benefit arrives.

That funding requirement is why the restructuring provision appears as a use of funds alongside the purchase price rather than being absorbed by operating cash flow. If the sources and uses structure is unfamiliar, what a sources and uses table is sets it out.

Step 6: Size the Pricing Lever Correctly

This is where the arithmetic asymmetry lives. A pricing diagnostic found the spare parts and service book priced off an outdated list. Affected revenue is 40% of the top line, the price increase is 3.5%, expected volume attrition is 2.0% of affected revenue, and contribution margin on lost volume is 25%.

  • Affected revenue = €400.0m × 0.40 = €160.0m
  • Gross price uplift = €160.0m × 0.035 = €5.6m, all of which reaches EBITDA
  • Lost revenue = €160.0m × 0.020 = €3.2m
  • EBITDA lost on attrition = €3.2m × 0.25 = €0.8m
  • Net pricing effect = €5.6m − €0.8m = €4.8m

Pro-forma revenue = €400.0m + €5.6m − €3.2m = €402.4m.

Applying the Contribution Margin

The step candidates get wrong is deducting the full €3.2m of lost revenue from EBITDA instead of the €0.8m contribution margin on it. Volume you no longer serve also no longer consumes materials, labour or freight. Because a price increase carries a 100% margin while the volume it costs you leaves at only the contribution margin, this pricing action would tolerate more than seven times the assumed attrition before it turned EBITDA-negative — which is the substantive point about pricing worth making in an interview.

Be ready for the commercial follow-up too: why the aftermarket? Because installed-base customers buying replacement parts face high switching costs and the part price is small relative to the cost of machine downtime. That is where pricing power sits in industrials, and identifying it is a commercial due diligence output rather than a modelling exercise.

Step 7: Build the EBITDA Bridge, and Leave the Cash Out

Run-rate EBITDA = entry EBITDA + procurement + headcount + pricing.

€12.0m + €7.6m + €13.0m + €4.8m = €37.4m, a 9.3% margin on pro-forma revenue of €402.4m.

The €48.8m working capital release enters this bridge at €0.0m. It is the largest single number in the whole plan and it contributes nothing to EBITDA. Putting it in the bridge means it also gets valued at the exit multiple, turning €48.8m of one-time cash into €244m of phantom enterprise value. This is the error that most often ends an operational improvement interview, and it is why the plan is always presented as two exhibits: an EBITDA bridge for the recurring earnings levers and a net debt walk for the balance sheet effects.

Sanity-check the endpoint before moving on. Taking the margin from 3.0% to 9.3% lands the business at the bottom of the 9% to 12% peer range, not above it. That is deliberate: a plan projecting best-in-class margins for a business that has never managed its own cost base will not survive an investment committee, and volunteering the peer comparison shows you know it.

Step 8: Carry the Plan Through to a Return

Now translate operations into an equity outcome. Entry multiple 5.0x, debt raised €20.0m, three-year hold, exit at the same 5.0x so that no multiple expansion is credited.

  • Entry enterprise value = 5.0x × €12.0m = €60.0m
  • Total uses = €60.0m purchase price + €15.0m restructuring + €5.0m fees = €80.0m
  • Sponsor equity = €80.0m − €20.0m debt = €60.0m
  • Exit enterprise value = 5.0x × €37.4m = €187.0m
  • Exit net debt = €20.0m − €48.8m = −€28.8m, i.e. €28.8m of net cash
  • Exit equity value = €187.0m + €28.8m = €215.8m
  • MoM = €215.8m / €60.0m = 3.60x
  • IRR = 3.60x ^ (1/3) − 1 = 53.2%

Then interpret it rather than stopping at the number. There is no multiple expansion and no revenue growth in this return, and leverage at 1.7x entry EBITDA is barely a third of a conventional buyout. The entire 53.2% comes from operational execution plus the balance sheet clean-up — which is both the signature of the turnaround model and the reason these assets trade at 5.0x while healthy peers trade at 9x or more. The seller is not being naive; it is pricing the execution risk the buyer is assuming. For the return arithmetic itself, how to calculate MoM and IRR in a private equity interview and the MoM and IRR case cover the shortcuts worth memorising.

The Four Follow-Ups You Should Expect

"Why can't the €48.8m be valued at the exit multiple?" Because the multiple applies to run-rate EBITDA, and the release never touches EBITDA. It enters through net debt instead. There is also a durability argument: an earnings improvement recurs annually and a buyer will pay a multiple for it, whereas a working capital release happens once and cannot be repeated, since days cannot fall below zero.

"Defend your 6.0% on 70%." Answered in Step 4 — spend cube for the addressable share, pilot tender and prior carve-out experience for the rate, and an explicit acknowledgement that the payables extension works against the price negotiation.

"What if you phase the levers over three years?" The exit equity value barely moves, because exit is priced off final-year run-rate EBITDA. What changes is the middle: years one and two generate far less EBITDA, interest cover is tighter, covenant headroom is thinner, and the assumption that operating cash flow is neutral stops holding. A realistic phasing might be 40% of the uplift in year one, 80% by year two, 100% by year three. The honest conclusion is that the headline IRR is roughly unchanged while the probability of achieving it falls — which is precisely why committees discount unphased plans.

"Why won't a lender size debt off the €37.4m?" Because lenders underwrite delivered earnings. At close the business has earned €12.0m, so €20.0m of debt is already 1.7x on the only number that exists. Sizing at 3.0x of €37.4m would imply €112m of debt against €12.0m of actual EBITDA, which is over 9x and unfinanceable. This is why turnaround deals are equity-heavy, and why the refinancing comes later — once the margin is proven, the business supports real leverage, at which point a dividend recapitalisation can pull equity forward before exit. The constraint itself is worked through in the Debt Capacity case.

How to Structure the Answer Under Time Pressure

If you get ninety seconds rather than fifteen minutes, compress rather than skip. State the margin gap, name the four levers, flag immediately that three are EBITDA and one is cash, give an approximate size for each, and land on a run-rate EBITDA and a rough money multiple. An interviewer would far rather hear a correctly structured estimate than a precise calculation of the wrong thing.

The same discipline applies to the broader paper exercise format, where the constraint is deliberately time rather than data — see what a paper LBO is and practise it on the Paper LBO case. And once you can quantify a plan, the natural next skill is presenting it: Writing an Investment Thesis covers how the operational plan becomes the operational leg of a one-page thesis that an investment committee will actually read.

Work the full exercise yourself on Operational Improvement: What Aurelius Does, where every input above is given and the model answer shows each step with the figures substituted in.