Operating Diligence in PE: KPIs, Unit Economics, and What Models Must Prove
Commercial diligence beyond QoE: KPI families by business model, how findings map into LBO cases, red flags, interview answers, and practice habits.
PE · 5 min read
Quality of earnings tells you whether historical EBITDA is real. Operating diligence tells you whether future EBITDA is achievable. In PE interviews, candidates who only discuss add-backs miss half the job. Sponsors underwrite drivers: price times volume, retention, utilization, same-store sales. Not just a management hockey stick. This guide covers the KPI categories diligence teams actually request, how they should flow into an LBO operating model, and how to talk about them in an interview without sounding like a buzzword generator.
QoE vs operating diligence
| Lens | Question | Typical outputs |
|---|---|---|
| QoE | Was reported EBITDA normalized correctly? | Adjusted EBITDA bridge, working capital peg |
| Operating | Can the plan actually happen? | KPI dashboards, cohort data, market studies |
| Commercial | Will customers stay and pay? | NPS, churn, contract backlog, win rates |
Your model's revenue build should name the same drivers the diligence memo cites. If the IC deck says "3% price, 2% volume" but your Excel only has one blended growth line, you cannot stress the thesis.
How interviewers probe this
They may not hand you a full VDR. They will say: "Management guides 12% revenue growth. What would you diligence?" A good answer names KPIs and what would make you cut the case.
KPI families by business model
Subscription / recurring revenue
| KPI | Definition | Model link |
|---|---|---|
| ARR / MRR | Recurring revenue run-rate | Revenue starting point |
| Gross churn | Logo or revenue lost | Downgrade line in forecast |
| Net revenue retention | Expansion minus churn on cohort | Above 100% = land-and-expand story |
| CAC / LTV | Cost to acquire vs lifetime value | Sanity-check sales and marketing spend |
| Rule of 40 | Growth % + EBITDA margin % | Quick quality screen |
Interview point: NRR above 100% means existing customers grow faster than churn. LBO cases can underwrite lower new-logo pressure, but only if cohorts support it.
Industrial / distribution / manufacturing
| KPI | Definition | Model link |
|---|---|---|
| Volume / units | Shipments, tons, units | Separate price vs volume |
| Capacity utilization | Output ÷ max capacity | Capex trigger for expansion |
| Gross margin per unit | Price − direct cost | Margin bridge |
| Backlog | Orders not yet shipped | Near-term revenue visibility |
Consumer / retail
| KPI | Definition | Model link |
|---|---|---|
| Same-store sales (SSS) | Comp locations growth | Split new units vs SSS |
| Average ticket / traffic | Decompose revenue | Marketing efficiency |
| Inventory turns | COGS ÷ inventory | Working capital forecast |
| Four-wall EBITDA | Store-level profit | Roll-up underwriting |
Software / tech-enabled services
| KPI | Definition | Model link |
|---|---|---|
| Bookings vs revenue | Contract signed vs recognized | Deferred revenue roll |
| Implementation backlog | Services revenue visibility | Hiring plan |
| R&D / capex as % sales | Reinvestment rate | Maintenance vs growth capex |
| Gross margin | Subscription vs services mix | Margin bridge quality |
From diligence findings to the operating case
A credible LBO forecast usually documents:
- Baseline: LTM adjusted EBITDA from QoE
- Bridge: price, volume, mix, new products, cost programs
- Initiatives: quantified with timing (not all Day 1)
- Risks: customer concentration, contract renewals, regulatory
| Initiative type | Modeling treatment |
|---|---|
| Pricing increase | Phased over 2 to 4 quarters (customer pushback) |
| Plant consolidation | One-time cost + run-rate savings after a delay |
| Sales force expansion | S&M upfront, revenue lag 2 to 3 quarters |
| Procurement savings | Often 50% to 70% of vendor quote in Year 1 |
| Cross-sell / attach | Tied to NRR / cohort evidence, not a flat plug |
Example bridge language
"Base case revenue is 3% price and 2% volume on the core book, plus $8M of Year 2 cross-sell that only stays if NRR stays above 105%. Cost-out is $6M run-rate starting month 9 after $2M of costs to achieve."
That sentence is more PE than "revenue grows 8%."
Red flags that kill deals (and interview stories)
- Customer concentration: top customer above 20% of revenue without a long contract
- Negative cohort trends: older vintages churn faster
- Margin expansion with flat headcount and no automation story
- Working capital release every year forever (not sustainable)
- CapEx below depreciation indefinitely (asset quality risk)
- Bookings growth while deferred revenue and cash collections lag
- Capacity already maxed with no capex in the plan
Strong candidates narrate one metric that worried them and how it changed the case (lower entry multiple, more equity, tighter covenants, or walking away).
How this shows up in technical interviews
"Walk me through how you'd diligence this software business."
ARR growth, churn/NRR, CAC payback, gross margin, S&M efficiency, cohort chart, deferred revenue, customer concentration.
"What would make you cut EBITDA in the case?"
Missed price increases, lost contract, utilization cap, delayed synergy capture, higher churn.
"How does diligence differ from the banker model?"
Banker model sells the company. PE diligence tries to break the plan and rebuild conservatively.
"What KPI would you put on the IC dashboard monthly?"
Pick ones that match the thesis. For a subscription buyout: NRR, gross churn, net new ARR, gross margin. For an industrial: volume, utilization, unit margin, backlog.
Tie-in to returns attribution
Operating diligence validates the EBITDA growth leg of returns attribution. If growth in the model assumes 500 bps of margin expansion without a headcount or pricing bridge, deleveraging and multiple expansion cannot save a weak operational story.
A clean integrated answer:
"I underwrite returns attribution only after the KPI bridge supports the EBITDA path. Otherwise I am just allocating a fantasy multiple."
Practice habits that make this real
- Take any LBO case and replace blended revenue growth with price/volume/mix
- Add a one-page KPI dashboard assumption tab
- Write three diligence questions that could falsify the base case
- Cut Year 2 EBITDA by 10% and restate MoIC/IRR so you feel the sensitivity
Falconcrest-style institutional PE cases reward full-model coherence. Use them to practice naming drivers, not only hitting a returns cell.
Practice on SheetRank
Apply what you learned with drills, guided Academy builds, and live deal underwriting.