The Model Works, But Will the Engine Start?
You have the Quality of Earnings (QoE) report in hand. The adjusted EBITDA looks defensible. The Total Addressable Market (TAM) is growing. On paper, the asset is a winner. But you aren’t buying paper. You are buying a complex machine made of code, people, and processes—and historically, that machine is far more fragile than the spreadsheet suggests.
We have entered the era of the "Operational Audit." With median holding periods stretching to 5.7 years according to recent Bain & Company data, the "flip it quick" strategy is dead. You cannot financial-engineer your way out of a broken delivery model or a tech stack held together by duct tape.
Operational Due Diligence (ODD) has shifted from a checkbox exercise to a forensic investigation. Why? Because the data shows that over half of fund failures result from operational breakdowns, not investment strategy. If you are a PE Operating Partner, you know that the skeletons in the closet aren't financial—they are structural.
The "Black Box" Problem
Founders are experts at selling the vision of their operations. They speak of "agile teams" and "proprietary tech." But without a rigorous ODD framework, you are buying a black box. You might acquire a company where:
- Tribal Knowledge is the only operating system (if the CTO is unavailable, the product dies).
- Technical Debt acts as a 30% tax on every new feature.
- Scalability is a myth because every new customer requires custom engineering.
The following 12 questions are designed to break open that black box. They are not polite. They are diagnostic.
The 12-Point Diagnostic Framework
Divide your inquiry into three buckets: Process Scalability, Technical Reality, and Commercial Efficiency. If the target management team cannot answer these with data, you are looking at a distressed asset priced as a platform.
Bucket 1: The Scalability & Process Test
Goal: Identify Key Person Dependency and "Hero Culture."
1. "Show us your 'Key Person Dependency' map."
Do not ask if they rely on key people. Ask to see the map. Which critical workflows (billing, deployment, sales engineering) stop if one specific person is unavailable? If the Founder or CTO is the only person who can approve a deployment, you have a bottleneck, not a business.
2. "Where is the SOP for your most profitable service line?"
We don't want to see a high-level flowchart. We want the Standard Operating Procedure (SOP) that a junior hire uses to execute the work. If it doesn't exist, or hasn't been updated in 2024, you are buying undocumented tribal knowledge, which trades at a discount.
3. "What is your Revenue Per Employee (RPE) trend over the last 8 quarters?"
Efficiency should increase with scale. If RPE is flat or declining while revenue grows, the company is scaling linearly (hiring bodies to solve problems) rather than exponentially (using software/process). This is a margin killer.
4. "How does your forecast accuracy correlate with actuals?"
Ask for the variance report. A company that consistently misses its own internal forecast by >15% does not have a market problem; it has a data hygiene problem.
Bucket 2: The Technical & Product Reality
Goal: Quantify the "Tax" on Future Growth.
5. "What percentage of engineering hours are allocated to 'Keeping the Lights On' (KTLO)?"
If they say "0%," they are using inconsistent definitions. If they say "15%," they are elite. If they say "50%+," you are buying a maintenance project. McKinsey data suggests 31% of acquired codebases are riddled with critical debt that threatens valuation.
6. "Show us the Open Source License Audit."
Proprietary tech built on "Copyleft" (GPL) libraries can be a legal landmine. You need to know if their IP is actually theirs to sell. This is often the single biggest blocker in tech due diligence.
7. "What is the Recovery Point Objective (RPO) and when was it last tested?"
Founders will show you a backup policy. You need to see the test log. A backup that hasn't been restored in 12 months is a wish, not a strategy.
8. "How much custom code is required to onboard a $100k ARR client?"
If the answer is "none, it's configuration," you have a SaaS product. If the answer is "200 hours of engineering," you have a services firm masquerading as software. Valuation multiples should be adjusted accordingly.
Bucket 3: The Commercial & Human Engine
Goal: Validate Growth Quality.
9. "What is the CAC Payback Period by channel?"
Aggregate CAC hides sin. You need to know if the "growth" is coming from efficient inbound (6-month payback) or expensive paid outbound (24-month payback). High-growth startups often burn cash on inefficient channels to pump the top line before a sale.
10. "What is Net Revenue Retention (NRR) excluding price increases?"
Are customers staying and buying more because they love the product, or simply because you raised prices? Strip out the inflationary gains to see the true product-market fit.
11. "What is the attrition rate of hires made in the last 12 months?"
High turnover in the first year signals a "bait and switch" culture or poor onboarding. Bad hires bleed EBITDA through recruiting fees and lost ramp time.
12. "Who are the top 5 customers, and do we have a 'Whale Trap' scenario?"
If Customer A accounts for 30% of revenue and demands a custom roadmap, they effectively own the company, not you. This concentration risk is a primary deal-breaker.
The 30-Day "Mock Audit" Action Plan
Whether you are on the buy-side evaluating a target, or the sell-side prepping a portfolio company for exit, you cannot afford to wait for the bankers to find these issues. You must run a "Mock Audit" 6-12 months before the transaction.
Step 1: The Documentation Sprint (Days 1-10)
Force the target to build a "Data Room" that isn't just PDF contracts. Demand a live folder of "System Evidence." If they claim to have a secure SDLC (Software Development Life Cycle), ask for screenshots of the Jira workflows and GitHub pull request rules. Evidence beats intent.
Step 2: The Technical Deep Dive (Days 11-20)
Bring in a third-party technical assessor. Do not rely on the CTO's word. Run a code scan (using tools like SonarQube or Black Duck) to quantify the technical debt. Assign a dollar value to the remediation. If it costs $2M to fix the code to a scalable standard, that comes off the purchase price.
Step 3: The Margin Stress Test (Days 21-30)
Recalculate the "Adjusted EBITDA" by stripping out the "Heroics." If the founder is working 80 hours a week and doing the jobs of a VP of Sales and a Product Lead, you need to add back the cost of hiring those two executives. Often, a $5M EBITDA business is actually a $3M EBITDA business once you pay market rates for the labor currently being donated by the founders.
Conclusion: Don't Buy the Narrative, Buy the Machine
As Bain & Company noted in their 2025 Global Private Equity Report, "margin growth" accounted for just 6% of value creation in software deals over the last decade. That era is over. In the current interest rate environment, you cannot rely on multiple expansion or sheer revenue growth to generate returns.
You must engineer efficiency. And you cannot engineer what you do not understand. Use these 12 questions to turn the lights on. If the answers scare you, you have two choices: walk away, or price the risk into the deal.
