Mistakes to Avoid

BAD: Memorizing assumptions without understanding their business logic

At a 2024 H.I.G. Capital interview, a candidate used a 7x entry multiple because “that’s what the case study showed.” When the interviewer asked why 7x versus 9x, the candidate had no response. The interviewer—a Principal—later noted that “this candidate could do the math but couldn’t think about the deal. That’s not a PE skill.”

GOOD: For any assumption you make, prepare a one-sentence business justification. “I assumed 7x because comparable transactions in the healthcare services sector traded at 6.5x-7.5x in the prior 18 months, and this company’s EBITDA growth profile justified the midpoint.”

BAD: Racing to completion at the expense of accuracy

At a Q2 2023 EQT interview, a candidate completed his model in 22 minutes—fastest in his cohort. He had three arithmetic errors that changed the IRR calculation by 400 basis points. The interviewer specifically noted that “speed without accuracy is worse than slow and correct, because it suggests a candidate who optimizes for output over quality.”

GOOD: Aim for 85-90% completion with 5 minutes to review. At an Oaktree interview in 2024, a candidate who finished with 4 minutes remaining caught and corrected two errors in the debt schedule, which the interviewer cited as “exactly the behavior we want—quality control in a time-boxed environment.”

BAD: Completing the model and waiting silently

At a 2023 Bain Capital second-round, a candidate finished his model and sat quietly for 90 seconds before the interviewer asked him to walk through it. The walkthrough was disorganized and failed to highlight his key assumptions. The interviewer later said the candidate “didn’t seem to understand that the test is the beginning of a conversation, not the end of an evaluation.”

GOOD: Complete the model, take 10 seconds to organize your thoughts, then say: “I’ve finished the base case. Here’s what I built and what the key assumptions were. The IRR comes to 22%, which is driven primarily by our leverage assumption of 5x—we reach full debt paydown by year four, which generates a 2.6x MOIC. Happy to walk through any section in more detail or discuss sensitivities.”

---

FAQ

How much does the LBO paper test factor into the overall hiring decision at PE firms?

The paper test typically represents 20-30% of the first-round evaluation, but its weight is disproportionate to its score. At a 2024 Blackstone first-round, a candidate with a “Weak” technical score received a “Strong No Hire” despite strong behavioral responses, because the paper test was treated as a gating requirement. Conversely, at a Q3 2023 KKR loop, a candidate with an “Average” paper test score but “Exceptional” deal sense discussion in follow-up questions advanced to the second round. The test is a filter, not a score—it eliminates candidates who lack foundational technical competence but doesn’t automatically select those who advance.

Should I use pencil or pen during the LBO paper test?

Use a mechanical pencil with soft lead (0.5mm, grade B or 2B) and an eraser. At a 2024 TPG interview, a candidate using a pen was unable to correct an arithmetic error cleanly, which created visual noise that made it difficult for the interviewer to follow his logic. The interviewer noted that “paper test readability is a proxy for modeling discipline.” At Houlihan Lokey’s 2023 campus recruiting, a senior banker told candidates that “if I can’t follow your model when I’m sitting next to you, the model is wrong, even if the math is correct.”

What if I can’t finish the LBO model in the allotted time?

Incompleteness is not automatically disqualifying if the completed sections are correct and you can explain your approach. At a 2023 Advent International interview, a candidate completed 80% of the model but correctly identified the key drivers of return sensitivity in his incomplete debt schedule. The interviewer—a Director—specifically noted that “he understood what mattered, even if he didn’t finish the mechanics.” The critical error is leaving the model in a state of confusion—half-completed sections with no clear logic. Better to complete 70% cleanly and be able to explain your approach than to rush and produce 100% of a broken model.amazon.com/dp/B0GWWJQ2S3).