Modeling Test Day: What to Do in the First Five Minutes

Most people who run out of time on a modeling test lose it at the start, not the end. They open Excel before they know what they're building.

A modeling test is a timed exercise where you build something in Excel from a prompt. Usually it's a simple LBO or a three-statement model, sometimes with a short written recommendation at the end. You might get a blank workbook, a partly built template, or a PDF packet with a few pages of company information and a list of assumptions. The clock is usually somewhere between one and three hours.

The instinct is to start typing right away, because the clock is running and typing feels like progress. That instinct is what causes most of the trouble. Candidates who run short on time rarely do so because they were slow at Excel. They run short because they built the wrong thing for twenty minutes, or built the right thing in an order that forced them to rip it up halfway through.

Five minutes of setup fixes most of that. Here's how to spend them.

Minute one: read the whole prompt before touching Excel

All of it, start to finish, including the last page. Prompts often bury something that changes the build near the end. A line saying the sponsor wants a dividend recap in year three, or that the revolver is undrawn at close, or that the final deliverable is a one-paragraph investment recommendation and not just a returns table.

As you read, mark three things on the page or in a scratch tab:

• The time limit and how the file gets submitted. If you need to email it back by a certain time, know that before you start, not with two minutes left.
• What's already built for you. A template with the income statement laid out saves you real time, but it also means you have to work inside someone else's structure. Check whether their rows already have formulas or are just labels.
• Every number in the prompt that you'll need. Purchase multiple, leverage, interest rates, growth rates, margins, capex as a percentage of revenue, exit multiple, holding period.

Minute two: find the deliverable and work backward

Ask what the reviewer will actually look at first. On an LBO test it's almost always IRR and MOIC, sometimes under a couple of exit scenarios. On a three-statement test it's a balance sheet that balances and a cash flow statement that ties to the change in cash.

Once you know the output, the path to it is mostly fixed. For a standard LBO it runs roughly like this:

1. Sources and uses at close. Purchase enterprise value plus fees on one side, debt tranches plus sponsor equity on the other. They have to match.
2. Projected revenue down to EBITDA.
3. EBITDA down to free cash flow available for debt paydown, which means subtracting cash interest, cash taxes, capex, and the increase in net working capital.
4. A debt schedule that uses that cash to pay down debt each year.
5. The exit. Exit multiple times final-year EBITDA gives exit enterprise value, and subtracting remaining net debt gives exit equity value.
6. Returns. Exit equity divided by initial equity gives MOIC, and the IRR follows from that and the holding period (plus any interim dividends).

Writing that list down takes thirty seconds and saves you from the most common sequencing error, which is polishing the income statement for fifteen minutes before realizing the debt schedule needs a line you didn't build.

Minute three: put every input in one place

Before you build anything, set up an assumptions block at the top of the sheet or on its own tab. Every hardcoded number from the prompt goes there once. Everything else in the model references those cells.

This feels slow. It isn't. It's what lets you change an exit multiple or an interest rate at the end without hunting through forty cells, and a lot of tests ask for exactly that kind of sensitivity in the last section. It also makes your model readable to whoever grades it, and they are reading it.

• Hardcoded inputs in blue font, formulas in black. It's the standard convention and graders notice when it's missing.
• Label every input with its unit. "5.0x" and "5.0%" are easy to mix up when you're tired.
• If the prompt gives a number you'll derive from other numbers (an equity check implied by total price minus debt, for example), calculate it rather than typing it. Then it updates if something upstream changes.

Minute four: decide what you'll simplify

You won't build a perfect model in the time you have, and nobody grading it expects you to. What they want is a model that's correct where it matters and clearly simplified where it doesn't. Make those calls now, on purpose, instead of discovering them under pressure at minute fifty.

Interest and the circular reference

This is the single biggest time trap on an LBO test. If you calculate interest on the average debt balance for the year, you create a loop. Interest depends on the ending balance, the ending balance depends on how much cash went to paydown, and that cash depends on interest. Excel will either throw a circular reference warning or, if iterative calculation is turned on, quietly solve it until one bad input blows the whole sheet up with errors that are painful to clear.

The simple fix is to calculate interest on the beginning-of-year balance. It breaks the loop, it's a widely accepted simplification on a timed test, and it only moves the answer slightly. If the prompt specifically asks for average balances, then use them, but add a circuit breaker toggle (a cell that switches the average calculation off) so you can recover if the sheet errors out.

Other reasonable shortcuts

• Annual periods instead of quarterly, unless the prompt says otherwise.
• Taxes as a flat rate on pre-tax income, with no deferred tax or loss carryforward detail unless it's given.
• Working capital as a single net line driven by a percentage of revenue, not separate receivables, inventory, and payables, if the prompt doesn't break it out.
• A cash sweep that pays down the most senior prepayable debt first, rather than a detailed waterfall across every tranche.

Whatever you simplify, write it down. A short line in the assumptions block like “interest on beginning balance to avoid circularity” tells the reviewer you made a choice. Without it, the same simplification can look like you didn't know better.

Minute five: set up the skeleton and save

Lay out the timeline across the top: a closing column (often labeled year 0) and then each projection year. Put the row labels down the left for the sections you listed in minute two, with no formulas yet. Now the whole model exists as an empty frame, and every minute after this is just filling it in.

Then save, with a sensible file name, and keep saving every few minutes. Excel crashing mid-test happens, and it's a terrible way to lose half an hour.

One quick check before you move on: confirm calculation is set to automatic (Formulas, Calculation Options). If a previous user left it on manual, your numbers won't update and you can lose a lot of time wondering why a correct formula shows the wrong answer.

After the first five minutes

Build in the order you wrote down, and build the whole chain roughly before making any part of it pretty. A complete model with plain formatting beats a beautiful income statement attached to nothing. If you get stuck on one piece, hardcode a placeholder, flag it in a different color, and keep moving. You can come back.

Leave real time at the end for checks. How much depends on the length of the test, but holding back the last ten minutes or so of a one-hour test is a reasonable rule. Use it for these:

• Sources equal uses.
• On a three-statement model, assets equal liabilities plus equity in every period, and ending cash on the cash flow statement matches cash on the balance sheet.
• Debt never goes negative. If your sweep pays down more than the balance, cap it with a MIN.
• Returns pass a sanity check. A plain LBO with modest leverage and no multiple expansion returning 60% a year usually means a sign error or a missing subtraction somewhere, often the exit debt.
• The written answer, if there is one, actually states a recommendation and the main reason for it.

For the specific errors that tend to show up in LBO answers, see common LBO interview mistakes and how to avoid them. And if you're still reaching for the mouse, the Excel shortcuts post is worth an hour of practice before test day. Speed at the keyboard buys back exactly the kind of time these five minutes cost.

Common questions