Find the errors in this three-statement model
This version of the integrated three-statement model has six mistakes in it, of the kind a quickly generated, unreviewed model tends to carry. Some break a check. Others leave every check at zero and only show when an assumption changes. Find all six, correct them in the file, and make the model reconcile again.
Reviewing a model is a different skill from building one, and it is the one a first year in a deal team uses most. A model that balances is not a model that is right: three of the six mistakes here leave every check at zero.
The workbook
Download the model with its mistakes
The same company, sheets and drivers as the integrated three-statement model: three historical years, five forecast years, working capital, PP&E, debt and equity schedules, and a Checks sheet. The cover lists the reported figures the history was keyed in from, and the conventions the model claims to follow.
What to look for
Six habits a reviewer brings to every model. Each finds at least one of the mistakes; none says where it is.
Tie the history to its source
The cover lists the reported figures the history was keyed in from. A historical year that disagrees with its source is the most common and the least forgivable mistake, because nothing downstream can be right.
Read the checks, then follow the first one that fails
A balance sheet that does not balance tells you that something is wrong, not where. Find the earliest year and the earliest line where the check moves, and work back from there.
Check the signs on the cash flow statement
An asset that grows uses cash; a liability that grows provides it. Say each line aloud with its sign and ask whether the business has more cash or less because of it.
Look for a number where a formula belongs
Click along a forecast row. Every cell should hold the same formula shifted one year. One typed figure in a row of formulas is right today and wrong the moment anything changes.
Change an assumption and watch what does not move
Switch the scenario, raise a rate, change the amortisation. A line that should respond and does not is reading the wrong cell, even if its value looks right in the base case.
Check every schedule against the convention it claims
The cover states how interest, depreciation and the revolver are meant to work. Read each schedule against those statements, and follow capital expenditure from the cash flow statement to the asset it buys.
Keep a list as you go: the sheet, the row, the years, what is wrong and what the line should be. That list is what a reviewer hands back, and it is how you will check yourself at the end.
Check your work
Correct the mistakes in the file, save it as .xlsx and upload it in the grading panel on the model page. It is recalculated and marked the way a reviewer would mark it: outputs, links, formulas rather than typed numbers, whether it holds when an assumption changes, and whether every check closes. Anything still wrong is named by sheet, row and year.
A free account has one graded submission in total, shared with the starter workbooks, so upload when you think you have all 6. Pro can fix a finding and resubmit.
Upload your corrected fileRead the schedules behind it
Each page sets out what the schedule carries, how it links to the statements, and the mistakes it usually holds.
Working capital schedule
Receivables, inventory and payables in days
Capex and depreciation (PP&E roll-forward)
Opening, plus capex, less depreciation, closing
Debt schedule
Tranches, interest, amortisation and the sweep
Cash flow statement
Indirect method, closing on balance-sheet cash
Then build the debt schedule yourself from the starter workbook, or open the worked model in the browser and change the inputs. The method behind the hunt, step by step, is in how to audit an AI-built financial model.