What does a lender want from my model that I wouldn’t expect?
They want to see how you think. Owners plan for growth. Lenders plan for repayment. So they test the downside: what happens to repayment if a new plant starts six months late, if a big customer pays late, if prices drop, if costs run over. A model that only shows the plan tells them you haven’t thought about the risk.
They are also sharp at spotting inconsistencies. They read the model next to the business plan, the capex list and your actuals. If those tell different stories, they’ll assume you don’t know your business. And every number needs a source: a quote behind each machine, a contract or a study behind each price.
Your own bank asks less for a working-capital line: the cash cycle, the security, and the last two years. For a term loan, to build a plant or buy a company, it tests the model the same way. A development bank reads everything, and takes months.
Where do these models break when the lender’s team tests them, and what does one they trust do instead?
There’s a mistake I see all the time: revenue grows by a fixed percentage a year. Revenue doesn’t just happen. It comes from capacity, sales effort and price, and the model has to show those. Otherwise it’s a spreadsheet of wishes.
After that, models break in three places:
- The start. A new plant ramps up over months, and cash goes out long before it comes in.
- Stock and slow payers. Stock ties up cash before the first sale, and customers who pay late tie it up after.
- The first year of repayment. The loan looks fine on average, but that year the cash barely covers it.
A model the lender trusts plans for low, medium and high demand, and says what the company does in each case. It answers real questions: what happens to repayment if sales dip by 10%, by 15%, by 50%?
Keep the drivers few. I worked with a company whose model had around 80 assumptions, and nobody could say which of them moved the answer. Leadership didn’t trust it, and rightly so. Start with the main drivers of revenue, cost and cash. Add detail only where it changes a decision, and put a source behind it when you do.
The lender meeting is in a month. What do you do first, and what do I walk in with?
We prepare the company for the lender. First, the question the lender is going to ask. What they’ve requested so far tells you what worries them: the ramp-up, the margins, the capex, the cash in year one. Then: can we trust the actuals? If not, they get fixed before the model is built. Four weeks is enough for the actuals, the model and the story if the books are usable. If they aren’t, the first two weeks go on the books.
Then just enough structure for the model to answer that question. You walk in with a story the numbers can defend: here’s what’s working, here’s what we’ve fixed, here’s what’s next and how we’ll measure it. The base case and the downside, side by side.