Skip to content

RESOURCES · BLOG

What a lender's reporting cycle costs, and what it should

Rui Luis · Published · 4 min read

A thin off-white measuring scale with fine tick marks across a dark dotted background: a white pointer sits about two-thirds of the way along, an ember-orange limit mark stands near the right end, and a small glowing orange triangle is lit above the scale well before the pointer

Short answer: Most owners have never counted what the lender package costs: pulling the figures, calculating the covenants, filling in the bank's template, and answering the questions when a number looks off. Count it before anything else. Most of the work is rules a computer can apply the same way every time. The covenant figures are exactly the kind of number a model should never judge, and the conversation with the banker stays with you.

This is for owners and CFOs of family companies with a credit facility and regular reporting covenants: a borrowing-base certificate, a compliance certificate, aged receivables and payables, or quarterly statements.

What goes into a lender reporting cycle?

The details vary by bank and by facility, but most cycles have the same parts:

  • Close the month or the quarter.
  • Pull the trial balance, aged receivables and payables, inventory, and the bank balances.
  • Calculate the covenant ratios the agreement defines: often a debt service coverage ratio, a leverage or debt-to-equity ratio, and a current or working-capital ratio.
  • Work out the borrowing base, if the facility has one: eligible receivables and inventory, less the exclusions the agreement lists.
  • Re-key all of it into the bank's own template or certificate.
  • Have the CFO or owner review it and sign.
  • Send it, then answer the account manager's questions when something doesn't match last period.

Every one of those steps happens on a date, the same way each time. That is what makes it a loop worth looking at.

How do we count what it costs?

The same four numbers we use for any loop (rule 01), applied to this one:

  1. Hours per cycle. Everyone who touches it, including the owner's review and the time spent answering the bank's questions afterwards.
  2. Days from close to submission. How long after the books close does the package reach the bank?
  3. Questions or corrections per cycle. How often does the bank come back, or does a figure change after it was sent?
  4. Who can build it. If the person who usually does it is away at quarter-end, can anyone else?

Count one cycle honestly, as it actually runs, not as the procedure says it should. Most owners are surprised by the third and fourth numbers more than the first.

Which steps can run on their own?

Sort every step: delete it, turn it into code, give it to a model, or leave it with a person (rule 03). For a typical lender cycle the sort looks like this:

Step Sorted to
Re-key figures into the bank's template Delete (fill the template from the source data)
Pull the trial balance, agings, inventory and bank balances Code
Calculate each covenant ratio as the agreement defines it Code
Calculate the borrowing base and its exclusions Code
Compare each ratio to its limit and flag any that are close Code
Explain any figure that moved sharply since last period Model drafts, person edits
Review, decide and sign Person
Talk to the banker Person

Notice how little goes to a model. Covenant arithmetic is defined in the credit agreement, line by line; it is a rule, not a judgement. If a model is calculating a ratio or deciding whether it is acceptable, it is in the wrong place (rule 04). The one place a model helps is the same as in a board pack: drafting a plain explanation of what moved, for a person to check and sign.

Why do near-misses matter more than hours?

Because the expensive moment in a lender relationship is not the time spent on the package. It is finding out late that a ratio is about to breach.

When the calculation runs as code every time the books close, the comparison to each limit runs with it. A ratio drifting toward its limit is flagged when it starts to drift, not when someone builds the certificate at quarter-end. That gives the owner time to act, and time to call the bank before the bank calls them. The rule doesn't decide anything; it just makes sure a person sees the number in time to decide.

What should we not automate?

The conversation with your banker. The judgement about whether to ask for a waiver or an amendment, or how to explain a bad quarter. Anything the agreement leaves to discretion. These stay with the owner and the CFO, and a well-built loop gives them better information sooner rather than taking the decision away.

It is also worth saying plainly: for a company with a simple facility and one short certificate a quarter, counting may show the loop is small. If the numbers don't make the case, that is a real answer (rule 14).

Where do we start?

Count your next cycle. Write down the hours, the days from close to submission, the questions the bank asked, and who could have done it if the usual person were away. Those four numbers tell you whether there is anything here worth building.

If your lender package runs through one person and a spreadsheet every quarter, book a 30-minute call. No deck, just the work that piles up.

Frequently asked questions

Will our bank accept a package that was assembled automatically?

The bank receives the same package it always has, in the same format, signed by the same person. What changes is how the figures are pulled and checked before anyone signs. Ask your account manager if you want to be sure; most care about accuracy and timing, not how the spreadsheet was built.

What happens if we are close to a covenant?

That is a person's call, and it should be. A rule can tell you earlier that a ratio is drifting toward its limit, and by how much. What to do about it, and what to say to the bank, stays with the owner and the CFO.

We report monthly, not quarterly. Does this still apply?

More so. A monthly cycle is the same loop twelve times a year instead of four, so every hour and every correction counts three times over.

← All posts