Fix borrower reporting at the source

A monitoring platform reads what the borrower sends: the compliance certificate, the borrowing base, the monthly KPI pack. When those numbers are assembled by hand, the platform shows the same errors, faster.

That holds for a cash flow lender reading a quarterly certificate and for an asset-based lender reading a weekly borrowing base. The cadence differs, and the problem sits in the same place.

Covenants exist so a lender can act early, and acting early takes numbers that arrive early. When a credit goes wrong, the first question is usually why the numbers did not show it sooner. The answer often sits upstream of the lender, in how the borrower closes its books and builds its reports.

Lenders get earlier warning by instrumenting the borrower's own systems, under an information protocol both sides agree to at the start.

Borrower data breaks in 3 places, all upstream of the lender. Each is the ordinary result of a finance team running on spreadsheets and goodwill. The controls that catch them also make a double-pledged receivable or an invented invoice harder to hide.

Manual closes. Accruals are estimated, journal entries post late, and the month closes when the controller has time. Every package built on the close inherits its delays and its estimates, so each monthly number reaches the lender already old.

Spreadsheets. When the borrowing base lives in a workbook, eligible receivables are filtered by hand, links break when someone inserts a column, and last month's file becomes this month's template. An error there overstates availability, and it can survive several certificates before a field exam tests the detail.

Adjustments that drift. The credit agreement defines EBITDA and the add-backs it allows, and many of those add-backs take judgment. Quarter by quarter the claimed add-backs grow, each defensible alone, while the support behind them thins. Soon nobody can trace the number on the certificate back to the ledger.

Each break is small. Together they mean the lender monitors a story about the business, assembled by hand once a month.

Covenants test the past. A quarterly certificate arrives weeks after the quarter closes, and it reports a ratio built from adjusted numbers.

The operating data moves first, and 3 signals are worth reading every week.

Cash conversion. Days to collect stretch, suppliers shorten terms, and inventory days creep up. Each comes straight from the ledger and the warehouse system. When the cycle lengthens, the business needs more revolver to stand still.

Working capital. Receivables and inventory build faster than sales, and the revolver starts funding the gap. The detail matters more than the total: receivables aging past terms, and stock that has not moved in a quarter.

Drift. Adjusted EBITDA holds while operating cash flow falls. Some of that gap is working capital, interest, and taxes. The rest is add-backs, and add-backs that grow faster than the business deserve a question.

A lender who sees drift early can ask about it in a phone call. A lender who first sees it on the certificate is already in a negotiation.

Each of these signals lives in the borrower's systems every week. Read the same way each week, from the source, they reach the lender well before the ratio moves.

Covenants test the past. The operating data moves first.

The borrower's own systems already hold all of it. What the lender needs is a process that reads them the same way every week, without someone building a file.

The same goes for the 13-week cash flow. Rebuilt in a spreadsheet each Monday, it is a forecast of a forecast.

Built from the ledger, its first weeks come from open receivables and payables, and the rest from the plan. What earns a lender's trust is the variance report, each week's forecast against actual cash from the bank feed.

Instrumenting a borrower's systems takes the borrower's cooperation, so the terms are agreed in writing before any work starts. The protocol answers 5 questions.

The protocol protects both sides. The lender knows what it will receive and when, and that the review window has an end. Management knows that nothing reaches the lender before management has seen it.

The borrower gets something back too. Ad hoc requests from the lender shrink, since the answers arrive on schedule, and when the company needs a waiver, it asks with numbers the lender already trusts.

It also sets limits. The lender receives outputs on the agreed schedule, and the data stays in the borrower's systems.

In practice, the read can be a single page: each signal, its trend, and the system it came from. A lender can scan it in a minute and knows which number to ask about.

The last step is to build the lender package from the same source as the board pack: the same ledger, the same KPI definitions, the same close.

When the sponsor's board pack and the lender's package come from one source, every conversation starts from the same numbers. An amend and extend still turns on pricing, fees, and new equity from the sponsor, and those terms get argued on numbers both sides accept.

The plan gets better too. Tested against margin by customer and SKU, cash conversion, and what each initiative must deliver, it becomes a plan the credit committee can check.

One source also gives the next field exam less to reconstruct. Eligibility rules live in the system, applied the same way each week, with a log of every receivable excluded and why.

Monitoring starts in the borrower's ledger. Fix the data there, agree in writing who sees what and when, and every platform downstream reads better numbers.