Concentration Risk: The Spreadsheet Every Factor Maintains
Mid-market factors run concentration risk on a spreadsheet because the LMS does not report it the way credit committee needs to see it. Here is the fix.
TL;DR. Every mid-market factoring firm maintains a concentration risk spreadsheet because the LMS does not report exposure the way credit committee actually reviews it. The sheet rolls up debtors by industry, layers in aging and advance rate, and tracks rolling-portfolio percentages against policy thresholds. When that sheet is wrong, the firm turns down profitable deals or takes on quiet over-exposure that surfaces as a charge-off six months later. The fix is not a software project. It is treating the sheet as real infrastructure with ownership, versioning, and a reconciliation routine.
Your factoring firm has a concentration risk spreadsheet. It lives in Excel, the senior credit analyst maintains it, and it is the only place leadership can actually see how exposed the book is to ABC Trucking right now and how that compares to your 15% single-debtor policy. Your loan management system does not show this view. It has the underlying data, but the rollup, the industry buckets, the aging overlay, and the rolling-portfolio percentages all live in the sheet. When the analyst is out, nobody actually knows where the firm stands.
Why the LMS Does Not Show It
Walk into any mid-market factor running FactorSoft, CADENCE, or one of the other factoring LMS platforms and ask the credit manager to show you concentration. She will pull up a report that gives you outstanding by debtor, sorted descending. That report is correct. It is also not what credit committee wants.
Credit committee does not look at concentration by debtor name alone. It looks at concentration by industry, by aging bucket, by recourse vs non-recourse, and by rolling 30-60-90 day exposure against policy. The LMS knows all of that data. It just does not roll it up the way the committee defines its buckets.
Take industry. Your LMS captures debtor industry as either a NAICS code or a free-text field, depending on the platform. NAICS is too granular for credit committee. The committee thinks in operator categories: transportation, staffing, oil and gas services, government contractors, manufacturing subs, food processing. Those buckets do not map cleanly to NAICS. So the analyst opens her spreadsheet, pulls the LMS export, and maps every debtor to her firm's industry taxonomy by hand.
Same problem with aging. The LMS shows you aging by invoice. Credit committee wants aging by debtor, weighted by exposure, with a separate column for invoices over 90 days that have been bought but not yet reserved. That is a spreadsheet calculation.
Recourse vs non-recourse exposure splits differently again. The LMS tracks it correctly at the contract level, but credit committee wants a single view: net exposure if every recourse client paid us back tomorrow, gross exposure if no client paid us back. Two columns. Both in the spreadsheet.
The rolling-portfolio view is the one nobody can fake. Credit committee wants to see how concentration moved over the last 30, 60, 90 days, not just where it stands today. That requires snapshots. The LMS does not natively keep snapshots of concentration by industry in the format the committee uses. The analyst saves a copy of her spreadsheet every Monday morning, and the comparison view is built off those snapshots.
The Four Cuts the Spreadsheet Actually Runs
Every concentration sheet I have seen at a $40M to $80M factor runs four cuts of the same portfolio. The cuts are not novel. The work is in maintaining them weekly without errors.
By debtor. Top 20 to 50 debtors by outstanding exposure, with percent of portfolio next to each. Anything over 5% gets a yellow flag. Anything over 10% gets a credit committee line item. The threshold varies by firm, but industry research suggests 15 to 25% of capital as the typical single-name ceiling for secured lenders.
By industry. Same total exposure, sliced by the firm's operator industry taxonomy. A firm heavy in trucking will run staffing, oil and gas services, government contractors, manufacturing subs, and "other" as their five primary buckets. A firm heavy in staffing will cut it differently. The point is the buckets reflect how the firm thinks about correlated risk, not how the Census Bureau categorizes businesses.
By aging bucket. Total outstanding split into current, 1-30, 31-60, 61-90, and 90+. The 90+ column carries a separate concentration calculation because a stale invoice on a single debtor can flip that debtor from green to red without changing the total exposure number at all.
By rolling-portfolio percentage. Two snapshots: 30 days ago and 60 days ago. The committee wants to know if any debtor or industry concentration is trending up. A debtor that was 4% of the book 60 days ago and is now 9% deserves attention even if neither number breaches policy.
That is the sheet. Four cuts, one portfolio, updated weekly.
What Goes Wrong
The sheet is fragile in three predictable ways.
Stale data. The analyst pulls the LMS export on Monday morning. By Wednesday afternoon there have been three large advances, two early payoffs, and a debtor concentration breach that nobody sees because the sheet is still showing Monday's snapshot. The fix is not "automate the pull." The fix is admitting the sheet is a weekly view and routing real-time concentration questions through a different surface, usually a quick LMS query the credit manager runs ad hoc.
Bucket disputes. Two analysts will categorize the same debtor into two different industries. A staffing firm that places nurses might be "staffing" to one analyst and "healthcare" to another. The committee assumes the buckets are stable. In reality they drift quietly as new analysts join and as the firm picks up clients in adjacent verticals. The fix is a written taxonomy with examples, owned by the credit manager and reviewed annually.
Single-keeper risk. The sheet is maintained by one person. When she takes a week off, nobody else can produce it. When she leaves, the firm spends three months reconstructing the methodology from her file history. This is the single largest hidden risk inside the concentration risk function, and it is invisible until the day it is not.
I have walked into firms where the credit manager had not taken more than three consecutive days off in five years, partly because she could not hand off the sheet. That is not a sustainable operations posture. The International Factoring Association frames this as part of the underwriting-at-scale problem: as a factor grows, the analysis one person could carry in her head becomes a documented system, or it becomes a liability.
Fixing It Without Replacing the LMS
The instinct is to replace the LMS. Do not. Mid-market factors have invested years of process knowledge into their current platform. The concentration sheet is not evidence the LMS is wrong; it is evidence the LMS does not bend to your firm's specific industry taxonomy and reporting cadence. No off-the-shelf LMS will bend to it. Every factor's sheet looks slightly different because every factor's risk policy is slightly different.
What you actually need is for the sheet to stop being a sheet and start being infrastructure.
That means three things.
Ownership. One named credit manager owns the methodology. A second analyst can rebuild it from documentation if the owner is unavailable. The taxonomy, the bucket definitions, the aging calculations, and the rolling-snapshot routine all live in a written runbook, not in the owner's head.
Versioning. Every Monday's snapshot is preserved with a clear date stamp. The committee should be able to compare any two arbitrary weeks without asking the analyst to reconstruct an old view from LMS history.
Reconciliation. Once a quarter, the sheet's portfolio total ties out to the LMS's portfolio total to the dollar. Discrepancies are explained or fixed. Most concentration sheets I have seen have a quiet $50K to $200K variance against the LMS that has accumulated over time. That variance is the kind of thing a real auditor will catch eventually.
If you want to layer AI on top of any of this, the right place is the bucket assignment. Categorizing several thousand debtors into your firm's industry taxonomy is the kind of work an AI agent can take on reliably, because it is high-volume, rule-based, and tolerant of human review. The committee meeting itself is not what needs automation. The reconciliation is not what needs automation. The bucket work is.
What This Looks Like in Practice
A $50M factor I spoke with last quarter had been running their concentration sheet the same way for nine years. The credit manager was retiring in eighteen months. The owner could not name a single colleague who knew how the industry taxonomy was constructed, which advance-rate columns fed into the rollup, or where the 90+ aging carve-out was calculated. The sheet was a four-tab Excel file with formulas pointing across tabs to other formulas.
We did not rebuild the sheet. We documented it. The credit manager spent six hours over two weeks walking us through every bucket definition, every formula, every "exception" she made when a debtor did not fit the taxonomy cleanly. The output was a 14-page runbook, a written taxonomy with examples, and a Monday-morning checklist a second analyst could follow.
The sheet stayed in Excel. The firm did not replace the LMS. What changed was that the firm no longer had a single point of failure inside its risk function.
The Granular Read
Mid-market factors run on systems that were not built to roll up exposure the way their committees actually deliberate. The spreadsheet that bridges the gap is doing real work, and it deserves real engineering treatment. If your firm is one credit manager away from not being able to produce the concentration view leadership relies on, that is the conversation to have. We have done this kind of operator-knowledge capture for mid-market manufacturers and field-service businesses, and the playbook ports cleanly to specialty finance. Book 30 minutes and we will walk through what a runbook looks like for your sheet.
FAQ
What concentration threshold should a mid-market factor use?
The most common single-debtor policy ceiling is 10 to 15% of portfolio for unsecured exposure, lower for higher-risk industries. Industry-level concentration ceilings vary by firm. Transportation-heavy factors often set transportation at 40 to 60% of portfolio with hard caps on individual sub-segments. Whatever your numbers, the policy should be written down and reviewed annually, not carried in the credit manager's head.
Can FactorSoft, CADENCE, or another LMS replace the spreadsheet?
Not entirely. LMS platforms report by debtor reliably. They do not natively roll up by your firm's specific industry taxonomy, run aging-weighted concentration, or maintain weekly snapshots in the format credit committee uses. The spreadsheet is bridging a real gap that the LMS was never designed to fill.
How often should the sheet be reconciled to the LMS?
Once a quarter at minimum, the portfolio total should tie to the dollar. Anything over a $25K variance should be investigated and either fixed or documented with a written reason.
What is the right ownership structure for the sheet?
One credit manager owns the methodology and produces it weekly. A second analyst should be able to produce it from a documented runbook within a half-day if the owner is unavailable. That second analyst should rehearse the production at least quarterly, not just read the runbook.
Where does AI actually help here?
Industry bucket assignment for new debtors. Reviewing the LMS export, identifying which existing taxonomy bucket each new debtor belongs to, and flagging the ambiguous cases for human review. The committee deliberation, the policy thresholds, and the quarterly reconciliation are human work.
Keep Reading
- Why Specialty Finance Is the AI Vertical Nobody Is Targeting. Where the AI services money is actually going across factoring, ABL, equipment finance, and commercial lending, and why the obvious targets keep missing the mark.
- The Carrier Appetite Chart Living in Your Broker's Head. The other operator-credible spreadsheet that runs a business, and why the parallel applies anywhere senior judgment is being stored in Excel instead of infrastructure.