Why your MCA portfolio looks healthy when recent deals are not

A merchant cash advance funder we work with had a portfolio that looked fine by every summary metric.

His servicing platform showed $6.2 million outstanding across 340 active deals. The blended collection rate was 94 percent. Twenty-eight deals had been marked as a loss over 18 months, which seemed manageable for a book that size. He used those numbers to approve a new $800,000 funding round in Q3 2025.

Eight months later, that quarter’s deals had produced 19 defaults and were tracking toward $340,000 in losses. The aggregate number had barely changed. The aggregate number was the problem.

What the blended collection rate does not show

A blended collection rate adds up every deal in the portfolio and divides total collections by total outstanding. The number is accurate. What it does not reveal is whether the deals you funded this quarter will perform like the ones you funded two years ago.

The issue is time. Deals funded three years ago have matured, paid down, and either closed cleanly or defaulted. Deals funded three months ago have not. When both groups are combined into one rate, the long-tenured performing deals lift the number and conceal what is happening in more recent fundings.

Grouping deals by the period in which they were funded, sometimes called vintage analysis or cohort analysis, separates those groups and makes the comparison possible.

Four signals that vintage analysis surfaces

Underwriting drift. When approval criteria shift, default rates follow. That shift will not appear in the blended number for 6 to 12 months, because new deals take time to default. A quarterly vintage comparison shows the signal earlier. If deals funded in Q3 are defaulting at twice the rate of deals funded in Q1, something in the approval process changed in Q3.

Seasonal patterns in merchant performance. Some merchant categories run differently in certain quarters. A restaurant does more volume in December than in February. Mixed into a blended rate, that variation disappears. Separated by funding period and compared to the same quarter of the prior year, seasonal patterns become visible and plannable.

The track record of a specific broker. If a new broker who sources deals joined your network, all of their deals land in one or two recent vintages. The aggregate number cannot isolate their book from the rest of the portfolio. A vintage report can. If their deals default at a materially different rate, the vintage data shows it.

The right moment to slow capital deployment. If your most recent three vintages are underperforming relative to the same quarters in the prior year, that is a signal to slow deployment and review criteria before funding the next round. A blended rate will not provide that signal until the losses have already accumulated.

What one portfolio looked like by vintage

Here is what the funder’s book looked like once we split it by quarter of origination.

The Q3 2025 vintage was running at a 79 percent collection rate at six months. The blended portfolio rate was 94 percent. The 15-point gap was invisible in the aggregate report.

We traced the Q3 2025 problem to a one-month period when approval criteria were loosened for a specific merchant category. That change produced one quarter of deals that will drag collections for the next 12 to 18 months.

Why this matters

Three problems develop when a funder only watches the blended number.

The loan-loss reserve is underfunded. A loan-loss reserve, the amount set aside to cover expected defaults, needs to reflect how each vintage is actually performing. When recent vintages are failing faster, the reserve should be larger. The blended rate keeps that gap invisible until it is too late to build the reserve gradually.

Capital decisions are based on outdated signals. Underwriting from 18 months ago says little about whether deals funded last quarter will perform. Deployment decisions running off an 18-month blended rate are running off a number that mostly reflects the past.

Performance reports to capital partners are incomplete. A capital partner funding 30 to 40 percent of your deals wants to know how their share of each funding round performed. A blended rate does not give them that breakdown. Vintage-level reporting does.

What proper vintage tracking looks like

For MCA clients we work with, the portfolio report is split by quarter of origination. Each vintage is compared at the same point in its life cycle to the same period of the prior year.

The loan-loss reserve calculation, sometimes called CECL (Current Expected Credit Losses), runs off vintage-level default rates rather than a blended rate. A reserve calculated from a blended 6 percent default rate is wrong when a recent vintage is running at 21 percent. That difference can represent several hundred thousand dollars in capital that should not be deployed until the reserve is corrected.

Producing this report takes one additional column in the monthly portfolio export. That column is the one that actually indicates whether the business is performing well or drifting toward a problem.

Best practices for MCA portfolio reporting

A few habits that keep vintage analysis useful over time:

• Compare each vintage at the same point in its life cycle. The 90-day collection rate for Q1 2025 should be compared to the 90-day rate for Q1 2024, not the current rate on a two-year-old book.
• Create a separate cohort when a major change happens. New underwriting criteria, a new broker source, or a new merchant category each represent a distinct set of deals worth tracking on their own.
• Track each capital partner’s share by vintage. If a partner funded 40 percent of Q3 2025 deals and that vintage is failing at 21 percent, their specific exposure is not reflected in a blended 6 percent default rate.
• Update the loan-loss reserve quarterly using vintage-level rates. Annual updates with a blended rate leave the reserve wrong for most of the year.
• Run the vintage comparison monthly. Early signals in a 90-day cohort are worth catching at 90 days, not 12 months later.

Three questions worth asking

If you run an MCA shop and are not certain how your recent vintages are performing, three questions to bring to whoever handles your accounting:

1. What is the collection rate on deals funded in the past six months, measured separately from deals funded in prior periods?
2. How is the loan-loss reserve calculated, and does it use vintage-level default rates or a blended portfolio rate?
3. If approval criteria changed in the past 12 months, is that period tracked as its own cohort?

If those answers are uncertain, the blended portfolio rate may be masking a real problem in the recent book. Send your deal book export from the past four quarters alongside the matching loss data. We will pull the vintage split and tell you where the aggregate number and the current reality diverge.