The CIM tells you what the seller wants you to look at. The deal tells you what the seller wants you to ignore. After diligencing 100 independent mortgage banks and closing 70-plus of them, I've come to believe that most PE diligence on lenders gets the surface layer right and the operating layer wrong — and the operating layer is where the deal compounds or grinds.
This is not a knock on diligence teams. The generalist financial-services analyst can model GAAP-to-economic earnings, normalize an MSR portfolio, and stress-test a warehouse line. That's the floor. The problem is that an IMB doesn't fail at the floor. It fails one level down — in the assumptions about loan officer capacity, branch P&L variance, and operational throughput that nobody on the deal team has ever actually run a P&L against.
Here is what I look at first when a sponsor brings me into IMB diligence.
Capacity per LO, not LO count
Every CIM tells you the loan officer headcount and the average production per LO. Almost no CIM tells you the distribution — and the distribution is the entire signal. In the IMBs I've diligenced, the top 20% of LOs typically produced 60–75% of the volume, the middle 60% produced the rest, and the bottom 20% produced essentially nothing while still consuming branch overhead. If you mark to the average, you've already mispriced the deal.
The right question isn't "how many LOs do they have." It's "if the top decile leaves in the next 18 months, what's left?" Then back out compensation parity, recruiter activity in the seller's footprint, and the median LO's tenure on platform. That tells you what you're actually buying. In several deals I worked, the answer was: a building, a license, and a very expensive operations team. That's not always a deal-breaker. But it should be priced like what it is.
Branch P&L variance — the average lies
Same problem one layer up. A 30-branch IMB with $40k average branch contribution sounds healthy until you realize five branches are throwing off $200k each, twelve are roughly breakeven, and the other thirteen are losing money the seller has been masking with corporate overhead allocation choices.
I've never seen a seller volunteer this view. They don't have to be hiding it; they often haven't built the reporting that surfaces it, because the firm grew through acquisition and the legacy GLs never got rationalized. The diligence work is to rebuild branch-level economics from production data, allocate true cost-to-serve (not the seller's allocation methodology), and ask which branches you'd close on day 31 if you had to. The answer to that question — and how the seller reacts to it — tells you more about the platform than any quality-of-earnings memo will.
MSR marks: how aggressive, and against which assumption?
MSR economics are where I see the most diligence variance, because the marks themselves can be defensible while the assumption stack underneath them is loose. Prepay speeds matched to a benchmark that doesn't reflect the seller's actual book composition. Servicing cost-per-loan that hasn't been refreshed since the platform was a fraction of its current size. Recapture assumptions that count on a retail channel the seller is quietly losing.
The right diligence isn't "is the mark fair." It's "what does the mark assume, and would the operating team behind this servicing portfolio actually deliver those assumptions?" That's an operator question. It comes down to specific people, specific systems, and a specific recapture engine — not a model.
Technology debt as a modernization tax
Most IMBs in the $100M–$5B production band are running an LOS implementation that was best-in-class in 2017, three layers of integrations that nobody fully owns, and a point-of-sale tool the LOs hate but tolerate. None of this is in the CIM as a risk. It shows up as a modernization tax in years two and three when the new sponsor wants to scale.
The operator's diligence question: what would it cost — in dollars, in distraction, in attrition risk — to bring this stack to where it needs to be in 24 months? That number is rarely zero, and it's almost always more than the sponsor's initial post-close capex assumption. Better to surface it pre-LOI than to find out in the first board meeting after close.
The recruiter signal
The single best leading indicator on an IMB's health isn't in the financials. It's what local recruiters say about the seller. Recruiters know which LOs are taking calls, which branch managers are looking, which compensation packages are unraveling, and which cultural changes are pushing producers out the door. Six months before any of this hits the production reports, the recruiters know.
You will not get this from the seller. You will get it from making three or four phone calls in the seller's footprint to mortgage recruiters who don't know you're working a deal. Most diligence teams skip this. It is the cheapest, highest-yield call you can make in the entire process.
GSE and warehouse relationships
The seller will tell you the GSE relationships are clean and the warehouse lines are committed. Both can be true on paper and broken in operating reality. I've worked deals where Fannie Mae was carrying performance issues that hadn't yet generated a formal letter, where a warehouse provider had been quietly tightening covenants, and where a take-out investor had moved from preferred-pricing to standard pricing without anyone in finance flagging the margin impact.
These relationships don't break in diligence. They break six months in, when the new owner is already capitalized, the operating plan is already in motion, and the cost of the surprise is fully borne by the deal. The operator's job in pre-LOI is to call the people on the other end of those relationships and ask, on background, what they're seeing. Most of them will tell you. Few diligence teams ask.
What this means for sponsors
None of the above replaces traditional diligence. It supplements it. The financial work, the legal work, the QofE — all required. What I'm describing is the layer underneath that, where the actual operating risk lives, and where generalist teams typically have neither the relationships nor the muscle memory to look.
The cost of getting this wrong is asymmetric. A $200M IMB acquisition that loses its top three branches in year one isn't a 5% problem. It's a deal-thesis problem. The cost of doing this layer of diligence properly is a few weeks of operator time pre-LOI and a few hundred phone calls. I've never seen a sponsor regret making that investment. I've seen plenty regret skipping it.
The market is watching every IMB before a sale process opens. The sponsor's edge is whether their diligence sees what the market already knows.
If you're working a deal in the IMB or specialty-lending space and you want an operator on the diligence side, that's the work I do. I've sat in the COO chair on a national platform, run a turnaround on a PE-owned specialty lender, and led a successful sell-side exit. I know what's worth pricing, what's worth fixing, and what's worth walking away from.