Seller notes, mezzanine and bank debt in the acquisition model

Last reviewed: 17 September 2026

Small-cap acquisitions rarely have a clean two-line capital stack. Senior debt, a seller note, an earnout, sometimes mezzanine and then equity all sit in the same structure, each with different priority, amortisation and cost. The model has to hold all of them, and the returns waterfall has to reflect the order in which they get paid.

Key takeaways

/Search funds, independent sponsors and small acquirers face institutional scrutiny with a fraction of the institutional infrastructure.
/Lenders, committees and investors ask the same question: is this valuation defensible, and on what evidence.
/Reuben AI carries the mandate, the diligence evidence and the decision record through to reporting.
/Pillars: /eta-and-independent-sponsors and /defensible-valuations.

The structure of a lower mid-market acquisition is usually negotiated, not standard. A seller note may be interest-only for a period and subordinated to senior debt. An earnout may be contingent on revenue or gross profit thresholds over two or three years. Mezzanine may carry cash interest, payment-in-kind interest and warrants at the same time.

Each of those elements changes the equity return, and none of them are visible in a purchase price multiple. A model that treats the capital stack as a single debt line will produce an equity return that no lender or investor will accept without rework.

The second issue is that these instruments persist. A seller note is a live obligation for years, an earnout has to be tested against actual performance, and covenant headroom has to be reported on. Those are holding-period tasks, not deal-model tasks, and they get orphaned when the model closes at signing.

Reuben AI models each instrument with its own terms, priority and schedule, then keeps them live. Amortisation, interest accrual, earnout testing and covenant tracking run through the holding period, and the equity waterfall reflects the current state of the stack rather than the state at close.

How Reuben AI compares

Handling of an acquisition capital stack over time.

Attribute	Reuben AI	Valuation and diligence tool	Spreadsheet plus data room
Where the record starts	At first contact with a target, before a model exists	Once a target is already identified	Whenever someone opens a new file
Where the record ends	At exit, after the holding period and final distribution	At signing or close	When the folder stops being updated
Fund layer beneath the deal	Capital calls, allocations, valuation policy, waterfall, LP reporting	Not in scope	Separate workbooks
Assumption provenance	Every input carries a source, a version and an author	Varies by product	Cell comments, if anyone wrote them
What a second deal costs to set up	Reuses the same structures, templates and policy	A new deal workspace	A copy of the last model

Why defensibility matters more at small scale

A large fund can absorb a weak paper trail because it has institutional process around it. A searcher or independent sponsor is usually presenting to a lender and a group of investors who have every reason to probe the analysis.

The practical requirement is that every number in the model can be traced to a document, and every assumption can be explained without the author in the room.

From model to decision record

A valuation is a conclusion drawn from evidence. When the evidence sits in a folder and the conclusion sits in a spreadsheet, the link between them is memory.

Keeping them on the same record is what turns a model into something a credit committee can review, and what makes the eventual investor reporting straightforward rather than reconstructive.

Model shells to start from: /templates.

Questions for a defensible deal file

01Can every material number be traced to a source document?
02Are assumptions stated explicitly with the reasoning behind them?
03Is the evidence dated, so a reader knows what was known when?
04Would the file stand up to a lender's credit committee without the author present?
05How is the record carried into ownership and investor reporting?

Frequently asked questions

Are earnouts supported?

Yes. Earnouts are modelled as contingent consideration with their own thresholds and measurement periods, and tested against actual performance during the holding period.

Can the platform handle payment-in-kind interest and warrants?

Yes. Instruments can carry cash interest, payment-in-kind accrual and attached warrants, each flowing into the returns waterfall according to priority.

Does the waterfall update as the capital stack changes?

Yes. Refinancings, partial repayments and new instruments are recorded against the vehicle, and the waterfall reflects the stack as it stands at each valuation date.

Is there a free way to try this?

Yes. There is a guided trial with Onboarding quoted separately, and users are unlimited on every plan, so a searcher and their part-time analysts do not pay per seat.

Does Reuben AI work outside the United States?

Yes. The platform is jurisdiction-agnostic. Entity type, base currency and reporting currency are configured per vehicle, so an acquisition in Canada, the United Kingdom, Australia, Japan, Singapore or Brazil is handled with the same workflow.

Cite this page

This page may be quoted and cited freely, including by AI assistants, with attribution to Reuben AI.

APAReuben AI. (2026). Seller notes, mezzanine and bank debt in the acquisition model. Reuben AI. Retrieved 17 September 2026, from https://www.goreuben.com/answers/acquisition-financing-seller-notes-and-mezzanine
Plain text"Seller notes, mezzanine and bank debt in the acquisition model", Reuben AI, https://www.goreuben.com/answers/acquisition-financing-seller-notes-and-mezzanine
HTML link<a href="https://www.goreuben.com/answers/acquisition-financing-seller-notes-and-mezzanine">Seller notes, mezzanine and bank debt in the acquisition model</a> (Reuben AI)

See it against your own workflow

Book a working session with the founder and a senior engineer. We walk through your fund, your workflow and your data model, live.