The exit math:
every dollar off EBITDA costs ten — unless it buys a turn.

There is a reflex in every owner-led business: protect EBITDA, because at a 6× multiple every dollar you spend costs ten at exit. The reflex is right. The conclusion — that you therefore never spend on the data layer — is where it goes wrong. Some spending doesn't just dent EBITDA; it moves the multiple. And the multiple is the larger number.

Two effects, not one

Exit proceeds are EBITDA times a multiple. Most operating decisions touch only the first term. A price increase, a headcount cut, a better vendor contract — they move EBITDA and leave the multiple alone. That is the world the reflex was built for, and inside it the reflex holds.

Data infrastructure is one of the few line items that touches the second term. Multiples aren't fixed. A buyer pays more turns for a business it can verify — where the margin number, the delivery number, and the inventory number survive diligence because they come from one model, not three reports. Spending to make the business verifiable dents EBITDA slightly and raises the multiple. On the arithmetic below, the second effect is roughly seven times the first.

This comparison is the board slide

Illustrative numbers, an $8M-EBITDA shop, held flat against the same shop after a $150K investment in the model layer. Same business, same year — the only difference is whether the numbers are defensible when a buyer's analyst pulls them apart.

Illustrative · multiples vary by sector, size, and buyer · the mechanism does not

Run this comparison on your own numbers — and weigh an ERP upgrade, a migration, analytics, the platform, or AI against each other — in the interactive exit model.

The $150K comes straight off EBITDA, exactly as the reflex fears — $8.0M becomes $7.85M. But the multiple moves from 6.0× to 7.0×, and applied to nearly $8M of earnings, that one turn is worth almost $7M. The spend returns roughly forty-six times its cost, and none of that return shows up on the income statement — it shows up on the term sheet.

Why the multiple actually moves

A buyer discounts what it cannot verify and pays up for what it can. The turn is not a reward for prettier dashboards; it is what disappears from the risk column when diligence can trace every important number back to a single, documented model. These are the findings that hold a multiple down — each one an unpriced risk a buyer must either resolve or discount for:

Finding	What diligence sees
Margin defined three ways	Standard cost vs. last-actual vs. blended — three defensible answers, one board question.
On-time delivery on three dates	Promise vs. ship vs. invoice date. Each report picked its own.
Costing rules changed in the ERP	The report never heard about it.
Measures copied across spreadsheets	Each copy drifting quietly from the last.
Refresh path owned by a departed vendor	The model works until the day it doesn't, and nobody has the keys.

Clear all five and the business stops presenting as a story a buyer has to take on faith. It presents as a system a buyer can inspect — and that is the entire difference between the sixth turn and the seventh.

The same asset, seen from the exit

This is the exit-facing view of the argument in Own the platform, not the portal. A reporting layer that stands apart from the ERP — documented, in a mainstream environment, verifiable without a vendor's login — is the same asset whether you are running the business or selling it. It shortens diligence, it survives the buyer's integration, and it is one of the few investments that supports the multiple instead of only the EBITDA line.

The ordering is the same one that governs everything else here: define the numbers, build the trusted model, and the exit math takes care of itself. A business sells the sixth turn on its story. It sells the seventh on its system.