No single deal sinks a SaaS company's margin. Leakage does. A point of discount here, a free quarter there, a renewal that went out flat instead of up — each is defensible in the moment and invisible alone. Add a year of them together and you've funded a problem nobody decided to have. Here's where that money goes, and how to catch it before it's signed.

What is margin leakage in SaaS?

Margin leakage is the gap between the margin a deal should earn and the margin it actually earns, created by concessions and terms granted during the sale and at renewal. It's not fraud and it's rarely one bad call. It's the accumulated cost of small, reasonable-looking give-aways that never get measured together.

The trap is that every individual concession has a story. The 22% discount closed the quarter. The free onboarding got the logo. Net-60 was "just this once." Each is a fine decision in isolation. Margin leakage is what you see only when you stop looking at deals one at a time.

Margin leakage vs. revenue leakage

People use the two terms interchangeably. They're related but they leak in different places and different people own the fix.

Revenue leakage Margin leakage
What leaks Money you earned but never collected Money you collected but gave away in the deal
Typical causes Billing errors, failed payments, un-invoiced usage, missed renewals Over-discounting, free scope, soft uplifts, costly terms
Where it hides The billing system The negotiated deal and the renewal
Who fixes it RevOps / Finance / billing ops The deal desk and deal review

Industry estimates commonly put revenue leakage at 1–5% of ARR — real money, and worth a billing audit. But that figure only counts revenue you were owed and didn't capture. Margin leakage is the amount you did capture and then handed back at the table. It doesn't show up as an error, so it rarely gets audited at all — which is exactly the part a deal desk exists to govern.

Where deals quietly lose margin

Here's the map. Most leakage in a B2B SaaS deal comes from a short, repeating list — and almost all of it is decided in the last two weeks of a negotiation, when everyone wants the deal done.

Leak Where it hides What it quietly costs
Over-discounting past the defensible line The final-week price concession Every point off list is a point off gross margin, permanently
Free add-ons and unpriced scope "We'll throw in onboarding / an extra module" Real delivery cost, and a precedent for the renewal
Soft or missing renewal uplift Renewals sent flat to avoid a hard conversation 5–7 points of compounding growth you never get back
Long payment terms (net-60/90) The order form's fine print Cash-flow cost — you finance the customer for a quarter
Under-priced year one in a ramp Multi-year deals that start cheap "to land" A discount that never fully catches up over the term
Price protection and MFN clauses Legal redlines nobody scored commercially A cap on every future price increase for that account
Over-provisioned usage commitments Volume tiers discounted on peak, not actual use Margin priced against usage that never arrives

None of these are exotic; they're the standard furniture of enterprise deals. The problem isn't that they happen — it's that nobody compares the give-away to what the company agreed to before.

Why margin leakage is so hard to see

Four things keep it invisible:

  • No deal looks wrong. Each concession is small and locally justified; the damage shows only in aggregate, and nobody's job is to look there.
  • The cost is delayed. A soft renewal uplift or a net-90 term doesn't hurt this quarter. It hurts three quarters out, long after the deal was celebrated.
  • It's spread across owners. The rep grants the discount, Legal accepts the clause, Finance eats the payment term. No one sees the whole leak.
  • The evidence is unstructured. Discounts may sit in the CRM, but terms live in PDFs, order forms, and email. You can't spot a pattern you can't query.

That last point is the real one. The knowledge that would prevent the next leak — what you granted last time, and what it cost — exists; it's just trapped in documents nobody reads before the next negotiation.

How to find margin leakage in your own deals

You don't need a new tool to start — just the deals you've already closed and harder questions than "did we hit the number."

  1. Measure realized discount against list, by segment — and trend it. A stable company-wide average hides drift inside segments; enterprise discounts creeping from 15% to 25% over a year is your leak, on a chart.
  2. Score deals on true gross margin, not ACV. A $200k deal with net-90 terms, a free module, and a capped uplift can be worth less than a clean $160k.
  3. Audit renewals for uplift capture. What share of renewals went out with a price increase, and what was the average? Flat renewals are silent leakage.
  4. Read the terms commercially. Pull payment terms, price caps, and MFN clauses from signed order forms and put a dollar cost on each — Legal cleared them; nobody priced them.
  5. Compare each to your own precedent. The only honest benchmark for "was this concession okay" is what comparable past deals got.

How to stop margin leakage before it's signed

You stop leakage the same way you find it — comparing every non-standard deal to your own history — except before the ink dries. That's the whole job of a deal review: a fast, structured second look at the deals that can move the number the wrong way, each concession measured against precedent while there's still room to hold the line.

It's simple to describe and hard to run by hand. Someone has to assemble the comparable past deals, the real terms, and the true margin for every reviewed deal, every week, without a team — which is what makes most deal reviews degrade into a rubber stamp.

That's the standing function Precedent runs for you. It turns your own deal history into a one-page review brief on every non-standard deal before sign-off — the defensible discount at that ACV, the terms that cost you last time, the concession you're about to repeat — delivered in Slack within two hours. Not a tool you have to feed and not a hire you have to make; the deal-review function itself, answering the margin questions with your history instead of someone's memory of it.