Every concession feels like it costs you once. You give 10% to get the deal over the line, you book the revenue, you move on. Clean trade.

It isn't. The moment you give something because a buyer asked for it, you've told their brain there's more on the table. So they ask again. And again. You didn't close a deal — you trained a customer. And what you trained them to do is rely on a discount you may not be able to give next time.

Here's where it turns from a cost into leakage. If you can't or won't hold the concession at renewal, they don't renew at a higher number. They leave. The price you gave to get in the door becomes the price you have to keep hitting forever, or the relationship is over.

Discounting isn't the problem. Discounting without a system is.

Start with the rule most CFOs run on: total concession on a deal should stay under 10–15% of its value. That ceiling isn't arbitrary, and it's the one number in this piece I'd stake my name on. I've sat in the rooms where deals get defended to the Sequoias and Lightspeeds of the world, and 10–15% is the line you can hold without flinching. Stay under it and the discount is a rounding error in the board deck. Go past it and you're not defending a discount anymore, you're defending your judgment. Take a $1M deal and call the ceiling 15% for simplicity: a $150K budget.

Most teams track the price discount against that ceiling and stop there. But a concession is never just price. The roadmap feature you committed to build. The liability cap you raised. The net-90 payment terms. The three seats you "threw in." The most-favored-nation clause buried on page 14. Every one of those has a cost, and they all draw down the same $150K. Add them up and a "10% discount" is often a 25% concession — you just never priced the rest of it.

Now the part nobody models. You didn't give $150K once. You set a reference price. At renewal the buyer expects the same $150K — plus the roadmap promise, plus the terms. So the real cost isn't $150K. It's $150K per year, for the life of the account. On a three-year deal, that one concession is $450K. And if you try to claw it back, you don't renew flat. You churn the logo you paid to win.

I watched a CFO demand a specific discount because "he always gets that much." Our list was already 30% under the competition — the extra 10–15% put us at nearly half their price, fully loaded. We acquiesced.

They never rolled the product out. Churned inside a year. The rep got paid on a three-year commit; we collected less than twelve months of revenue; and the customer left hating us. The discount didn't win the deal. It funded the exit.

Even "the company that doesn't discount" discounts. A cap of ~15% mid-year quietly becomes 30%+ by late January as quarter- and year-end pressure builds. The cost isn't any single deep deal — it's that enterprise buyers now know the cadence and wait for it. Predictable discounting trains your entire market to stop paying list.

Paddle/ProfitWell put hard numbers on it: discount-led "saves" churn at 70–80% anyway. When a customer threatens to leave and you hand them 50% off, you haven't retained them — you've delayed the churn by one to three months and confirmed the price was negotiable all along.

The 15–20% annual-commit discount is what a good concession looks like. You give margin and get something back: cash up front and materially lower churn. That's give-to-get, written straight into the pricing model instead of negotiated away in a room.

A security startup I know priced against incumbents charging $150K per test by offering $50K for twelve. Disruptive — and a trap. They now have ten pilots and nowhere to go at renewal. Move to $75K, still half the incumbent, and that's $250K back on the table they gave away for nothing. Underpricing is a concession too. It's just one you make to yourself, before the buyer even asks.

STEP 1

Before any negotiation, know the number below which the deal isn't worth doing. Everything above it is negotiable; nothing below it is. Write it down before you're in the room, because you will not think clearly once the buyer is across from you and your VP is asking why the deal isn't closed yet.

STEP 2

No concession leaves the table without something coming back: a longer term, prepayment, a case study, a reference, a logo, a roadmap item you get to deprioritize. Give it away for free and the buyer learns there's always more if they push. Make them trade for it and they learn your price actually means something. It's the same dollar amount either way. The difference is what you've trained them to do next time.

One deal taught me the cleanest version of this. Procurement wanted year one at $0 — they'd be running us and the incumbent in parallel during migration, and they swore every other vendor had already agreed to free. One of those vendors was public. I'd read their filings, and I knew that wasn't true. So instead of fighting the number, I sat down with our finance team and restructured around it: $0 in year one, with years two and three stepped up so total contract value landed exactly where we needed it. They got the optic they had to walk back to their own CFO. We kept the economics and locked a multi-year commit. The ramp even mirrored the real migration — our usage climbing as theirs wound down. Nobody left money on the table. We just moved it across the term.

STEP 3

The framework matters more than its exact contents. A discount matrix that's slightly wrong but consistently applied beats a perfect one you abandon at quarter-end. And price the full concession — discount, terms, roadmap, legal — against one ceiling, and decide as a total. You can change the system over time. You can't run without one.

This is where the precedent is set. Whatever you give to land the account becomes the floor you're expected to hold forever. Land at a price you can defend at renewal — not the lowest price that closes this quarter.

Your single best moment to convert a concession into value. Never hand the discount back for free — trade it for a longer term, an expansion, or a multi-year lock. If you have to hold the price, get something for holding it.

The temptation is to undercut hard. Resist it. You need to be cheap enough that switching is an easy yes. You don't need to be the cheapest, and being the cheapest actively hurts you: it caps what the account is ever worth at renewal and leaves you nothing to grow into. Win on the gap, not the floor.

Procurement is paid to extract concessions; they'll ask whether or not they need to. Assume the ask is ritual, anchor high, and make every give conditional. The buyer who "always gets that much" is testing whether your price is real. Show them it is.

A concession feels like the customer-friendly move. It isn't. It tells your buyer your price was never real, and it sets up a renewal you're structurally going to lose. The discount you hand over to build trust is the same thing that quietly erodes it, because now every number you give them is suspect. Holding your price and making them trade for value is the version of the relationship that's still standing in three years.

If you have customers asking for discounts and you want a framework to decide when to give them, grab time here.