Last week of the quarter, and a buyer says the magic words: "Your competitor came in cheaper." The rep, eyes on quota, offers 25% off to close today. The deal lands, the rep hits target, and nobody notices that a quarter of the margin just evaporated on a deal that would have closed at 10%. Multiply that by every rep, every quarter, and you have the most expensive habit in commercial operations, the one a deal desk exists to govern.

The quick version

  • A deal desk is a small cross-functional function (sales, finance, legal, ops) that reviews and approves non-standard deals, unusual discounts, custom terms, odd packaging, before they go out.
  • Discounting governance is the rulebook the desk runs on: who can approve what level of discount, on what evidence, within what time.
  • A discount is far more expensive than it looks. McKinsey's work on pricing found that for the average large company, a 1% cut in price knocks roughly 8% off operating profit when volume holds, because the whole discount comes straight out of the margin, not the revenue.
  • Set a tiered approval matrix, give each tier a deadline, and require a reason that isn't "the customer asked", so speed and margin stop being a trade-off.

The idea in depth

A deal desk is easy to caricature as bureaucracy, another box to tick before a rep can close. Run well, it's the opposite: the mechanism that lets you say "yes" fast to the deals that deserve it and "not at that price" to the ones that don't, without every exception becoming a private negotiation between a rep and a VP.

Salesforce describes a deal desk as a cross-functional team, drawing on sales operations, finance and legal, that handles non-standard deals: the volume discounts, multi-year commitments, custom payment terms and bundled extras that fall outside the standard playbook, serving as the hub for pricing, approvals and compliance (Salesforce, "What is a Deal Desk?"). It puts a structured decision where there used to be an improvised one.

Why a point of discount costs more than a point of price

Here is the number every commercial leader should have tattooed somewhere visible. In McKinsey's classic analysis of pricing, Marn, Roegner and Zawada showed that on the average income statement of an S&P 1500 company, a 1% improvement in price, volume unchanged, lifted operating profit by around 8%, roughly three times the effect of a 1% gain in volume (McKinsey, "The Power of Pricing"). The arithmetic runs both ways: a 1% cut in price drags operating profit down by about the same 8%.

The reason is that a discount has no costs to share the pain. Sell one more unit and the extra revenue is partly eaten by the cost of making it; discount, and the whole giveaway comes out of profit, because the costs were already fixed. A 25% discount on a 40%-margin product doesn't cut your profit by 25%, it cuts it by more than half. Discounting feels cheap to the rep because the cost lands on a line they never see.

So the move is to make margin visible at the moment of decision. Put the gross-margin impact of the discount on the approval screen, in pounds and as a share of profit, not just a percentage off list. A rep who sees "this 20% discount gives away £14,000 of margin" negotiates differently from one who sees "20% off."

flowchart LR
    L(["List price
£100k"]) --> I(["Less on-invoice
discount −15%"])
    I --> O(["Less off-invoice
rebates, terms, freight"])
    O --> P(["Pocket price
what you actually keep"])
    P --> M(["Margin: the discount
came entirely from here"])
The price "waterfall": list price erodes step by step to the pocket price, and every deduction lands on margin. McKinsey's pricing work found off-invoice leakage alone often runs into double-digit percentages of list. Leaders Loop

That waterfall matters because most discounting isn't a single number. McKinsey's "pocket price" research traced how list price erodes through layers, on-invoice discounts, then off-invoice rebates, payment terms, freight and allowances, until the price the company keeps is far below the headline. A deal desk is where you watch the whole waterfall, not just the line on the quote.

Discounting governance: rules beat heroics

The heart of the desk is a discount-approval matrix: a simple grid for which role signs off which level of discount. A common shape, reps approve small discounts themselves, managers approve the middle band, finance signs off the deep ones, so scrutiny rises with the margin at stake. The pattern is widely used; the thresholds are yours to set from your own margins, not a law to copy.

The idea sits on solid commercial ground. In Monetizing Innovation, Madhavan Ramanujam and Georg Tacke argue that discounting should be tightly controlled and used as a last resort, after value is established, not as the opening move when a buyer pushes back. Left ungoverned, discounting becomes the default tool because it's the easiest lever a rep can reach. Governance replaces "whatever it takes to close" with "here is what we'll trade, and what we won't."

A deal desk doesn't slow deals down. It replaces a hundred private negotiations with one public rule.

So the move is to write the matrix down and attach two things to every tier: a time limit and a required reason. A time limit (same-day for tier one, 48 hours for the complex deals) stops the desk becoming the bottleneck reps fear. A required reason, competitive displacement, multi-year commitment, strategic logo, genuine volume, turns "the customer asked" into evidence, and lets you see why margin is leaving, not just that it is.

There's a quieter benefit, too. A clear matrix protects your good reps from your loudest customers. "I'd need to take that to our deal desk, and they'll ask what we're getting for it" is a far stronger position than a rep negotiating alone against their own quota clock. The rule does the pushing back, so the relationship doesn't have to. Governance like this pairs naturally with the broader work of revenue operations.

The honest limitation: governance can curdle into bureaucracy

A deal desk earns its keep on complex, non-standard deals. Point it at every deal and it becomes exactly the drag its critics fear, routine renewals stuck in a queue, deals dying of slow approval, and a workaround culture where people structure things to dodge the desk. The cure for ungoverned discounting is not "approve everything centrally"; it's approve by exception. Standard deals at standard prices should sail through. Spend the desk's scarce attention on the deals that actually move margin.

And the central figure deserves a caveat. The "1% price → 8% profit" result is an average across large public companies on a representative income statement, an illustration of operating leverage, not a promise about your business. A company with very high gross margins (much software) or very low ones (much retail) will see a different multiple. Use it to make the principle land, discounting is leveraged, and the leverage works against you, then run the number on your own P&L.

So the move is to govern by exception and tune to your own economics. Set thresholds generous enough that the bulk of deals never touch the desk, and recompute the profit impact of a discount from your real gross margin, not a borrowed headline.

A worked example

The figures below are illustrative, chosen to show the mechanics rather than to report a real company.

A B2B software company sells a platform with a 75% gross margin. Standard list for the mid-market tier is £100,000 a year. Historically, reps could discount "to close," and the average deal went out at 28% off. Finance notices net revenue retention is fine but new-business margin is quietly sliding, and stands up a deal desk with a three-tier matrix.

Tier one (0–10%): the rep approves it, instantly, no review. Tier two (10–20%): the sales manager approves within the day, and the system asks for a reason code. Tier three (20%+): finance reviews within 48 hours, and the request must name what the company gets in return, a multi-year term, a reference commitment, real volume.

flowchart TD
    A(["Deal request
with discount %"]) --> B{"How deep
is the discount?"}
    B -->|"0–10%"| C(["Rep self-approves
same day"])
    B -->|"10–20%"| D(["Manager approves
+ reason code • within 24h"])
    B -->|"20%+"| E(["Finance reviews
+ what do we get back? • 48h"])
    C --> F(["Closed, margin tracked"])
    D --> F
    E --> F
A three-tier discount-approval matrix: scrutiny and the burden of proof rise with the margin at stake, and every tier has a deadline so the desk never becomes the bottleneck. Illustrative thresholds. Leaders Loop

Two things happen. First, the deep discounts that used to be reflexive now need a justification, and a chunk of them turn out not to have one, the buyer accepts 12% when 28% is no longer the default. Say the average discount falls from 28% to 18%, a 10-point recovery. On a £100,000 deal, that's £10,000 of revenue landing entirely on the 75%-margin line: roughly £10,000 of profit per deal that used to walk out the door. Second, the deals that do get the deep discount now come with a multi-year term or a reference attached, so even the giveaways buy something back.

The rep who feared the desk would cost them deals finds the opposite: standard deals close faster because tier one is instant, and the hard negotiations now have finance's authority behind them instead of resting on one stressed rep at quarter-end. The bottleneck never appears, because most deals never reach the desk at all.

Frequently asked questions

What's the difference between a deal desk and discounting governance?

The deal desk is the who, the cross-functional function (sales, finance, legal, ops) that reviews non-standard deals. Discounting governance is the rulebook it runs on, the approval thresholds, reason codes and time limits that decide what gets waved through and what gets scrutinised. You can have light governance without a formal desk (a rule in your CPQ tool), but a desk without governance is just a meeting.

Won't a deal desk slow our sales cycle down?

Only if you point it at the wrong deals. Govern by exception: standard deals at standard prices should self-approve instantly, and the desk should only see the genuinely non-standard ones. Attach a deadline to every approval tier (same-day for small discounts, 48 hours for complex ones) and the desk speeds the hard deals up, by giving them a clear path, rather than slowing the easy ones down.

How do we set the discount thresholds?

Start from your gross margin, not a competitor's matrix. Work out the discount level at which a deal stops being worth winning, and place your top tier below it. Look at your historical deals: where does discounting cluster, and where does it spike without a good reason? Set tier one generous enough that the bulk of deals never need approval, and reserve the deepest tier for discounts large enough to genuinely move your margin.

Who should own the deal desk?

Most often revenue operations or finance, because the desk's job is to protect margin and consistency across the whole sales org, interests a single sales manager can't be expected to hold against their own quota. Sales must be a full partner, though, or the desk becomes an adversary reps route around. The healthiest version feels like an ally that helps reps win the right deals, not a checkpoint that exists to say no.

Is discounting always bad?

No. A discount in exchange for something, a multi-year commitment, a flagship reference, real volume, faster payment, can be a smart trade. The problem is the unearned discount: margin given away for nothing because it was the easiest way to end an awkward conversation. Governance isn't about banning discounts; it's about making sure every one of them buys something back.

Related in the Toolkit

Where to go next