Discount Approval Matrix and Governance

Executive summary

A discount approval matrix is a table that states, for each type of deal, who may approve a discount, up to what level, and what they must see before they say yes. Done well it protects margin without slowing every deal. Done badly it adds a queue, trains reps to split deals into smaller pieces, and tells leadership nothing about where margin went.

Three design choices decide which one you get. First, anchor the matrix to a margin floor, not just a discount percentage, because the same 12% discount is harmless on a high-margin line and destructive on a thin one. Second, route by dollars as well as percentage, because a 6% discount on a very large deal can cost more than a 20% discount on a small one. Third, record every approved exception with a reason, because the exception log is the only data that tells you whether the policy is working.

This page covers the matrix, the evidence an approver needs, the exception record, the monthly review, and the mistakes that make a policy decorative. For why governance alone does not hold realization, read Why Price Realization Keeps Drifting: a discount policy controls the form concessions take, and a matrix that ignores off-invoice items just moves the leak. For the revenue-side diagnostic that quantifies how much is leaking, read Price Waterfall Analysis.

When to use this

Build or rebuild the matrix when:

• Discounts are approved case by case in chat or email with no consistent rule.
• Average realized price is falling and nobody can say which deals caused it.
• Sales leaders approve their own team's exceptions.
• A price increase is planned and you need to protect the new list. See Price Increase Playbook.

Do not use a matrix as a substitute for fixing incentives. If commission pays on gross revenue, reps will find the cheapest route to a closed deal, and a matrix will change the route, not the pressure.

Diagnostic questions

Answer these from the CRM, the quoting tool, and the ERP before you design anything.

1. What is the average discount from list on closed deals over the last 12 months, by product line and by segment?
2. What share of deals carries any discount, and what share carries a discount above the threshold at which approval is currently required?
3. Who currently approves discounts, how long does approval take, and is the approval recorded in the system of record?
4. What is the contribution margin or gross margin floor by product line, and is it written down?
5. Do the discounts on invoices match the discounts that were approved, or are there differences at invoice time?
6. How much concession leaves through off-invoice routes (rebates, free freight, extended terms, service credits) that the current approval does not cover?
7. Do renewals go through the same approval as new business?
8. Are deals being split, re-dated, or re-papered to stay under a threshold?
9. Is commission calculated on gross revenue, on margin, or on realized price?
10. How many exceptions were approved last quarter, and how many have a recorded reason?
11. When an approver declines, what happens, and how often does the deal then close at the requested discount anyway?

Question 8 and question 11 expose whether the existing policy is real. If deals are being split or declined discounts end up approved at a higher level, the matrix has a design problem, not a compliance problem.

Framework: floor, authority, evidence, record

1. Floor: set the discount ceiling from margin

Start with the margin floor, the lowest pocket margin you will accept on a deal for a given product or service line. The floor converts to a maximum discount for any list price:

Margin after discount % = (List margin % − Discount %) ÷ (100% − Discount %)
Maximum discount % = (List margin % − Floor %) ÷ (100% − Floor %)

With a list margin of 40% and a floor of 30%, the maximum discount is (40 − 30) ÷ (100 − 30) = 14.3%. Check it: a list price of 100 and cost of 60 becomes a price of 85.71 after a 14.3% discount, and (85.71 − 60) ÷ 85.71 = 30.0%. A line with a 25% list margin and the same 30% floor has no discount room at all. That is why a single company-wide cap like "reps may discount up to 10%" is wrong for at least one of your lines.

A second formula shows what a discount costs in volume terms:

Volume gain required to break even % = Discount % ÷ (List margin % − Discount %)

At a 40% margin, a 10% discount needs 33.3% more volume to leave contribution dollars unchanged, and a 20% discount needs 100% more. Put this table on the approver's screen. It changes the quality of the conversation.

2. Authority: tiers by percentage, by dollars, and by margin

An authority tier says who can approve a deal that falls in a band. Use three triggers, and escalate to the highest tier any trigger reaches:

Tiers are normally four: rep (within a pre-approved band), sales manager, finance partner or pricing owner, and CFO or CEO for breaches of the floor or strategic exceptions. Keep the number of tiers small, because every added tier adds queue time.

Separation of duties matters. The approver of a discount should not be the person whose commission or quota depends on the deal. At the top tier, include someone from finance.

3. Evidence: what the approver sees

An approver who sees only "15% requested, please approve" will approve. Require the request to carry:

• List price, requested price, discount percent and dollars.
• Resulting pocket margin versus the floor for this line.
• Total customer revenue and margin over the last 12 months, and whether the account is above or below target pocket margin.
• The competitive reason, in one line, and the competitor's price if known.
• What the customer gives in return (volume commitment, term, payment terms, reference, scope).
• Any off-invoice items attached to the same deal.

If a field is missing the request does not route. This is the part of the process that takes discipline, because without it the matrix is a signature ritual.

4. Record: the exception log is the data

Every approved discount above the rep's band goes in a log with the fields in the template section. The log drives the monthly review, supports the audit trail a buyer or lender may ask for, and is what lets you test whether discounting is buying anything: win rate, deal size, or terms.

Step-by-step process

Weeks 1 to 2: Baseline

1. Pull 12 months of closed deals with list price, invoice price, discount, line, segment, rep, and approver. Owner: Sales Ops. Output: deal extract.
2. Compute realized discount distribution by band, by line, and by rep. Owner: FP&A. Output: distribution table and the share of discount dollars per band.
3. Reconcile approved discounts to invoice discounts. Where they differ, find why. Owner: Finance. Output: reconciliation with a list of causes.
4. Quantify off-invoice concessions by type. Owner: Finance. Output: off-invoice schedule.

Weeks 3 to 4: Design

5. Set the margin floor by line. Owner: CFO. Output: floor table and maximum discount by line.
6. Draft the authority tiers using percentage, dollars, and floor triggers. Owner: CFO and Head of Sales. Output: matrix draft.
7. Define the evidence package and the request form. Owner: Sales Ops. Output: required fields in the CRM or quoting tool.
8. Define service levels. For example, a target response time for each tier. Owner: Head of Sales. Output: written approval service levels.

Weeks 5 to 6: Build and pilot

9. Configure approvals in the quoting tool or CRM, so routing and records are automatic where possible. Owner: Sales Ops. Output: configured workflow and test cases.
10. Pilot with one team for two to three weeks. Collect cycle times and rejected requests. Owner: Head of Sales. Output: pilot readout and changes.
11. Align incentives. Decide whether commission will use realized margin, or at least reduce payout for discounts above a band. Owner: CFO with HR. Output: decision memo. Compensation changes involve HR and sometimes legal review, so follow your plan documents and agreements.

Weeks 7 to 12: Roll out and hold

12. Roll out to all teams with a one-page guide per tier. Owner: Head of Sales. Output: trained teams and a published matrix.
13. Start the monthly discount review. Agenda below. Owner: CFO. Output: first review minuted.
14. Close the off-invoice gap. Extend approval to rebates, freight, terms, and credits. Owner: CFO. Output: policy covering all concession types.
15. Audit at 90 days. Compare the exception log to invoices and check for deal splitting. Owner: Internal audit or Controller. Output: audit memo and fixes.

Templates

Discount approval matrix (copy into a spreadsheet)

Set X, Y, Z, A, B, and C from your own deal distribution and floor, using the method in the worked example. Escalate to the highest tier any column reaches.

Exception log

Monthly discount review agenda

1. Average discount, discount dollars, and realized price by line versus plan and prior month (10 min).
2. Exceptions above tier 1: count, dollars, reasons, outcome (10 min).
3. Reps and teams outside the band, and win rate for each (5 min).
4. Invoice versus approval differences and off-invoice concessions (5 min).
5. Rule changes needed and decisions for sales and finance (5 min).

Metrics and KPIs

Setting targets by method. Use your own history before anything external. Your trailing best realized price by line is a target you have already reached. Your margin floor is a contractual-style limit that you set once and hold. Win rate by discount band tests whether deeper discounts buy anything; if win rate is flat across bands, the deeper bands are paying for nothing. Do not copy a discount-rate benchmark into your policy. Compare to named listed peers only where a public source reports it.

Common mistakes

1. Writing the matrix on percentage alone. A single percentage cap ignores line margin and deal size.
2. Approvers with a stake in the deal. If the approver is paid on the revenue, the matrix only adds delay.
3. Slow approvals. If tier 2 takes a week, reps will give the customer a smaller discount in writing and an informal promise afterward. Publish and measure service levels.
4. Ignoring off-invoice concessions. The deal that cannot get a discount gets free freight, extended terms, or a rebate. Cover all of them in the same matrix, as discussed in Why Price Realization Keeps Drifting.
5. Deal splitting. Contracts get split to fit under a dollar trigger. Test for it by looking for clusters of deals just below each threshold, and apply the dollar trigger by customer per period rather than by single quote.
6. No exception log, or a log nobody reads. The review is the control.
7. Treating renewals as exempt. Renewals are where price quietly re-anchors. Put them through the same matrix.
8. Leaving incentives alone. A matrix without a compensation change is a speed bump.

Worked example

Illustrative. All numbers are made up to show the method. They are not benchmarks.

A company reviews 100 closed deals, each with a list value of $50,000 ($5.0M total list). The distribution of realized discounts:

Average discount. 677.4 ÷ 100 = 6.774%. Discount dollars. $5.0M × 6.774% = $338,700.

Where the dollars sit. Deals at 10% and above are 22 of 100 deals (14 + 6 + 2) but carry (170.8 + 105.0 + 48.0) ÷ 677.4 = 47.8% of discount dollars. Almost half the cost is in under a quarter of the deals, so approval effort at tier 2 and above is well spent, and approval effort on deals under 5% (52 deals, 23.8% of dollars) is mostly overhead.

Margin floor. The line has a 40% list margin and a 30% floor. Maximum discount = (40 − 30) ÷ (100 − 30) = 14.3%. The 8 deals above 15% (6 + 2) are therefore breaching the floor on this line, and 14 more are between 10% and 15%, close to it.

Cost of the deepest tiers. If the 8 deals above 15% had been capped at 15% (just above the 14.3% floor, to keep the arithmetic simple), the discount points fall by 6 × 2.5 + 2 × 9.0 = 33.0, so average discount falls from 6.774% to 6.444%, and discount dollars fall by $5.0M × 0.33% = $16,500. That is a modest figure, and it is the honest answer: here the policy's value lies less in capping extremes than in the middle bands, where 25.2% of discount dollars sit and where a pocket-margin view and a required trade could change decisions.

Volume test. At a 40% margin, a 10% discount needs a 33.3% volume gain to break even, and a 15% discount needs 60.0%. If win rate in the 10 to 15% band is no better than in the 5 to 10% band, those 14 deals are paying for nothing.

Resulting matrix (example thresholds). Tier 0: up to 5% and under $2,500 of discount. Tier 1: up to 10% and under $5,000. Tier 2: up to 14.3% (the floor), any dollars. Tier 3: anything beyond the floor. The thresholds come from the distribution and the floor, not from a rule of thumb.

AI prompt assistance

Provide your own data and verify output against the CRM and ERP before changing policy.

1. Build the discount distribution.

I will paste a CSV of 12 months of closed deals with list price, invoice price, product line, customer segment, rep, and approver. Compute average discount, discount dollars, and share of discount dollars by band (0-5, 5-10, 10-15, 15-20, 20+). Show the formulas and flag any row where invoice price is above list.

Check: recompute one band by hand and tie total list and invoice values to the ERP.

2. Calculate maximum discount by line.

For each product line below, I give the list price, unit cost, and target margin floor. Compute the maximum discount that keeps pocket margin at or above the floor, and the volume gain needed to break even at 5%, 10%, and 15% discounts. Show the formula for each.

Check: verify two lines manually using price minus cost divided by price.

Implementation guide

Governance: the CFO owns the floors and the review, the Head of Sales owns the matrix and approval service levels, Sales Ops owns the workflow and data, and Internal audit or the Controller tests compliance. Report average discount, floor breaches, invoice-to-approval variance, and approval cycle time monthly.

What to expect: the effect on realized price takes two to three quarters, because concessions settle on rebate and renewal calendars.

Downloadable assets

The matrix, exception log, request form, and review agenda above are designed to be copied into a spreadsheet or configured in your quoting tool. The price waterfall builder computes list-to-pocket leakage from your own figures. The EBITDA improvement calculator can translate a change in realized price into EBITDA.

Related resources

• Price Waterfall Analysis: list price to pocket margin, and the cost-side blind spot.
• Why Price Realization Keeps Drifting: why tightening on-invoice discounts relocates the leak.
• Price Increase Playbook: protecting a new price list with an exception budget.
• CPQ and Quote Discipline: where the matrix is enforced in the quoting process.
• Margin Expansion Playbook: pricing in the context of the other margin levers.

Frequently asked questions

What is a discount approval matrix?

A discount approval matrix is a table that states who can approve a discount for each type of deal, up to what level, and with what evidence. It sets authority tiers by discount percentage, discount dollars, and resulting margin against a floor. Its value comes from consistent routing and a record of every exception.

How do I set the maximum discount for a product line?

Compute it from the margin floor: maximum discount equals list margin minus floor, divided by one minus the floor. With a 40% list margin and a 30% floor the maximum is 14.3%. Use your own costs and floors by line, because a single company-wide cap overstates room on thin lines.

Who should approve discounts?

Approve at tiers: a pre-approved band for the rep, then sales manager, a finance or pricing partner, and the CFO for floor breaches. The approver should not be paid on the deal being approved. Include finance at the upper tiers so margin is judged by someone who sees the account and line view.

How do I stop reps from splitting deals to avoid approval?

Apply the dollar trigger by customer over a period instead of by single quote, and test for clusters of deals just under each threshold. Pair the matrix with incentives that reward realized margin. Treat a flagged cluster as a question for Sales Ops to examine, not as proof of intent.

How often should discount governance be reviewed?

Review monthly: average discount, exceptions above tier 1, reps outside the band, invoice-to-approval differences, and off-invoice concessions. Audit the exception log against invoices at 90 days. Realized price responds over two to three quarters, so judge trends, not single months.

How this works in a System of Performance

A discount policy is only as good as the record of what was approved, what was invoiced, and what the customer gave in return. In most companies that record is spread across the CRM, the quoting tool, the ERP, and email. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. If you want to see where your own discount dollars concentrate, bring it to us as a launch partner.