Routing Deep Discount Approvals to the Right Person

For example, a flat rule of "anything over 10% needs manager approval" treats a small deal and a strategic enterprise account the same way, even though the dollar impact of a given discount percentage is wildly different between them. It also creates a single number every rep in the org learns to work right up against, which turns the threshold into a target rather than a genuine control.

Tiers fix both problems by scaling scrutiny to actual risk: a small discount on a small deal barely matters and shouldn't need a manager's time, while a deep discount on a large account deserves a specific person who owns margin actually looking at it before it goes out.

A workable structure usually has three levels: a self-serve tier a rep can apply without approval (small, standard discounts within a defined band), a manager-approval tier for anything deeper that still doesn't materially threaten margin, and a finance or deal-desk tier for discounts steep enough to need someone who owns the company's margin target directly. Where exactly the lines sit depends on your product's margin structure, not a generic industry rule.

The tiers should be based on the discount's actual dollar or margin impact, not just the percentage: for example, a percentage-only rule treats a small deal's 20% discount the same as an enterprise deal's 20% discount, when the second one might represent real money the first one doesn't.

A workable tier structure settles these points:

• Define a self-serve tier for small, standard discounts within a set band that reps can apply without asking anyone.
• Add a manager-approval tier for deeper discounts that still don't materially threaten margin on the deal.
• Send the steepest discounts to finance or a deal desk, where someone owns the company's margin target directly.
• Base tier boundaries on dollar or margin impact, not percentage alone, so a large account and a small deal aren't treated alike.
• Set a turnaround target for each tier so slower approvals don't push reps to structure deals around them.

Without automated routing enforced in the CPQ itself, approval becomes a matter of whoever the rep decides to ask, and reps learn quickly which approvers say yes fastest with the least scrutiny. A rep chasing quarter-end can route a deep discount to whichever manager is most likely to rubber-stamp it, and the tiered structure exists on paper without actually constraining anything.

Enforcing the tiers as a hard rule in the CPQ, where a quote literally can't generate without the right approval logged against the right tier, closes that gap. It's the difference between a policy and a control.

A tiered system that routes a deep discount to finance and then sits for four days waiting on a response just moves the bottleneck instead of removing it, and it pushes reps to structure deals specifically to avoid the slow tier. Set an explicit turnaround target for each tier, and make sure the approver group actually has the bandwidth to meet it, especially during the last week of a quarter when discount requests spike.

A fast, reliable approval process is what keeps reps from trying to route around the system. A slow one guarantees they will.

One more design choice worth getting right: what happens to a request that's still pending when the deal's close date arrives. Letting a deal quietly slip past its target date while waiting on approval is its own kind of failure, even if the approval eventually comes through correctly, since the delay itself becomes a cost nobody accounted for.

Set up a periodic review of every approved exception, not just the ones that get escalated or disputed. If one manager is approving a noticeably higher share of requests than their peers, or approving them faster than seems plausible given real review, that's worth a direct conversation before it becomes the default way reps route around real scrutiny.

The audit doesn't need to be adversarial. Often a manager who's rubber-stamping requests doesn't realize their approval pattern looks different from the rest of the team until someone shows them the comparison.

A rep needed to offer a discount just over the self-serve tier's ceiling to close a deal before quarter-end. Rather than routing it to their manager, they split the deal into two line items, each individually discounted just under the self-serve threshold, effectively delivering the same total discount without triggering the approval step the structure was designed to require.

The pattern only surfaced during a routine audit of approved (or in this case, never-submitted) exceptions, when someone noticed the unusually specific line-item split. The fix wasn't punishing the rep. It was tightening the tier logic to evaluate the total discount across a deal's line items, not each line individually, which closed the exact gap the workaround had found.