Sign in

Blog · Board and management reporting

Discount approval thresholds from price realisation data, not from a round number

How a commercial team sets the discount level at which a rep needs approval from what its own invoice lines show: the distribution of discounts given, the discount at which win rate stops improving, the reps and product lines where discounts cluster just under the current threshold, the revenue given away between the effective and the agreed discount, and the threshold per product line that follows, with the approval volume it will generate.

The short answerMost discount thresholds are a round number someone chose. The invoice lines and the outcomes show where the threshold should be: the distribution of discounts actually given, per product line; the discount level beyond which win rate no longer rises, from the outcomes; and the cluster of deals just under the current threshold, which is the threshold being gamed. The new threshold per line sits where the win-rate curve flattens, and the report says how many approvals a month it will generate, so the policy is set with the workload known.

A discount approval threshold of 15 percent was set in a meeting years ago. The invoice lines show a third of all discounts at 14.5, a win rate that stops improving at 10, and four hundred thousand dollars a year given away between the two. This guide sets out the threshold from the data: the distribution, the win-rate curve, the cluster, and the approval volume.

The measures

Per product line:

Discount distribution = share of invoiced lines in each discount band Win rate by discount band, from outcomes joined to the proposed discount Flattening point = the band beyond which win rate rises by less than a stated amount Cluster = share of lines within one point below the current threshold Revenue between = value of discount given above the flattening point

Per rep: the same, to find who prices to the threshold.

The rows you need

  • Invoice lines: line, product, customer, rep, list value, invoiced value.
  • Outcomes: opportunity, product, proposed discount, won or lost.
  • Current policy: threshold per line or overall.

Customer and rep identifiers only.

A worked distribution

Line A. Current threshold 15 percent.

Discount band Share of lines Win rate
0 to 5% 18% 24%
5 to 10% 22% 38%
10 to 12% 15% 41%
12 to 14% 11% 42%
14 to 15% 31% 42%
Over 15% (approved) 3% 44%

Win rate flattens at 10 to 12 percent. Nearly a third of lines sit in the last point below the threshold, at the same win rate as 10. Everything between 12 and 15 on those lines is discount that bought nothing.

The revenue between

Line Lines between flattening and threshold Discount given above flattening Annual
A 4,200 3 pts average $410,000
B 1,900 2 pts $90,000
C 800 0; flattens at the threshold none

The new thresholds, with the workload

Line Current Proposed Approvals per month at proposed Revenue protected
A 15% 12% 140 $410,000/yr
B 15% 13% 40 $90,000/yr
C 15% 15% unchanged

A hundred and forty approvals a month on line A is a workload the sales director accepts or delegates; the report says it before the policy changes rather than after.

Per rep

Rep Share of lines in the cluster Average discount, line A
R-04 58% 14.6%
R-11 12% 9.1%

Rep R-04 prices to the threshold. The conversation is about that, with the win-rate curve showing it buys nothing.

Where it goes wrong

One threshold for all lines. Too tight for one; too loose for another.

No win-rate curve. The threshold is a guess about what discount wins.

Cluster ignored. The threshold is the price.

Policy changed without the workload. Approvals flood; the threshold is quietly ignored.

Every year, thresholds from the lines

Mapped once, the invoice lines, the outcomes and the policy produce the distribution, the win-rate curve, the cluster, the revenue between and the proposed thresholds with their workload. Covirage builds this from the exports as they are. The metrics governance solution describes the setup, and the price realisation guide covers the unagreed discount the new threshold reduces.

Questions people ask

Why per product line?

Because the discount that wins a commodity line is different from the one that wins a specialised one, and one threshold across both is too tight for one and too loose for the other. The invoice lines carry the product; the threshold follows it.

What is the cluster under the threshold?

If the approval threshold is 15 percent and a third of all discounts are between 14 and 15, reps are pricing to the threshold rather than to the deal. The cluster is visible in the distribution, and it is the strongest evidence that the threshold is the price, not the ceiling.

How is the win-rate curve computed?

From outcomes joined to the discount on the proposal or invoice: win rate by discount band, per line. It usually rises with discount and then flattens, and the flattening point is where extra discount buys nothing. The threshold goes there, or slightly below.