On this page
The deal closed on the last day of the quarter, at 12% off, and the number was approved in an afternoon because the quarter needed it. A year later the renewal conversation opens, and it does not open at list. It opens at the discounted figure, which both sides now treat as the price. The concession was budgeted as a one-time cost of winning the deal. It turned out to be a repricing.
That is the pattern this essay checks against the measured record. The short version: a discount is kept in two memories and billed by a third mechanism. Buyers benchmark the last price paid and punish a price above that benchmark harder than one below it was ever rewarded. Salespeople anchor their next concession on their last one. And the fiscal calendar manufactures discounts on its own schedule, at a measured cost. What the record does grant, and where it runs out, is the useful part: because the folk layer above this question, where I checked it, cited no evidence at all.
Which standard practitioner myths obscure the long-term cost of discounting?
The folk layer says discounts “train customers to wait.” On the day I checked, page one of Google for that claim was a LinkedIn post, a Reddit thread, a Threads post, a Facebook post and five vendor blogs. Not one result cited a study. The single number on the page: “62% of shoppers now wait”: carried no source at all.
Above the folklore sits an advice layer with real pedigree: Harvard Business Review has been telling operators to discipline their discounts since at least “A Case for Discount Discipline” in 2006, and Hermann Simon’s Confessions of the Pricing Man carries the argument at book length. This layer is advice whose evidence base this piece does not audit, and it is anything but evidence-free. “Mind Your Pricing Cues” (2003) was written by Eric Anderson and Duncan Simester, the same researchers whose field experiments appear below. What has been missing, for the negotiated-B2B question, is the assembly: the measured findings, in one place, with their conditions attached. That is the whole job here.
Why do buyers permanently anchor to the lowest negotiated concession?
Reference prices are one of the oldest findings in pricing research: buyers judge a price against an internal benchmark formed, as Hernan Bruno, Hai Che and Shantanu Dutta put it, “from experience of previous prices either through purchase (i.e., paying the price) or observation.” Most of that research watched supermarkets. Bruno, Che and Dutta moved it into negotiated business markets, modelling the transaction records of a UK company selling processed timber to industrial customers: furniture makers, window manufacturers, decking contractors: 10,614 transactions across its most frequently bought products, from customers with repeated purchases, with the large contract accounts of a separate key-account channel excluded. Their reference measure is the price the customer paid last time.
Three findings carry the section. Reference-price effects exist in B2B transactions, on the quantity bought and on the price struck. The reaction is asymmetric: in their quantity model, “the coefficients for LOSS and GAIN are –1.32 (SE = .17) and .348 (SE = .15), respectively”: standardized values, which is why the authors read the gap as loss aversion. And the asymmetry deepens with familiarity: the more transactions a customer has with a salesperson, the stronger the response to a perceived loss and the weaker the response to a gain. The most familiar relationships punish a price above the benchmark hardest.
Figure 1The walk-back costs more than the concession earned
How strongly quantity responds when a B2B customer buys above the benchmark of their last paid price, against buying below it. The bars show absolute size; the signs are opposite.
Source: Bruno, Che & Dutta (2012), Table 4, Model 1 (in-text values, p. 650) — standardized coefficients, quantity equation, transaction records of one UK industrial timber supplier. Directional evidence from one homogeneous-product setting; the authors note (fn. 10) that unobserved lost sales mean loss aversion “may plausibly be stronger” than these estimates. Not a portable multiplier.
Now run the chain one step further, because this is where a discount becomes a ratchet. The concession you granted is a price the buyer paid, so it enters the formation set and becomes the new benchmark, and any later price above it is processed as a loss, the side the data punishes. Their pricing model carries both directions, and the record should be read whole: a buyer’s loss in the past transaction leads to a lower current price, and a past gain to a higher one. So the data does not say that one concession breeds the next by drift; on the gain side it even leans the seller’s way. What it says is narrower and sharper: the trap is the attempted recovery. The concession resets the benchmark; the walk-back is a loss event, punished asymmetrically in quantity; and a buyer who has just suffered such a loss pays less the next time. The teeth of the ratchet are the walk-backs. Assembling those findings into a ratchet is my reading, and the chain has stated limits: their setting is a homogeneous, frequently purchased input where prices are easy to compare, and the estimation excludes the large key-account contracts a reader here may care most about. In heavily customised deals the benchmark blurs, and this mechanism is an extrapolation.
The buyer’s benchmark also connects backward to where a price gets its meaning in the first place: the set the buyer assembles decides the band, and the prices they have actually paid are the strongest members of that set.
How do ad hoc discounts erode internal sales discipline and reference thresholds?
The second memory sits on your side of the table. Dominic Bergers, Mahsa Ghaffari, Giampaolo Viglia and Raffaele Filieri studied discount decisions in B2B software with two experiments and an interview study among B2B salespeople, and their question was precise: does a salesperson’s previous discount shape the discount granted to a new customer: one who never saw the earlier deal? It does. “The findings show the importance of reference discounts in determining the discount level given to a new customer.” Their interviews name the mechanism: uncertainty reduction. When nobody can say what the right number is, the last number stands in for it.
Bruno, Che and Dutta had already found the matching trace in transaction data: salespeople carry reference prices of their own, which shape the prices they strike. So the concession you approved for one account does not stay with that account. It becomes the anchor your team reasons from: at the salesperson level in both studies; whether it travels further through an organisation is not something either one measures. Bergers and colleagues add a lever: incentives reduce the reliance on the anchor. Which sounds like the fix, until you look at what incentives were measured doing elsewhere.
Why does end-of-quarter discounting impose an unmeasured structural tax on future margins?
Paul Oyer showed in 1998 that nonlinear sales compensation: quotas, accelerators, fiscal-year resets: gives salespeople “incentive to manipulate prices, influence the timing of customer purchases, and vary effort over their firms’ fiscal years,” and that manufacturers’ sales bulge at fiscal year-end. Ian Larkin measured the price side at a leading enterprise software vendor: salespeople timed deal closures to an accelerating commission scheme and “agree[d] to significantly lower pricing in quarters in which they have a financial incentive to close a deal, resulting in mispricing that costs the vendor 6%–8% of revenue.” His robustness checks rule out deliberate price discrimination by the vendor as the explanation.
This third mechanism is a standing cost, not a memory. A concession granted in December does not write anything into the calendar; the calendar effect exists because the compensation design does, deal by deal, quarter after quarter. Whether buyers learn the seller’s clock and wait for it is a different claim: the folk layer repeats it, and it is unmeasured in negotiated markets, as the next section shows. What Larkin’s number gives an operator is a price tag for high-powered incentives: a bill you either accept with open eyes or re-design away. Discovering it later, inside the margin bridge, is the worst option.
What does empirical research prove about buyers being trained to delay orders?
The training story has real evidence behind it: in markets with a tape. Four findings mark out what that tape shows.
In consumer categories, the lift is real and the persistence is not. Across 560 Dutch supermarket categories over four years, Vincent Nijs, Marnik Dekimpe, Jan-Benedict Steenkamp and Dominique Hanssens found strong short-run effects of price promotions: an average category elasticity of 2.21, median 1.75, that “rarely exhibit persistent effects,” dissipating in roughly ten weeks on average, with a long-term category impact of “essentially zero.” Koen Pauwels, Hanssens and S. Siddarth, on scanner-panel data for one perishable and one storable category, found “permanent promotion effects are virtually absent” for category incidence, brand choice and purchase quantity alike. A meta-analysis of 51 studies by Devon DelVecchio, David Henard and Traci Freling suggests that, on average, promotions do not affect post-promotion brand preference: with effects that can run either way depending on the promotion and the product.
In field experiments, the long-run sign flips by customer type. Eric Anderson and Duncan Simester ran three large-scale catalog experiments with durable goods and found deeper discounts increased future purchases by first-time customers and reduced them for established customers: with “increased deal sensitivity” among their named mechanisms, and the warning that prices set on short-run elasticity will be too low for an established base. This is consumer evidence. I refuse it as proof of B2B damage, so I refuse it as proof of B2B acquisition benefit too; it marks the direction the conditions point, nothing more.
Where B2B has a tape, timing can at least be read. Manufacturers promoting through retailers is a business-to-business market with decades of scanner data behind it: the record Robert Blattberg, Richard Briesch and Edward Fox synthesized into the field’s empirical generalizations. Forward buying: purchases pulled ahead of a deal: is among the mechanisms Anderson and Simester name in their own experiments, and it is on that trade tape, not in the negotiated world, that such timing is even readable. The one dynamic B2B study in this essay’s source base cuts both ways: Jonathan Zhang, Oded Netzer and Asim Ansari found buyers moving between a vigilant state, with heightened price sensitivity, and a relaxed one, and “the seller’s pricing decisions can transition buyers between these two states.” A concession can calm as well as train.
In the negotiated world: software, services, custom components: I found no persistence tape at all. No study I could locate measures whether negotiated concessions train buyers to wait across renewals. That is a bounded claim about a search, this campaign, this year; it is not a law. But it inverts the usual burden. The demand-side damage story is unproven in exactly the negotiated markets it is told about: while the two memories and the calendar bill, the mechanisms above, are measured. Discount policy should stand on the measured set.
| The mechanism | What is measured | Where it stops |
|---|---|---|
| The buyer’s benchmark | Reference effects on price and quantity; losses outweigh gains; a past loss lowers the next price | One homogeneous-product B2B setting; customised deals are an extrapolation |
| The seller’s anchor | Past discounts shape the discount to a new customer; incentives reduce the effect | Salesperson level, experiments plus interviews; organisational spread unmeasured |
| The calendar’s bill | Lower pricing in incentive quarters: 6–8% of revenue at one vendor; year-end sales bulges | A structural cost of comp design, not a memory of any one deal |
| ”Trained to wait” | Real where a tape exists: consumer categories, retailer forward buying | No persistence record found for negotiated B2B: a bounded search claim |
Table 1One concession, two memories, one bill
What the measured record actually supports about a negotiated discount, and where each finding stops.
Source: Author’s own assembly of the cited studies; each row’s conditions are in the text. Not a finding of any single paper.
Under what specific operational constraints can pricing concessions be commercially justified?
A fair reader now objects three times, and each objection deserves its full weight.
First: discounting is price discrimination, and price discrimination is how a seller captures willingness-to-pay that a single price leaves on the table. Correct. Nothing above argues against a designed concession structure: different prices for different segments, committed volumes, terms. The evidence indicts a narrower thing: the unpriced concession, granted ad hoc, tracked nowhere, timed by the seller’s quarter rather than the buyer’s case.
Second: we discount because the competitor does. Also real. In the consumer category data, the authors’ own conclusion is that promotions’ power lies in “the preservation of the status quo”: which I read as defence, at category level. But notice what a defensive concession is: a price for something, namely share held against a named attack. It belongs on the designed side of the line, priced and recorded like everything else there.
Third: comp design can fix the anchoring: Bergers and colleagues measured incentives reducing it. True, and the same instrument generates Larkin’s bill. Incentives move the anchor problem; they do not abolish the register problem. Hold both findings at once and the conclusion is narrower and more useful: pay for margin rather than volume where you can, and price the timing games you keep.
How can enterprise sellers govern the discount waterfall without public transparency?
The waterfall must have an owner
The diagnosis above asks why a concession can outlive the deal it was meant to close. The governance extension asks a different question: can the organisation reconstruct the price it actually realised, and can somebody explain every deduction?
The waterfall is not a decorative bridge from list to pocket. It is a record with a unit and a time boundary. Start with the declared list or reference price. Record the quoted price and each explicit discount. Add rebates, credits, freight or service allowances, payment terms, currency effects, and any other deduction that changes the realised exchange. Then record pocket price, the cost-to-serve boundary, and the declared contribution boundary separately. A lower pocket price is an observation; whether it was a good decision is a second question.
Marn and Rosiello’s pocket-price framing gives the waterfall its diagnostic lineage (Marn & Rosiello, 1992). Bergers and colleagues show why the seller-side reference discount belongs in the record, not only in a training conversation (Bergers et al., 2023). Oyer shows that nonlinear fiscal incentives can shape business seasonality and timing, while the exact mechanism and setting remain part of that study’s boundary (Oyer, 1998). Zbaracki and colleagues remind us that changing a price produces work beyond the mechanical menu update (Zbaracki et al., 2004).
Each deduction should therefore carry an owner, approval, counterpart, reason, effective date, expiry, and booking rule. The time series matters. A concession that appears once may be a bounded trade. The same deduction recurring across renewals may be a new reference price. A credit booked in one system but omitted from the customer-level bridge is not a small data defect. It changes the object being governed.
The illustrative 100-unit bridge below is a composition, not a target. Its first row is the 100-unit reference amount; its second row shows 82 units of pocket price after 10 units of discount, 5 units of rebate, and 3 units of credit. Replace it only with a locally defined, receipted dataset. The useful question is whether the bridge reconciles, has an owner, and changes the next approval decision.
Figure 2The 100-unit pocket-price bridge
The two rows keep the reference amount visible beside each deduction and make the reconciliation testable without presenting the values as a benchmark.
Source: Author's illustrative composition grounded in Marn & Rosiello (1992). Values are synthetic units, not a benchmark, recommendation, or observed ratio.
The fence rule follows from the two memories. If a concession must be given, tie it to a counterpart that costs the buyer something: committed volume, a longer term, faster payment, a named reference. The test of a real fence is exactly that cost. A counterpart the buyer would have provided anyway is decoration, and the price you fenced is simply the new price. Honesty requires a disclosure here: the fence rule is the pricing canon’s standard advice, and I found no study that measures it in negotiated markets. It has precisely the evidentiary status of the folklore this essay opened with. The difference is that you can test it: on your own book.
How should sales leadership extract true historical discount costs from enterprise billing data?
After one cycle of keeping the register, pull three columns per account: realized price against list, quantity, and the recorded counterpart. Include the accounts that churned: reading only survivors flatters every fence. Check the list-price history first: a “discount” against a list that quietly moved is drift, and it will masquerade as memory. Then read the trajectory: where a fenced concession lapsed, did the realized price step back up, and did the quantity hold? Where it did, your fences are real, and this essay’s warning is about somebody else’s book. Where price stayed down, volume thinned, or the next negotiation opened at the concession, you are looking at the ratchet: the buyer’s benchmark and your team’s anchor, doing what the record says they do.
The quarter that needed the deal will not remember the discount. The renewal will. The concession was never a moment; it was an entry: written into the buyer’s benchmark, copied into your team’s next offer, while the calendar keeps billing on its own schedule. The register is where you write it down yourself, before those three do it for you.
One entry the register cannot hold, because it is not a concession on the number at all: what you gave away when you chose the unit. A flat rate hands the customer’s usage swings to you, and the buyers who want that hardest are the ones with the biggest swings.
Where are the analytical boundaries of empirical discounting studies?
Boundary. The studies explain mechanisms of reference dependence and seller incentives, not a universal persistence rate for every B2B discount. Track discount history against renewal and margin in the relevant cohort.
Evidence base. The analytical frame also draws on these additional sources: Anderson and Simester 2003; Anderson and Simester 2004; Blattberg et al. 1995; DelVecchio et al. 2006; Marn and Rosiello 1992; Nijs et al. 2001; Oyer 1998; Pauwels et al. 2002; Simon 2015; Zbaracki et al. 2004; Zhang et al. 2014. The links identify the exact works; they support the mechanisms and boundary conditions discussed here, not every claim in isolation.
References
- Anderson, E. T., & Simester, D. (2003, September). Mind your pricing cues. Harvard Business Review, 81(9), 96–103. https://hbr.org/2003/09/mind-your-pricing-cues
- Anderson, E. T., & Simester, D. I. (2004). Long-run effects of promotion depth on new versus established customers: Three field studies. Marketing Science, 23(1), 4–20. https://doi.org/10.1287/mksc.1030.0040
- Bergers, D., Ghaffari, M., Viglia, G., & Filieri, R. (2023). Choosing the discount size in the software industry: How to incentivise the salesforce. Industrial Marketing Management, 109, 232–244. https://doi.org/10.1016/j.indmarman.2023.02.002
- Blattberg, R. C., Briesch, R., & Fox, E. J. (1995). How promotions work. Marketing Science, 14(3, Suppl.), G122–G132. https://doi.org/10.1287/mksc.14.3.G122
- Bruno, H. A., Che, H., & Dutta, S. (2012). Role of reference price on price and quantity: Insights from business-to-business markets. Journal of Marketing Research, 49(5), 640–654. https://doi.org/10.1509/jmr.09.0334
- DelVecchio, D., Henard, D. H., & Freling, T. H. (2006). The effect of sales promotion on post-promotion brand preference: A meta-analysis. Journal of Retailing, 82(3), 203–213. https://doi.org/10.1016/j.jretai.2005.10.001
- Larkin, I. (2014). The cost of high-powered incentives: Employee gaming in enterprise software sales. Journal of Labor Economics, 32(2), 199–227. https://doi.org/10.1086/673371
- Marn, M. V., & Rosiello, R. L. (1992, September–October). Managing price, gaining profit. Harvard Business Review, 70(5), 84–94. https://hbr.org/1992/09/managing-price-gaining-profit
- Nijs, V. R., Dekimpe, M. G., Steenkamp, J.-B. E. M., & Hanssens, D. M. (2001). The category-demand effects of price promotions. Marketing Science, 20(1), 1–22. https://doi.org/10.1287/mksc.20.1.1.10197
- Oyer, P. (1998). Fiscal year ends and nonlinear incentive contracts: The effect on business seasonality. The Quarterly Journal of Economics, 113(1), 149–185. https://doi.org/10.1162/003355398555559
- Pauwels, K., Hanssens, D. M., & Siddarth, S. (2002). The long-term effects of price promotions on category incidence, brand choice, and purchase quantity. Journal of Marketing Research, 39(4), 421–439. https://doi.org/10.1509/jmkr.39.4.421.19114
- Simon, H. (2015). Confessions of the pricing man: How price affects everything. Copernicus. https://doi.org/10.1007/978-3-319-20400-0
- Zbaracki, M. J., Ritson, M., Levy, D., Dutta, S., & Bergen, M. (2004). Managerial and customer costs of price adjustment: Direct evidence from industrial markets. The Review of Economics and Statistics, 86(2), 514–533. https://doi.org/10.1162/003465304323031085
- Zhang, J. Z., Netzer, O., & Ansari, A. (2014). Dynamic targeted pricing in B2B relationships. Marketing Science, 33(3), 317–337. https://doi.org/10.1287/mksc.2013.0842
Updated 3 changes
- Added a governed pocket-price waterfall with a synthetic 100-unit bridge. The exhibit makes unit, time boundary, ownership, approval, and reconciliation explicit without importing a benchmark or the guarded pocket-price lever.
- Source follow-up on 28 August 2026: The Ehrenberg, Hammond and Goodhardt sentence and reference were removed because the available official publisher record is an abstract page, not a lawful complete article. The claim is not carried without a full-text receipt.
- Refresh on 20 August 2026: The studies explain mechanisms of reference dependence and seller incentives, not a universal persistence rate for every B2B discount. Track discount history against renewal and margin in the relevant cohort.
