Price is the strongest profit lever, and the least managed
A price change flows almost entirely to profit, because costs stay the same. Extra volume has to pay its own variable costs first. Yet most companies manage volume daily and price once a year.
Source: Marn & Rosiello (1992), “Managing Price, Gaining Profit”, Harvard Business Review. Average economics of 2,463 companies.
The article has three parts: what price to set, how to raise it without losing volume, and what a promotion really does.
1What price should you set?
The price on the tag is not the price you earn
| Step | Per $100 of list price |
|---|---|
| List price | $100 |
| Invoice discount | −$8 |
| Volume rebate | −$5 |
| Payment terms | −$3 |
| Promo funding | −$4 |
| Pocket price | $80 |
Every concession looks small. Together they take 20% of the price. Track the pocket price, not the list price.
Illustrative example. The price waterfall concept: Marn & Rosiello (1992), Harvard Business Review.
Three anchors: cost sets the floor, value sets the ceiling
| Anchor | Role | What it does |
|---|---|---|
| Your cost | the floor | Cost-plus looks only here. It protects margin but never asks what the customer will pay. |
| Competitors | the reference | Competitor-based pricing copies the middle: useful as a reference, but it copies their mistakes too. |
| Customer value | the ceiling | Value-based pricing starts here. Strategy decides how much of the value you capture. |
Cost-plus never asks what the customer will pay
- Cost $180 + $120 margin = $300
- Competitor charges $330, so we charge $315
- The question “what is it worth to the buyer?” is never asked
- The product saves the buyer $1,600 a month: $300 captures under a fifth of the value
- If it saved only $150 a month, $300 would already be too much
- First value, then price
The pricing pyramid: the number sits on top of four decisions
- Value creation. What does the customer actually get?
- Price structure. What exactly does the customer pay for?
- Value communication. Does the customer see the value?
- Pricing policy. Who may negotiate, and how far?
- Price level. How much?
Source: Nagle & Müller, The Strategy and Tactics of Pricing.
A B2B software example:
| Layer | Example |
|---|---|
| Value creation | The product saves 80 manager-hours a month, about $1,600 |
| Price structure | $300 per company, or $30 per user, or 2% of transactions |
| Value communication | “You spend 120 manager-hours a month on this. We cut it to 40.” |
| Pricing policy | 5% for annual prepayment, 7% for volume, 10% for a reference case, 10% only with the commercial leader’s approval |
| Price level | $300 per month |
The structure alone can change the economics of the business without touching the product.
Perceived value = perceived benefits ÷ perceived sacrifices
- The price itself
- Time to implement
- Risk it will not take hold
- Training the team
- Justifying a new vendor to the boss
- Time and money saved
- Lower risk of errors
- Peace of mind for the team
- Status of using a modern tool
Source: Kent Monroe, Pricing: Making Profitable Decisions. Price also acts as a quality signal, not only as a barrier.
Elements of value: what customers pay for besides function
Thirty elements of value in four tiers: functional (14), emotional (10), life-changing (5) and social impact (1). Most useful for consumer brands: for each customer segment, pick the 3–4 elements people really pay for, and price and communicate against them.
Source: Almquist, Senior & Bloch (2016), “The Elements of Value”, Harvard Business Review.
Seven ways to find a price, and what each one cannot tell you
| Method | Question it answers | Main limit |
|---|---|---|
| Cost-plus | What must I charge to cover cost? | Ignores the customer |
| Competitor-based | What is the market reference? | Copies others’ mistakes |
| Value-based | What is it worth to the buyer? | Needs customer data |
| Van Westendorp | What price range feels acceptable? | Stated, not real, behaviour |
| Gabor–Granger | How does stated demand move with price? | Stated, not real, behaviour |
| Conjoint | Which features justify which price? | Costly; hypothetical choices |
| Live price test | What do customers actually do? | Needs traffic; fairness risk |
Van Westendorp: four questions give you an acceptable range
Ask each respondent four times: at what price would the product be…
- too cheap: so cheap I doubt the quality
- a bargain: cheap, but still a good deal
- getting expensive: I would still consider it
- too expensive: I would not consider it
Plot the cumulative curves. Where they cross marks the range of acceptable prices. It is a range to test, not the price to charge.
Source: Van Westendorp (1976), “NSS Price Sensitivity Meter”, ESOMAR Congress.
Surveys give a range. Behaviour gives the price
- Van Westendorp, Gabor–Granger, conjoint
- Fast and cheap to run
- Good for a first range and for ranking features
- People answer differently when no money is at stake
- Price tests with real customers
- A/B test, test region or test store
- Cleanest on new customers or new tiers
- Run it long enough for the full buying cycle
Basis: Miller, Hofstetter, Krohmer & Zhang (2011), Journal of Marketing Research: stated willingness to pay depends on the measurement method; methods with real money at stake track purchases better.
Price elasticity: use the average only as a sanity check
On average, a 1% price rise cut unit sales by about 2.6%. Individual products vary widely, so measure your own SKUs and segments.
Source: Bijmolt, van Heerde & Pieters (2005), “New Empirical Generalizations on the Determinants of Price Elasticity”, Journal of Marketing Research.
A price rise pays if volume falls by less than this
| Price increase | Margin 20% | Margin 30% | Margin 40% | Margin 60% |
|---|---|---|---|---|
| +3% | 13.0% | 9.1% | 7.0% | 4.8% |
| +5% | 20.0% | 14.3% | 11.1% | 7.7% |
| +10% | 33.3% | 25.0% | 20.0% | 14.3% |
Cells show the maximum volume loss that still leaves profit unchanged. Example: +5% on a 40% margin breaks even at an 11.1% volume loss.
2How to raise prices without losing volume
Slowly, selectively, and with the mechanics of perception on your side.
Companies keep only about a third of the increases they plan
The planned increase and the realised increase are two different numbers. Manage and report the second one.
Source: Simon-Kucher & Partners, Global Pricing & Sales Study 2017; “Why 97% of All Price Increases Fail” (Daus, 2017).
Why customers notice: three mechanisms of price perception
- Reference price. Customers remember the old price. The new price is judged against what they paid before, not against your cost.
- Just noticeable difference. Perception of change is proportional to the price level (Weber). Below the threshold, a change goes unnoticed.
- Loss aversion. Losses weigh more than equal gains, so a rise hurts more than the same discount pleases.
Gradual does not mean invisible: repeated small steps reset the reference price. The goal is a new normal the customer accepts, not a trick they miss.
Basis: Monroe (pricing and perceived value); Kahneman & Tversky (1979), Prospect Theory; Weber’s law.
Customers accept a rise they judge fair
- Profit-driven, silent, opportunistic
- Exploits a moment of high demand or no alternatives
- Looks like a pure margin grab
- Arrives without notice or reason
- Cost-driven, explained, shared
- Follows a visible cost increase
- The company shares some of the pain
- Reason, timing and added value are explained
Basis: Kahneman, Knetsch & Thaler (1986), “Fairness as a Constraint on Profit Seeking”; Campbell (1999), “Perceptions of Price Unfairness: Antecedents and Consequences”, Journal of Marketing Research.
The gradual ladder: six moves, in this order
Ways to raise revenue per customer without moving the tag
| Lever | How it works | Watch out for |
|---|---|---|
| Smaller discounts | Narrow the gap between list and pocket price | Sales pushback; track win rate |
| Good–better–best tiers | Move customers to a higher-value tier | Tiers must differ in real value |
| Usage or per-seat pricing | Price scales with value received | Customers fear unpredictable bills |
| Fees and extras | Charge separately for optional services | Hidden fees destroy trust |
| Pack-size reduction | Same price, less product | Less noticed, but trust risk if discovered |
On pack size: shoppers react to size cuts only about a quarter as strongly as to an equal price rise (ice cream).
Source: Çakır & Balagtas (2014), American Journal of Agricultural Economics.
Case: Costco raises fees rarely, modestly and with extra value
- About 8–9% each time, seven years apart.
- In 2024 the cap on the Executive 2% reward rose from $1,000 to $1,250, so value rose alongside the fee.
- A loyal base and a clear promise made the increase feel earned, not extracted.
Source: Costco Wholesale Form 10-K FY2024; Costco announcements and press coverage, 2017 and 2024.
Case: JCPenney removed discounts overnight and lost customers
Shoppers had learned the price from coupons and sales. “Fair everyday prices” removed the discount they used as their reference. Dropping a promotion cold is a price increase, too.
Source: JCPenney quarterly and annual results, 2012; contemporary press coverage.
How to announce an increase
Sales needs the same four answers before the announcement, not after the first angry email.
Set the stop rule before you raise the price
| Indicator | What to watch | Decision rule (set in advance) |
|---|---|---|
| Win rate or conversion | Drop versus the last 4–8 weeks | Pause if it falls below the break-even level |
| Discount requests per deal | Rising share of deals with exceptions | Tighten approval; do not let the rise leak |
| Cancellations by cohort | Churn of customers hit by the rise | Offer options; stop if churn exceeds the plan |
| Realised (pocket) price | Versus list price and the plan | Report this, not the list-price change |
| Volume by segment | Where the loss concentrates | Adjust the segment, keep the rest |
The break-even volume loss from the earlier table is the line between a good rise and a bad one.
3Promotions: growth, or moved demand?
What a promotion really does, and how to tell.
Every promotion bump has four components
| Component | Is it growth? | What it is |
|---|---|---|
| Incremental sales | Real growth | Extra volume that would not have happened without the promotion. |
| Brand switching | Not growth | The buyer took the same product from a competitor, or from your other SKU, at a lower price. |
| Purchase acceleration | Not growth | The buyer bought earlier, or stockpiled, what they would have bought later. |
| Post-promo dip | The bill | Sales fall below normal right after the promotion ends. |
Only about a third of a promotion bump comes from other brands
Earlier studies had suggested around three quarters of the bump was switching. A more careful decomposition of store data found the figure was one third. Most of the rest was shifted in time: customers bought sooner or bought more.
That does not make promotions profitable. Pulled-forward sales come back as a post-promo dip, and the margin given away is gone.
Source: van Heerde, Gupta & Wittink (2003), “Is 75% of the Sales Promotion Bump Due to Brand Switching? No, Only 33% Is”, Journal of Marketing Research.
The discount math: how much extra volume pays for a discount?
| Discount | Margin 30% | Margin 40% | Margin 50% |
|---|---|---|---|
| 10% off | +50% | +33% | +25% |
| 20% off | +200% | +100% | +67% |
| 30% off | never pays | +300% | +150% |
Cells show the extra unit volume needed just to keep profit flat. Example: 20% off on a 40% margin needs +100% volume.
Units up 80%, profit down: a promotion in numbers (illustrative)
| Normal week | Promotion week | |
|---|---|---|
| Price, unit cost $6 | $10 | $8 (20% off) |
| Units sold | 1,000 | 1,800 (+80%) |
| Gross profit | $4,000 | $3,600 |
| Later dip | — | −$1,200 (300 units × $4) |
Celebrating the peak week hides the loss. Count the baseline, the margin given away and the dip afterwards.
Measure the lift against a baseline, not against zero
| Method | How it works | Strength |
|---|---|---|
| Holdout group | Some stores or customers get no promotion | Strong |
| Geo experiment | Promote in one region, compare with a similar one | Strong |
| A/B test | Random split of online traffic | Strong |
| Before and after | Compare sales before and after the promotion | Weakest: seasonality and trend |
In large Facebook field experiments, non-randomised methods mis-estimated ad effects by a factor of three or more in about half of the studies.
Source: Gordon, Zettelmeyer, Bhargava & Chapsky (2019), Marketing Science.
Promotion mechanics: consumer versus B2B and distribution
- Volume discount
- Gift with purchase
- Limited time offer
- Bundles
- Loyalty programme
- Volume price scale
- Retro bonus: a discount after the fact for period volume
- Buy-back of unsold stock
- Payment delay as a hidden discount
- Co-op marketing budget
In B2B, count payment terms and retro bonuses as part of the promotion budget.
4Sales discipline: discounts as a tool, not a gift
In the room: quantify the value, then name the price
- “We have an AI platform with 37 features, API integration and a new analytics module.”
- The buyer cannot see why this costs $300.
- “You spend about 120 manager-hours a month on this. We cut it to 40. At $20 an hour that is $1,600 saved every month.”
- Now $300 looks like a fraction of the gain.
Price psychology: what is proven, and where it stops
| Effect | Source | Where it stops |
|---|---|---|
| Anchoring | Tversky & Kahneman 1974; Ariely, Loewenstein & Prelec 2003 | Fails if the buyer knows the market price |
| Decoy effect | Huber, Payne & Puto 1982; Frederick, Lee & Baskin 2014 | Needs a third, dominated option; weaker than often claimed |
| Charm pricing (.99) | Anderson & Simester 2003 (field experiments); Snir, Levy & Levy 2017 | Weaker for considered B2B buys; clashes with premium |
| Pain of paying | Prelec & Loewenstein 1998; Gourville 1998 | Small instalments help; hiding the total backfires |
Test any effect on a small sample before scaling. It is cheaper than believing a blog.
Eight questions before you change a price
- What value is the customer paying for?
- What is the measured elasticity, not the assumed one?
- How much volume can I lose before it hurts?
- Which reference price will customers compare with?
- Is the reason fair, and have I said it?
- What is my stop rule?
- What is the baseline for measuring the result?
- Start small: which segment goes first?
Key takeaways
- Start from value, not cost. Cost sets the floor. The customer’s value sets the ceiling.
- Raise slowly, selectively, fairly. Fix leakage, start where you are underpriced, explain the reason.
- Do the break-even math. Both for price rises and for discounts, before the decision.
- Measure incremental profit. Against a baseline, with the post-promotion dip included.
Sources and further reading
- Marn & Rosiello (1992). Managing Price, Gaining Profit. HBR.
- Nagle & Müller. The Strategy and Tactics of Pricing. Routledge.
- Almquist, Senior & Bloch (2016). The Elements of Value. HBR.
- Van Westendorp (1976). NSS Price Sensitivity Meter. ESOMAR.
- Miller, Hofstetter, Krohmer & Zhang (2011). How Should Consumers’ Willingness to Pay Be Measured? JMR.
- Bijmolt, van Heerde & Pieters (2005). New Empirical Generalizations on the Determinants of Price Elasticity. JMR.
- Simon-Kucher & Partners (2017). Global Pricing & Sales Study.
- Kahneman, Knetsch & Thaler (1986). Fairness as a Constraint on Profit Seeking. AER.
- Campbell (1999). Perceptions of Price Unfairness: Antecedents and Consequences. JMR.
- Çakır & Balagtas (2014). Consumer response to package downsizing. AJAE.
- van Heerde, Gupta & Wittink (2003). Is 75% of the Sales Promotion Bump Due to Brand Switching? JMR.
- Gordon, Zettelmeyer, Bhargava & Chapsky (2019). A Comparison of Approaches to Advertising Measurement. Marketing Science.
- Books: Simon, Confessions of the Pricing Man; Simon & Echter, Beating Inflation (2022).