Customers, Competition, and Pricing

More customers, a worse result

A small company sells appointment scheduling and customer reminders to other businesses. It has 300 subscribers paying $120 per month. Its sales manager proposes cutting the price to $110 and forecasts that subscriptions will increase to 340. More customers and more revenue sound like progress.

Serving a subscriber costs $40 per month. Does the proposed cut improve the business’s position?

The company and its numbers are fictional. We will compare two prices for the same service over the same month. At $120, the company would have 300 subscribers during that month. At $110, the forecast keeps all 300 and adds 40. Every subscriber pays the new price immediately; there are no customers still paying the old rate.

Assume spare capacity, unchanged quality, and no additional advertising, setup, or fixed cost. The $40 covers every incremental service cost at these volumes. There is no displaced business or difference in future returns, taxes, or financing. These assumptions isolate the pricing decision. They are questions to investigate in a real proposal, not facts about subscription businesses generally.

Chapter 2 asked whether an offered price covered the resources a decision would use. Here the firm chooses the price, and customers decide whether to buy. Costs alone cannot tell the manager how those customers will respond.

Customers compare alternatives

The service is worth buying only if it leaves a customer better off than the available alternatives. Those alternatives could include another provider, handling bookings internally, or accepting the inconvenience of doing without the service. A customer values the time saved or mistakes avoided, not simply the seller’s effort in producing it.

Willingness to pay is the most a customer would pay for the specified service rather than take the relevant alternative. It need not be identical across customers. A business with frequent appointment changes may value reminders more than a business with a few regular clients. That is a possible source of different valuations, not evidence about the customers in our numerical case.

A close substitute limits what the seller can charge. If another service supplies the same useful features with little effort required to switch, a small price difference may matter. If switching requires staff training or moving records, the customer must compare those costs with the saving. A cheaper rival is not automatically the cheaper complete alternative.

Time matters for the same reason. A customer unable to reorganize bookings this week may be able to switch before the next renewal. A manager should therefore ask whose response is being forecast and over what period. A modest immediate loss of customers does not establish that a price increase will remain attractive after they have time to adjust.

The firm’s customers are not the whole market

Foundations used a market containing all sellers of a defined product. Our 300 and 340 subscriptions belong to one seller. Its price cut could attract customers from rivals, bring previously unserved businesses into the market, or do both. An increase in this company’s sales need not be an equal increase in total market sales.

For now, hold rival offers, advertising, customer needs, and service quality fixed. The two price–quantity pairs describe the firm’s own conditional demand schedule. Changing its price while those conditions remain fixed is a movement along that schedule, not a shift in demand.

If a rival introduces a better service at the same price, customers may buy less from our provider at every price it could charge. That changes the schedule itself. The original forecast cannot simply be reused as though nothing else happened.

Likewise, an increase in an input cost across the market may put upward pressure on market prices, but it does not guarantee that this firm can pass through the full increase without losing customers. Rivals may have different costs, customers may substitute away, and adjustments may take time. “Our costs went up” explains why management wants more revenue. It does not establish customers’ willingness to provide it.

Put the complete price comparison on paper

At $120, monthly revenue is 300 × $120 = $36,000. Variable service cost is 300 × $40 = $12,000. Subtracting that cost leaves $24,000.

At $110, forecast revenue is 340 × $110 = $37,400. Variable service cost becomes 340 × $40 = $13,600. That leaves $23,800.

Here contribution means revenue minus variable service costs, before fixed costs. Table 3.1 compares the two complete scenarios.

On a small screen, scroll the table sideways to see all columns.

Revenue increases by $1,400, but serving the additional customers costs $1,600. Contribution therefore falls by $200. With all other relevant costs and benefits unchanged, profit also falls by $200. The contribution totals themselves are not total profit: existing fixed costs still have to be deducted.

There is another useful way to see the result. Each retained customer now pays $10 less, reducing contribution by 300 × $10 = $3,000. Each of the 40 additional customers contributes $110 − $40 = $70, adding $2,800. Thus:

Change in contribution = −$3,000 + $2,800 = −$200.

Counting only the new customers overlooks the lower price paid by existing customers. Counting only the discount overlooks the new business. Both changes belong in the decision.

How responsive are customers?

Forty additional customers tells us the size of the forecast change, but not its size relative to the existing business or the price reduction. Price elasticity of demand compares the percentage change in quantity demanded with the percentage change in price, holding other demand conditions fixed.

Use the midpoint method for this two-price comparison. It divides each change by the average of its starting and ending values. Using the same midpoint in either direction avoids getting a different responsiveness measure merely because we describe the price change backward.

First calculate the quantity change:

• Average quantity = (300 + 340) ÷ 2 = 320 subscribers.
• Change in quantity = 340 − 300 = 40 subscribers.
• Midpoint quantity change = (40 ÷ 320) × 100 = 12.5%.

Then calculate the price change:

• Average price = ($120 + $110) ÷ 2 = $115.
• Change in price = $110 − $120 = −$10.
• Midpoint price change = (−$10 ÷ $115) × 100 ≈ −8.70%.

Finally, divide the proportional quantity change by the proportional price change. Using the unrounded fractions:

Midpoint elasticity = (40 ÷ 320) ÷ (−10 ÷ 115) = −1.4375.

The negative sign records that price and quantity move in opposite directions. Its absolute value, 1.4375, says that the percentage quantity response is larger than the percentage price change over this interval. Demand is elastic over the interval. If the absolute value were below one, we would call it inelastic; exactly one is unit elastic.

For a price cut, elastic demand means the additional quantity more than offsets the lower price in the revenue calculation. With inelastic demand, revenue falls; with unit-elastic demand, it is unchanged. These revenue comparisons concern the same two prices and use midpoint percentages. For a price increase, the revenue directions reverse.

But elasticity contains no service-cost information. Our calculated elasticity correctly signals higher revenue, while the contribution calculation rejects the cut. If service cost were zero and nothing else changed, that revenue gain would also be a profit gain. At $40 per subscriber, it is not.

This elasticity is calculated from a forecast, not estimated from observed customer choices. Nor does it tell us the response at $90 or $150. Two stipulated points do not establish a constant elasticity, a complete demand curve, or the most profitable price. The manager has compared two proposals, not solved every possible pricing problem.

How many customers would make the cut worthwhile?

At the proposed price, each subscriber contributes $70. To match the original $24,000 contribution, the company would need:

Required subscribers = $24,000 ÷ $70 = 342 6/7.

A subscriber cannot buy six-sevenths of the full-month subscription in this example. At 342 subscribers, contribution is 342 × $70 = $23,940, still below the original result. At 343, it is 343 × $70 = $24,010. Thus 343 is the smallest whole-customer count that improves contribution under our assumptions.

The forecast of 340 misses that threshold by only three subscribers. That makes the quality of the forecast important, not the price cut automatically good or bad in practice. Management should ask whether there is credible reason to expect enough customers and whether any omitted acquisition or service expense would raise the threshold.

Suppose the response were 350 subscribers instead. Contribution would be 350 × $70 = $24,500, a $500 improvement. This is a sensitivity calculation: it shows what follows from a different input. It does not provide evidence that 350 customers will actually subscribe.

A forecast needs evidence

A customer survey, a previous price change, and a limited pricing test can each supply information, but they answer different questions. Saying a service sounds attractive at $110 is not the same as paying for it. A historical sales increase may have coincided with more advertising or a rival’s service failure.

The relevant comparison is what the same business would sell during the same period with and without the price change. Last month’s 300 subscribers need not be that comparison. Seasonal demand could have increased subscriptions even without a discount.

Where feasible and appropriate, randomly assigning comparable prospective customers to different price offers can make their resulting purchases more informative about price. The offers must otherwise be comparable, and customers’ exposure to one offer must not alter the response to the other. Record purchases, service costs, and subsequent retention, not merely clicks or expressions of interest.

Even a well-run short test has boundaries. Offers to new prospects do not by themselves reveal how existing subscribers react to repricing. A brief test may miss later cancellations, costly service requests, or rival responses. Use evidence to improve the comparison while keeping those limits visible.

Competition is part of this forecast, not just an industry label. Ask what rivals could offer, how quickly customers could move, and whether an attractive margin would draw additional suppliers. Entry requires resources and time, so it is neither instantaneous nor irrelevant. Chapter 4 develops strategic responses more explicitly; here they identify assumptions that may not remain fixed.

Service tiers: customers choose, management does not assign

A single price for a single service is not the only possible offer. Consider a separate fictional monthly scheduling service with two packages. Standard costs the customer $60. Priority costs $100 and adds reserved rapid-response support. The provider’s incremental monthly costs are $20 for Standard and $40 for Priority.

This is not a revised forecast for the earlier 300 subscribers. We will examine two individual customers, A and B. Their assumed monthly values appear in Table 3.2. Each can afford either package and can choose one or neither. Values are measured relative to the customer’s best outside alternative, whose net gain we set to zero.

The customer’s gain, or consumer surplus, is value minus price. It belongs to the customer, not the provider’s income statement.

On a small screen, scroll the table sideways to see all columns.

A chooses Standard: $90 − $60 = $30 is better than $105 − $100 = $5. B chooses Priority: $155 − $100 = $55 is better than $100 − $60 = $40. Both prefer the selected package to the outside alternative.

The provider does not need to label A and B at the point of sale. Their own choices sort them across the menu. This is self-selection. The numbers make the reasoning visible to us; they do not imply that a manager knows each customer’s private valuation.

Now raise only the Priority price to $120. B still values Priority at $155, so buying it gives a positive gain of $35. Yet Standard still gives B $40. B downgrades. “The customer values Priority above its price” is therefore not enough to predict a Priority sale.

The upgrade is worth $155 − $100 = $55 to B. Initially it costs $100 − $60 = $40 more than Standard. After the increase, it costs $120 − $60 = $60 more. B rejects the upgrade because its additional value no longer covers its additional price.

The use of different versions to induce customer self-selection is developed in Varian’s “Versioning Information Goods” (1997). The bounded lesson here is that an expensive option competes with the seller’s own cheaper option, not only with outside suppliers. Our service, values, and calculations are separate teaching constructions.

A workable menu must cover its consequences

Under the original menu, the provider receives $60 − $20 = $40 contribution from A’s Standard subscription and $100 − $40 = $60 from B’s Priority subscription. When the higher Priority price induces B to downgrade, contribution from B becomes $40, a $20 reduction. A larger posted price does not help if the customer buys a different package.

The original price gap is $40, while the service-cost gap is $20. The menu therefore does more than pass through added cost: it collects a larger contribution from the customer who values the upgrade more. This illustrates price discrimination, pricing differences that are not explained solely by cost differences. Different posted prices alone are not enough to establish it; better service can genuinely cost more to supply.

These unit contributions do not prove that offering both tiers beats a single-package alternative. That comparison also needs customer counts, the sales each package displaces, and any extra staffing or setup costs. A new cheaper tier can attract additional customers while also drawing existing customers away from a more profitable package.

The promised service distinction must be deliverable. If rapid responses require reserved staff time, that capacity has a cost. If Standard customers routinely receive the same priority service through exceptions, the reason to upgrade weakens. If package rules are confusing or restrictive, customers may prefer another provider or no subscription. Attempts to prevent sharing or circumvention also consume resources.

This brings pricing back to organizational design. Someone must decide which promises sales staff can make, authorize exceptions, and track the resulting service burden. A revenue target might reward the original $110 price cut even though it lowers contribution. A subscription-count target would be still less informative about that tradeoff. Later chapters examine these evaluation and compensation choices directly.

What should the manager recommend?

Keep the $120 price rather than adopt this particular $110 proposal if the stated forecasts and costs are correct. The proposed cut raises revenue but reduces monthly contribution by $200. That is a definite recommendation within a limited comparison, not a claim that $120 is the best possible price.

The next useful question is whether the customer forecast or service-cost assumption should change. A credible forecast of at least 343 full-month subscribers would reverse the comparison, provided no other costs or benefits change. A tiered offer is another possible proposal, but it needs its own customer-choice and cost analysis.

Study map

Identify the customer’s alternatives and the period of adjustment. Separate one seller’s demand from the whole market. Use midpoint elasticity to describe responsiveness and revenue effects, then include relevant costs. Check the forecast against the contribution threshold. For a menu, compare the customer’s gain from every available option. End with a recommendation, its assumptions, and the evidence most likely to change it.

Practice and discussion

These are public practice questions. Unless stated otherwise, retain the original assumptions and change only the input specified.

1. Change the response. At $110, suppose the first provider expects 350 subscribers. Calculate revenue, variable service cost, contribution, and the change from keeping $120. What is the smallest whole-customer count that improves contribution? Explain why finding that threshold does not validate the forecast.

2. Change the cost. Return to the original 300/340 subscriber comparison, but let service cost be $30 per subscriber at both prices. Calculate contribution under each price and recommend one of the two. Does the midpoint elasticity change? Explain why identical customer responsiveness can accompany a different pricing recommendation.

3. Question the evidence. After the company cuts its price, subscriptions rise from 300 to 340. A rival also experiences a service outage that month. What can and cannot be concluded from the increase? Describe one comparison that would help separate the price effect from the rival’s problem, and one limitation that would remain.

4. Check the menu. In the separate tier example, raise only Priority’s price from $100 to $120. Calculate both customers’ gains from both packages and identify their choices. Calculate the provider’s contribution from B before and after the change. Why does B’s $155 Priority valuation not guarantee a Priority sale?

5. Write a recommendation. In about 150 words, assess the original subscription price cut for someone proposing to evaluate it by revenue alone. Include the midpoint elasticity, the contribution change, and one assumption requiring evidence. Explain what each number establishes without claiming to have found the best possible price.

Check your arithmetic

For question 1, revenue is $38,500 and variable service cost is $14,000, leaving $24,500 contribution: a $500 gain. The minimum improving count is 343. For question 2, 300 × $90 = $27,000 and 340 × $80 = $27,200; the cut gains $200, while elasticity remains −1.4375. For question 4, A’s Standard/Priority gains are $30/−$15 and B’s are $40/$35. Both choose Standard; contribution from B falls from $60 to $40.

Source note

The subscription cases, tables, values, and exercises are original hypothetical constructions. Varian supplies a conceptual attribution, not evidence for these numerical inputs or their profitability. Neither example establishes an actual firm’s demand or an optimal pricing policy.