Business Models, Unit Economics and Growth Metrics

Introduction

Digital commerce has made it possible for firms to reach customers directly, test propositions rapidly, and scale distribution through data and technology. Yet revenue growth alone does not establish that an e-commerce or direct-to-consumer (D2C) business is economically viable. A firm may report rapidly increasing orders while losing money on every customer, or it may show positive first-order margins while its cash is trapped in inventory and customer acquisition.

This chapter develops an integrated framework for evaluating business models and growth. It connects revenue design with customer acquisition cost (CAC), customer lifetime value (CLV), contribution margin, payback period, cohort economics, retention, churn, average order value (AOV), and repeat purchase rate. The central managerial principle is that growth is valuable only when incremental customers generate sufficient contribution to recover acquisition and operating investments within an acceptable period.

Learning Objectives

By the end of this chapter, students should be able to:

1. Distinguish major e-commerce and D2C revenue models.
2. Calculate and interpret AOV, CAC, contribution margin, CLV, payback period, retention, churn, and repeat purchase rate.
3. Construct a cohort-level view of customer economics.
4. Evaluate growth quality rather than relying on revenue or vanity metrics.
5. Diagnose the effects of pricing, promotion, retention, channel mix, and fulfillment on profitability.
6. Recommend managerial actions using appropriately qualified financial and behavioral evidence.

1. Business Models in E-Commerce and D2C

A business model explains how an organization creates value for customers, delivers that value, and captures economic value for itself. In digital commerce, the model includes the customer proposition, channel, revenue mechanism, cost structure, data practices, and relationships with intermediaries.

1.1 Principal revenue models

These categories can be combined. For example, a D2C skincare firm may sell products, offer an auto-replenishment subscription, and earn revenue from a premium consultation service. The correct model depends on purchase frequency, product durability, customer involvement, switching costs, and the firm's operational capabilities.

1.2 Business-model economics

A model should be evaluated across four linked questions:

• Value proposition: Why does the customer choose this offer?
• Monetization: Who pays, how much, and when?
• Cost-to-serve: What variable costs arise from acquisition, payment, fulfillment, returns, service, and discounts?
• Scaling logic: Do higher volumes improve economics, or do they increase complexity and cost?

A marketplace, for instance, may have attractive gross margins after reaching scale, but it must invest in trust and safety, seller acquisition, dispute resolution, and technology. A subscription can improve forecasting, but a misleading renewal practice can increase short-term revenue while damaging long-term retention and regulatory compliance.

2. Unit Economics: The Customer as an Economic Unit

Unit economics examines revenue and costs associated with one economically meaningful unit, commonly an order, customer, subscription, or cohort. The unit must be defined consistently. A business that reports product gross margin but excludes shipping subsidies, payment fees, returns, and customer support may overstate its true contribution.

2.1 Revenue and average order value

For a period:

Net order revenue should specify whether it is measured before or after discounts, refunds, taxes, and shipping revenue. For managerial analysis, revenue is usually reported net of discounts, refunds, and returns, while sales taxes collected on behalf of governments are excluded.

If customers place multiple orders, revenue per customer is different from AOV:

The multiplication is valid only when the measures refer to the same period and customer population.

2.2 Contribution margin

Contribution margin measures the amount available to cover fixed costs, acquisition investment, and profit after variable costs.

Variable costs may include product cost, payment processing, pick-and-pack, outbound shipping subsidy, expected return cost, customer service, marketplace fees, and variable fulfillment labor. Fixed salaries, platform development, and headquarters rent are normally excluded from contribution margin unless they vary directly with the unit.

The formula is useful for short- to medium-term decisions when variable costs can be identified reliably. It does not by itself establish accounting profit because fixed costs remain to be covered.

Worked Example 1: Order contribution

A retailer records net revenue of $80 per order. Product cost is $28, payment cost is $2.40, fulfillment is $6, shipping subsidy is $4, and expected return cost is $3.

Variable cost per order = $28 + $2.40 + $6 + $4 + $3 = $43.40.

Contribution margin per order = $80 − $43.40 = $36.60.

Contribution margin percentage = $36.60 ÷ $80 × 100 = 45.75%.

If a promotional discount reduces net revenue to $68 while variable costs remain $43.40, contribution falls to $24.60 and the margin percentage falls to approximately 36.18%. The discount therefore changes both revenue and the rate at which orders contribute to fixed costs.

3. Customer Acquisition Cost and Payback

3.1 CAC

Customer acquisition cost estimates the cost of acquiring a new customer during a defined period.

The numerator may include paid media, agency fees, affiliate commissions, promotional incentives, creative production allocated to acquisition, and relevant sales labor. The analyst must state whether discounts, free samples, and onboarding costs are included. The denominator should contain genuinely new customers rather than all orders or all website visitors.

CAC is often calculated by channel. A blended CAC is:

Blended CAC is useful for portfolio decisions but can conceal unprofitable channels. Attribution is difficult because customers may encounter several touchpoints, organic demand may be incorrectly credited to paid media, and retargeting may receive excessive credit.

3.2 Payback period

Payback period is the time required for cumulative contribution from a customer to recover CAC.

When monthly contribution is approximately constant:

If contribution varies by month, the correct procedure is cumulative summation. Let CMₜ be contribution from a customer in period t:

Payback occurs in the first period T for which cumulative contribution is at least CAC. The constant-contribution formula is only an approximation and should not be used for strongly seasonal, subscription, or repeat-purchase businesses without qualification.

Payback is also a cash-risk measure. A business with positive lifetime economics may still fail if it pays for acquisition and inventory long before collecting enough customer contribution.

4. Customer Lifetime Value

CLV estimates the future contribution attributable to a customer relationship. It is not a single universally correct number; it depends on the time horizon, discounting, margin definition, retention model, and treatment of future costs.

4.1 Basic contribution-based CLV

For a customer observed over periods 1 through T:

Here, r is the discount rate per period and T is the chosen time horizon. The formula assumes that expected revenue and variable costs can be estimated for each period. It excludes CAC because CAC is compared against CLV rather than normally embedded within it.

A simplified steady-state model for a customer with monthly contribution m and monthly retention probability s is:

This expression assumes a geometric retention process, constant contribution, no finite horizon, no discounting, and no acquisition or reactivation effects. It is a model, not a fact about customer behavior. As retention approaches 1, the estimate becomes highly sensitive to small measurement errors.

If discounting is included and retention is constant:

The denominator requires r and s to be measured for the same period. These formulas are inappropriate when contribution changes materially with tenure, churn is non-constant, or customers have seasonal purchasing patterns. In those cases, a cohort model is preferable.

4.2 CLV:CAC interpretation

A common diagnostic is:

A ratio above 1 indicates that estimated lifetime contribution exceeds acquisition cost, but it does not guarantee liquidity, scalability, or satisfactory return on capital. A high ratio based on optimistic retention assumptions may be less credible than a lower ratio supported by mature cohorts. Managers should examine the absolute payback period, contribution definition, cohort maturity, and channel-specific variation.

5. Retention, Churn, and Repeat Purchase

Retention is the proportion of customers who remain active or purchase again during a defined interval. Churn is the proportion that becomes inactive or cancels. The definitions must identify the population, activity event, observation window, and censoring treatment.

For a subscription cohort:

If there are no acquisitions or reactivations within the interval, churn rate is approximately 1 minus retention rate. With new customers entering during the period, that simple relationship does not hold unless the denominator and population are controlled.

For transactional commerce, repeat purchase rate can be defined as:

The window must be long enough to observe the normal repurchase cycle. A 30-day window may understate repeat purchase for furniture but be suitable for consumables. Repeat purchase rate should not be confused with purchase frequency, which is the average number of orders per active customer.

Retention is not automatically evidence of satisfaction. Customers may remain because of contractual lock-in, lack of alternatives, or cancellation friction. Conversely, a customer may be highly satisfied but purchase infrequently because the product is durable. Behavioral measures should therefore be interpreted alongside complaints, returns, survey data, and customer support records.

6. Cohort Economics

A cohort is a group sharing a meaningful starting condition, such as first-purchase month, acquisition channel, geography, product, or campaign. Cohort analysis separates customer behavior from changes in the volume or composition of acquisition.

A useful cohort table records:

The figures are illustrative. A cohort should be compared with cohorts of similar maturity; a three-month-old cohort cannot fairly be compared with a twelve-month-old cohort using cumulative CLV. Cohort curves may reveal declining acquisition quality, improving product-market fit, seasonal demand, or the effects of a pricing change.

A mature cohort analysis commonly tracks cumulative contribution per acquired customer:

The cohort reaches payback when this value equals or exceeds cohort CAC. Analysts should use realized data for mature periods and clearly label forecasts for immature periods.

7. A Growth-Metric System

No single metric captures growth quality. A useful metric system connects acquisition, conversion, transaction value, retention, margin, and cash.

Acquisition and conversion

Traffic, qualified sessions, conversion rate, new customers, CAC, and channel incrementality indicate whether demand is being generated efficiently. Conversion rate is:

The denominator must be stable. Orders divided by users, sessions, and unique visitors produce different metrics.

Transaction economics

AOV, units per order, discount rate, return rate, gross margin, contribution margin, and shipping cost per order explain what each transaction contributes. Increasing AOV is not necessarily beneficial if it results from costly discounts or higher returns.

Relationship economics

Repeat purchase rate, order frequency, retention, churn, reactivation, and CLV measure whether the firm is creating durable customer value. These metrics should be segmented by product, customer type, channel, and acquisition campaign.

Financial and operational control

Contribution after marketing, payback period, inventory turnover, cash conversion, refund timing, and fulfillment reliability connect marketing metrics to business viability. A growth team should be accountable not only for orders but also for contribution and customer quality.

A practical North Star metric should represent customer value while remaining sensitive to economic quality. Examples include retained customers generating positive contribution within 90 days or cumulative contribution from newly acquired cohorts. Revenue alone is often too permissive; profit alone may be too delayed for experimentation.

8. Worked Integrated Example

Assume a D2C supplement brand acquires 2,000 new customers in April. Acquisition spending is $60,000, so:

CAC = $60,000 ÷ 2,000 = $30.

Each first order produces $54 of net revenue. Variable costs are product cost $20, payment $1.62, fulfillment $5, shipping subsidy $4, and expected returns $2.38. Total variable cost is $33, giving:

Contribution per first order = $54 − $33 = $21.

The first-order contribution does not recover the $30 CAC. Suppose 40% of customers place a second order within the next period. The second order has the same $21 contribution per ordering customer, so expected second-order contribution per acquired customer is:

0.40 × $21 = $8.40.

Cumulative expected contribution after two orders is $21 + $8.40 = $29.40, still below CAC. If 25% of the original cohort places a third order, expected third-order contribution is 0.25 × $21 = $5.25. Cumulative contribution becomes $34.65, exceeding CAC. The approximate payback point is the third-order period, subject to timing and actual cohort realization.

The example demonstrates why first-order profitability and customer profitability are different. It also shows why repeat purchase assumptions should be tested against observed cohort behavior rather than treated as permanent constants.

9. Illustrative Case: Verdant Home Direct

Background

Verdant Home Direct, a fictional D2C retailer of refillable household products, doubled paid-media spending over six months. Revenue increased by 82%, but cash losses widened. Management initially attributed the problem to fulfillment costs.

Problem

The firm reported a blended CAC of $24, AOV of $62, and a 52% gross margin. However, the gross-margin calculation excluded free shipping, payment fees, returns, and first-order discounts. Paid social appeared to be the largest growth channel.

Analysis

A revised contribution analysis produced the following approximate figures:

Paid social generated substantial order volume but attracted customers with low repeat purchase and high discount usage. Search customers had stronger economics, while email and organic customers were profitable but limited in volume.

Approach and solution

Management adopted five actions:

1. Reported contribution after shipping, payment, returns, and discounts.
2. Evaluated channels by mature cohorts rather than last-click revenue.
3. Tested product bundles that increased AOV without increasing return rates.
4. Introduced replenishment reminders based on expected consumption intervals.
5. Set a payback limit requiring new campaigns to recover CAC within six months unless strategically justified.

The company also reviewed consent management and promotional disclosures. Personalization was limited to data for which customers had an appropriate notice and lawful basis.

Outcome

In the illustrative scenario, paid-social spending was reduced and reallocated to high-intent search, referral, and lifecycle programs. Revenue growth slowed temporarily, but contribution after marketing improved, cash payback shortened, and the next mature cohort showed higher repeat purchase. The result illustrates a central distinction: optimizing for volume can conflict with optimizing for economically valuable growth.

Learning points

• Gross margin is not the same as contribution margin.
• Blended averages can conceal channel-level failure.
• Cohort maturity matters when estimating CLV.
• Retention interventions must improve customer value, not merely delay cancellation.
• Marketing, finance, merchandising, data, and operations must share metric definitions.

10. Limitations, Governance, and Ethical Issues

Measurement limitations

CAC is vulnerable to attribution error, organic-demand contamination, and inconsistent cost allocation. CLV is vulnerable to extrapolation from immature cohorts and survivorship bias. Retention may be inflated by involuntary churn being omitted or by defining activity too broadly. AOV can rise because of discounts, bundles, or a shift toward expensive products without improving contribution.

Managers should therefore publish metric definitions, time windows, inclusion rules, cohort maturity, and confidence or sensitivity ranges where appropriate. Scenario analysis is often more honest than a single precise CLV estimate. For example, a firm might report CLV under low, base, and high retention assumptions.

Experimentation and causal interpretation

An observed association between email exposure and higher CLV does not prove that email caused the difference; more engaged customers may be more likely to open email. Randomized holdout tests, where feasible, provide stronger evidence. Tests must account for statistical power, seasonality, interference between customers, and long-term effects. Short-term conversion gains may reduce margin or increase future churn.

Privacy and fairness

Customer-level data used for segmentation and CLV prediction should be collected, stored, and processed according to applicable privacy law and organizational governance. Firms should minimize unnecessary data, control access, provide appropriate transparency, and respect deletion or preference requests. Automated offers can create unfair treatment if pricing or eligibility models use protected or proxy variables. Explainability, monitoring, human review, and appeal processes are important where decisions materially affect customers.

Contemporary developments

AI can support demand forecasting, product recommendations, churn-risk scoring, creative testing, and customer-service triage. However, predictive accuracy does not guarantee incremental profit. A churn model may identify likely cancellers without showing which intervention will retain them. AI systems should therefore be evaluated using incremental contribution, calibration, fairness, privacy, and operational feasibility—not merely predictive accuracy.

11. Best-Practice Managerial Framework

A disciplined growth review can follow this sequence:

1. Define the economic unit and all metric denominators.
2. Reconcile orders, revenue, refunds, discounts, and accounting records.
3. Calculate contribution using fully specified variable costs.
4. Separate new, retained, reactivated, and existing customers.
5. Build cohorts by acquisition period and channel.
6. Estimate CAC, retention, repeat purchase, CLV, and payback using observed and forecast components.
7. Stress-test assumptions for margin, churn, returns, and acquisition cost.
8. Use experiments or credible quasi-experimental designs to evaluate interventions.
9. Monitor cash timing, inventory, service capacity, privacy, and customer welfare.
10. Scale only when incremental cohorts meet strategic and financial thresholds.

Chapter Summary

E-commerce and D2C growth must be evaluated as an economic system rather than as a collection of isolated marketing metrics. Business models determine how value is monetized, while unit economics identifies the revenue and variable cost associated with orders and customers. AOV describes transaction value; contribution margin describes the value remaining after variable costs; CAC measures acquisition investment; and payback indicates how quickly that investment is recovered.

CLV estimates future customer contribution but depends strongly on retention, margin, time horizon, discounting, and cohort maturity. Retention, churn, repeat purchase rate, and purchase frequency must be defined with precise populations and time windows. Cohort analysis is essential for distinguishing genuine improvement from changes in acquisition volume, seasonality, or customer mix. Managers should combine these measures with cash, inventory, fulfillment, privacy, and ethical considerations. The most valuable growth is incremental, durable, measurable, and contribution-positive.

Key Terms

Business model: The logic by which a firm creates, delivers, and captures value.

Direct-to-consumer (D2C): Selling directly to end customers without relying primarily on traditional intermediaries.

Unit economics: Revenue and cost analysis for a defined economic unit such as an order or customer.

Average order value (AOV): Net order revenue divided by the number of orders.

Variable cost: A cost that changes with the volume or characteristics of transactions.

Contribution margin: Revenue remaining after specified variable costs.

Customer acquisition cost (CAC): Acquisition and conversion cost divided by new customers acquired.

Blended CAC: Total acquisition cost divided by total new customers across channels.

Customer lifetime value (CLV): Estimated discounted or undiscounted future contribution from a customer relationship.

Payback period: Time required for cumulative contribution to recover acquisition cost.

Retention rate: Proportion of a defined customer population remaining active at a specified time.

Churn rate: Proportion of customers lost during a defined period.

Repeat purchase rate: Proportion of active customers making at least two purchases in a specified window.

Cohort: Customers sharing a defined starting condition or time period.

Contribution after marketing: Contribution remaining after acquisition or marketing expenditure.

Attribution: Assignment of credit for a conversion among marketing touchpoints.

Incrementality: The additional outcome caused by an intervention compared with a valid counterfactual.

Payback threshold: Maximum acceptable time or condition for recovering an investment.

Metric denominator: The defined population or activity count against which a rate is calculated.

Review Questions

Short Answer Questions

1. What is the difference between gross margin and contribution margin?
2. Why must CAC be calculated using new customers rather than total orders?
3. Define repeat purchase rate and explain why its observation window matters.
4. What is the purpose of cohort analysis?
5. Why does a positive CLV:CAC ratio not necessarily guarantee business success?

Long Answer Questions

1. Compare product sales, subscription, marketplace, and advertising-supported revenue models, including their principal risks.
2. Explain how AOV, contribution margin, CAC, CLV, and payback period should be used together in a growth decision.
3. Discuss the assumptions and limitations of the steady-state CLV formula.
4. Explain how attribution error and cohort immaturity can distort growth reporting.
5. Evaluate the ethical and governance issues involved in AI-based churn prediction and personalized offers.

Analytical / Application Questions

1. A retailer has CAC of $40 and first-order contribution of $28. Forty percent of customers place a second order with contribution of $25, and 20% place a third order with contribution of $25. Calculate expected cumulative contribution after three orders and determine whether the cohort has reached payback.
2. Design a cohort dashboard for a subscription D2C business. Specify at least eight measures, their definitions, and the decisions each measure would inform.
3. A paid-social channel has lower CAC than search but substantially lower retention and contribution. Explain how you would test whether shifting budget toward search creates incremental profit rather than merely changing attribution.

Multiple Choice Questions with Answers

Question 1

A retailer has net revenue of $100 per order and variable cost of $62. What is the contribution margin percentage?

A. 38%

B. 62%

C. 161%

D. $38

Correct Answer: A. 38%

Question 2

A firm spends $45,000 to acquire 1,500 new customers. What is CAC?

A. $15

B. $30

C. $45

D. $67.50

Correct Answer: B. $30

Question 3

Which condition is required for the simplified steady-state CLV formula to be appropriate?

A. Retention changes arbitrarily each month

B. Contribution and retention are approximately constant by period

C. CAC is included in every period's contribution

D. All customers make exactly one purchase

Correct Answer: B. Contribution and retention are approximately constant by period

Question 4

A cohort begins with 2,000 customers and has 1,500 active customers at month three. What is month-three retention?

A. 25%

B. 50%

C. 75%

D. 133.3%

Correct Answer: C. 75%

Question 5

Which metric specifically measures the time needed to recover CAC from cumulative customer contribution?

A. AOV

B. Churn rate

C. Payback period

D. Conversion rate

Correct Answer: C. Payback period

Question 6

A store generates $24,000 in net order revenue from 400 orders. What is AOV?

A. $40

B. $60

C. $96

D. $240

Correct Answer: B. $60

Question 7

Which statement best describes cohort analysis?

A. It combines all customers into one average regardless of acquisition date

B. It separates customers sharing a defined starting condition for comparison over time

C. It measures only website traffic

D. It replaces the need to calculate contribution margin

Correct Answer: B. It separates customers sharing a defined starting condition for comparison over time

Question 8

A subscription business starts a month with 1,000 customers and loses 80 during the month, with no new acquisitions. What is the churn rate?

A. 8%

B. 9.2%

C. 80%

D. 92%

Correct Answer: A. 8%

Question 9

Why can a high CLV:CAC ratio be misleading?

A. CLV never includes future contribution

B. CAC is always a fixed cost

C. CLV may rely on optimistic retention assumptions or immature cohorts

D. The ratio cannot be calculated for digital businesses

Correct Answer: C. CLV may rely on optimistic retention assumptions or immature cohorts

Question 10

Which method generally provides stronger evidence that a marketing intervention caused an outcome?

A. Last-click attribution alone

B. Comparing customers who self-select into exposure

C. A randomized holdout test

D. Increasing the reporting period without a control group

Correct Answer: C. A randomized holdout test