Mathematical Foundations for Analytics

Introduction

Business analytics converts observations into evidence for decisions. Although software performs much of the computation, analysts must understand the mathematical structures behind the output. Arithmetic supports reliable measurement; ratios and percentages support comparison; logarithms describe multiplicative change; functions represent business relationships; equations and inequalities express constraints; matrices organise data; and calculus explains marginal change and optimisation.

These topics are not isolated techniques. A revenue model may be a function, a break-even point an equation, a capacity limit an inequality, a customer dataset a matrix, and a pricing decision an optimisation problem. Sound mathematical reasoning also requires attention to units, assumptions, rounding and interpretation.

Learning Objectives

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

• perform and interpret arithmetic operations with appropriate units and rounding;
• calculate ratios, rates, percentage change and compound growth;
• use powers and logarithms in growth, finance and data contexts;
• construct and interpret linear, quadratic and exponential functions;
• solve equations, simultaneous equations and inequalities;
• represent business data and transformations using matrices;
• explain derivatives, marginal analysis and simple integration;
• apply these ideas to cost, revenue, profit and business-growth problems;
• evaluate assumptions, limitations and ethical risks in quantitative models.

1. Arithmetic, Units and Quantitative Reasoning

Arithmetic includes addition, subtraction, multiplication, division, powers and roots. In analytics, the operation must match the meaning of the variables. Addition is meaningful only when quantities have compatible units. For example, 240 units plus 60 units equals 300 units, whereas adding 240 units to $60 has no direct interpretation.

The order of operations is conventionally powers and roots, multiplication and division, then addition and subtraction. Parentheses override this order. A calculator result should be checked by estimating its magnitude and by checking units.

Rates and unit conversion

A rate compares a quantity with a unit of time or another denominator:

For example, if sales are $48,000 over 12 months, average monthly sales are $4,000 per month. A rate is not necessarily constant; an average can hide seasonal variation.

Dimensional analysis is a useful error check. If a machine produces 900 items in 6 hours, then:

Multiplying this rate by 2 hours gives 300 items. Units cancel or combine in the same way as numerical factors.

Accuracy, rounding and significant figures

Retain sufficient precision during calculations and round the final result. Premature rounding can materially affect compound calculations. Monetary results are usually reported to two decimal places, while percentages may be reported to one or two decimal places depending on the decision context. A reported value should not imply greater accuracy than the measurement supports.

2. Ratios, Percentages and Growth

A ratio compares two quantities. If a firm has 240 online orders and 160 store orders, the online-to-store ratio is:

Ratios are often normalised by division. The online share of total orders is 240 ÷ 400 = 0.60, or 60%.

A percentage is a ratio per hundred:

If 36 of 120 invoices contain errors, the error percentage is 30%.

Percentage change

For an initial value x₀ and a final value x₁, percentage change is:

This formula assumes x₀ ≠ 0 and uses the initial value as the reference base. If monthly cost increases from $80,000 to $92,000, the increase is ($12,000 ÷ $80,000) × 100% = 15%.

A decrease of 20% followed by an increase of 20% does not return to the original value. Starting at 100, the sequence is 100 × 0.80 × 1.20 = 96. Percentage changes compound because each change applies to a new base.

Compound growth

If a quantity grows at a constant rate r per period for n periods, its future value is:

Here V₀ is the initial value, r is expressed as a decimal, and n is the number of periods. The model assumes the rate is constant and growth is reinvested or compounded each period.

A $50,000 customer base growing by 8% annually for three years becomes:

The compound annual growth rate (CAGR), when initial and final values and the number of periods are known, is:

CAGR summarises a changing path as a constant equivalent rate. It does not show volatility or intermediate performance.

Margins and percentage points

A margin is often a percentage of revenue. If revenue is $200,000 and profit is $30,000, profit margin is 15%. A rise from 20% to 24% is an increase of 4 percentage points, but a relative increase of 20% because 4 ÷ 20 = 0.20. These expressions should not be confused.

3. Powers and Logarithms

A power represents repeated multiplication. In aⁿ, a is the base and n is the exponent. The rules include:

A logarithm is the inverse of exponentiation:

The conditions are b > 0, b ≠ 1 and x > 0. Common logarithms use base 10; natural logarithms use base e ≈ 2.71828 and are written ln(x).

Important laws are:

The change-of-base formula is:

Business applications

Logarithms are useful when solving for time or growth rate. If sales follow V_t = 20,000(1.05)^t, the time to reach $30,000 satisfies:

Thus the threshold is reached during the ninth period if observations occur only at whole periods. Logarithmic transformations also reduce right-skewed scales and convert exponential relationships into linear ones, but transformed results must be interpreted on the original business scale.

4. Functions and Models

A function assigns each permitted input exactly one output. In business models, the input may be quantity, price, time or advertising expenditure, while the output may be cost, demand, revenue or profit. Function notation y = f(x) identifies x as the independent variable and y as the dependent variable.

The domain is the set of permissible inputs; the range is the resulting set of outputs. A model can be mathematically valid but commercially invalid outside its realistic domain.

Linear functions

A linear function has the form:

a is the intercept and b is the slope. The slope is the change in output per one-unit increase in input. If fixed cost is $12,000 and variable cost is $18 per unit, total cost for q units is:

The units of the intercept are dollars and the units of the slope are dollars per unit.

Revenue, cost and profit

If price per unit is p and quantity sold is q, revenue is:

If price is constant, this is R(q) = pq. Profit is revenue minus total cost:

For the cost function above and a price of $50:

Quadratic and exponential functions

A quadratic function has the form:

It can model diminishing returns, capacity effects or revenue under a price-demand relationship. An exponential model has the form:

where a is the initial level and b is the multiplicative factor per unit of x. It describes constant proportional growth or decline, not constant absolute change.

A model is an approximation. Analysts should inspect residuals, validate assumptions and avoid extrapolating far beyond observed data.

5. Equations, Simultaneous Equations and Inequalities

An equation states that two expressions are equal. Solving it means finding values that satisfy the equality. Equivalent operations must be applied to both sides, and denominators may not be zero.

Break-even analysis

Break-even occurs when profit is zero, or revenue equals cost. With fixed cost F, price p and variable cost per unit v,

Rearranging gives:

The denominator is contribution per unit. If p ≤ v, each additional unit fails to cover variable cost and there is no positive break-even quantity under this model.

For F = 12,000, p = 50 and v = 18,

At 500 units, profit is 32(500) − 12,000 = $4,000.

Simultaneous equations

Two unknowns can be found from two independent equations. Suppose a firm sells basic and premium subscriptions. Let b and p be quantities:

Substituting b = 400 − p gives 4,000 − 10p + 25p = 7,000, so p = 200 and b = 200. A solution requires consistent equations and sufficient independent information.

Inequalities and constraints

Inequalities express limits such as budget, capacity or minimum service levels. Multiplying or dividing an inequality by a negative number reverses its direction. If capacity is 2,000 units and demand is at least 1,200, the feasible condition is 1,200 ≤ q ≤ 2,000. Business optimisation often seeks the best feasible value rather than an unrestricted mathematical maximum.

6. Matrices for Analytics

A matrix is a rectangular arrangement of numbers. A matrix with m rows and n columns has dimension m × n. A data matrix may contain observations in rows and variables in columns, although the convention should always be documented.

For example, monthly observations of units and advertising expenditure can be represented as:

The first column might be units and the second advertising dollars. Mixing units without labels is a governance risk.

Matrices can be added only when they have the same dimensions. Scalar multiplication multiplies every element by the scalar. Matrix multiplication is defined when the number of columns in the first matrix equals the number of rows in the second. If A is m × n and B is n × p, then AB is m × p, with entries:

Matrix multiplication is generally not commutative: AB ≠ BA. This matters when transformations are applied in sequence.

Matrix representation of equations

The subscription example can be written:

In applied analytics, matrices support spreadsheet models, linear regression, portfolio calculations, recommendation systems and machine-learning algorithms. A matrix does not automatically establish causation; it is a representation whose meaning depends on variable definitions and data quality.

7. Introductory Calculus for Business Decisions

Calculus studies change and accumulation. It is useful when a business quantity varies continuously or when a manager wants to assess marginal effects.

Derivatives and marginal analysis

The derivative of f(x) at x is the instantaneous rate of change:

This definition requires a meaningful limiting process and a function that is differentiable at the point. Common rules include:

For total cost C(q) = 12,000 + 18q, the marginal cost is C′(q) = 18 dollars per additional unit. For C(q) = 12,000 + 0.02q² + 10q, marginal cost is C′(q) = 0.04q + 10. It is a local approximation, not necessarily the exact cost of a large discrete increase.

A stationary point occurs where f′(x) = 0. The second derivative helps classify it: if f′′(x) > 0, the point is locally a minimum; if f′′(x) < 0, it is locally a maximum, subject to domain and endpoint checks.

Optimising profit

Suppose price depends on quantity according to p(q) = 100 − 0.2q. Revenue is:

If cost is C(q) = 1,000 + 20q, profit is:

Then:

Setting the derivative to zero gives q = 200. Since π′′(q) = −0.4 < 0, this is a local maximum. The model is meaningful only for non-negative quantities and prices; at 200 units, price is $60 and profit is 80(200) − 0.2(200)² − 1,000 = $7,000.

Integration and accumulation

A definite integral accumulates a rate over an interval:

If C′(q) is marginal cost, integrating it from q₁ to q₂ gives the variable cost change, provided the function and units are correctly specified. The fundamental theorem connects integration and differentiation:

In practical analytics, discrete sums may be more appropriate when transactions occur in whole units or at irregular times.

Applied Case: Launching a Subscription Service

The following is an illustrative case, with invented figures.

Background and problem

A small software firm is considering a subscription service. Monthly fixed operating cost is $12,000, variable service cost is $18 per subscriber, and the proposed monthly price is $50. Management wants to know the break-even volume, the effect of growth, and whether capacity is sufficient.

Analysis

The contribution per subscriber is $50 − $18 = $32. Break-even volume is:

The service expects 500 subscribers in month one. Monthly operating profit is:

Suppose subscribers grow by 6% monthly from an initial 500. After six months:

If technical capacity is 650 subscribers, expected demand exceeds capacity in month six. The inequality for feasible operation is V_t ≤ 650. Solving 500(1.06)^t ≤ 650 using logarithms gives approximately t ≤ 4.45, so capacity is exceeded during the fifth month if monthly observations are whole periods.

Approach and outcome

Management could increase server capacity, moderate acquisition spending, or introduce a tiered plan. The model indicates a profitable launch at 500 subscribers but also identifies a capacity constraint. A sensitivity analysis should vary price, churn, variable cost and growth rate rather than treating estimates as facts.

Learning points

The case combines unit rates, contribution, equations, compound growth, logarithms and inequalities. It also shows why a mathematically correct result is not automatically a good decision: the assumptions, timing, capacity and customer response require validation.

Limitations, Governance and Professional Practice

Mathematical models simplify reality. Common limitations include constant-growth assumptions, linear relationships outside their valid range, omitted variables, measurement error and dependence among observations. Analysts should document the source of each input, units, time period, rounding rule and intended domain.

Ethically, analysts must not select a formula merely because it supports a preferred decision. Growth projections can affect staffing, investment and customer pricing, so uncertainty should be communicated. If models use customer or employee data, privacy, access control, retention and applicable data-protection requirements must be considered. Automated tools and AI systems can calculate or generate formulas, but they may propagate unit errors, hallucinated assumptions or biased data. Human review, reproducible calculations and independent checks remain essential.

Best practice is to: define variables before calculating; use dimensional analysis; maintain full precision until final reporting; test boundary cases; compare model output with observed data; perform sensitivity analysis; and present results with assumptions and uncertainty.

Chapter Summary

Arithmetic provides the operational language of quantitative work, while units and rounding protect interpretation. Ratios and percentages support comparison, but percentage changes must use an explicit base and compound changes must be multiplied rather than simply added. Powers and logarithms describe and solve multiplicative growth.

Functions represent relationships among business variables. Linear models are useful for fixed and variable costs, while quadratic and exponential models capture curvature and proportional growth. Equations determine unknown quantities such as break-even volume; inequalities represent feasible business constraints. Matrices organise data and systems of equations, but their dimensions and variable meanings must be checked. Derivatives measure marginal change and support optimisation, while integrals represent accumulation. Every model remains conditional on assumptions, data quality, domain and ethical governance.

Key Terms

Ratio: Comparison of two quantities.

Rate: Quantity expressed per unit of another quantity, often time.

Percentage change: Change relative to an initial reference value.

Compound growth: Growth in which each period’s change applies to the updated value.

CAGR: Constant annual equivalent growth rate over several periods.

Logarithm: Inverse operation of exponentiation.

Function: Rule assigning each permitted input one output.

Domain: Set of permissible input values.

Range: Set of output values produced by a function.

Slope: Change in output per unit change in input.

Fixed cost: Cost that does not vary with activity within a relevant range.

Variable cost: Cost that changes with activity.

Contribution: Price minus variable cost per unit.

Break-even point: Activity level at which revenue equals total cost.

Inequality: Mathematical statement expressing an ordered constraint.

Matrix: Rectangular array of numbers or variables.

Derivative: Instantaneous rate of change.

Marginal cost: Change in cost associated with an additional unit, often represented by a derivative.

Integral: Mathematical accumulation over an interval.

Sensitivity analysis: Examination of how output changes when assumptions vary.

Review Questions

Short Answer Questions

1. Why must units be checked when adding or multiplying business quantities?

1. Distinguish percentage points from percentage change.

1. State the conditions required for log_b(x) to be defined as a real logarithm.

1. What does the slope of a linear cost function represent?

1. Explain why matrix multiplication is not generally commutative.

Long Answer Questions

1. Explain compound growth and derive the CAGR formula. Include a business example.

1. Describe how revenue, cost and profit functions are constructed and interpreted.

1. Explain break-even analysis, including its assumptions and limitations.

1. Discuss how derivatives support marginal analysis and business optimisation.

1. Explain how matrices can represent a system of simultaneous business equations.

Analytical / Application Questions

1. A product has fixed cost $25,000, price $80 and variable cost $30. Calculate break-even quantity and profit at 1,000 units. Interpret the results.

1. A customer base of 2,400 grows by 7% per quarter. Estimate its size after eight quarters and determine whether it exceeds a capacity of 4,000 customers.

1. Demand is q = 1,000 − 5p, where p is price. Derive revenue as a function of price and use differentiation to identify the price that maximises revenue, subject to a non-negative demand constraint.

Multiple Choice Questions with Answers

Question 1

A cost rises from $80,000 to $92,000. What is the percentage increase?

A. 12%

B. 15%

C. 18%

D. 92%

Correct Answer: B. 15%

Question 2

A value of $50,000 grows by 8% annually for three years. Which expression gives the future value?

A. 50,000 + 0.08 × 3

B. 50,000 × 0.08^3

C. 50,000 × 1.08^3

D. 50,000 × 1.24

Correct Answer: C. 50,000 × 1.08^3

Question 3

What is the break-even quantity when fixed cost is $12,000, price is $50 and variable cost is $18?

A. 240 units

B. 375 units

C. 600 units

D. 1,200 units

Correct Answer: B. 375 units

Question 4

Which condition is required for a real logarithm log_b(x)?

A. b = 1 and x > 0

B. b > 0, b ≠ 1 and x > 0

C. b < 0 and x ≥ 0

D. x may be any real number

Correct Answer: B. b > 0, b ≠ 1 and x > 0

Question 5

In the linear function y = a + bx, what does b represent?

A. The domain

B. The intercept

C. The slope

D. The maximum output

Correct Answer: C. The slope

Question 6

If total cost is C(q) = 12,000 + 18q, what is marginal cost?

A. 12,000 dollars per unit

B. 18 dollars per unit

C. 18q dollars

D. 12,018 dollars

Correct Answer: B. 18 dollars per unit

Question 7

A matrix A has dimensions 3 × 2 and matrix B has dimensions 2 × 4. What are the dimensions of AB?

A. 2 × 3

B. 3 × 4

C. 2 × 4

D. 3 × 2

Correct Answer: B. 3 × 4

Question 8

For π(q) = 80q − 0.2q² − 1,000, what quantity satisfies π′(q) = 0?

A. 80

B. 100

C. 200

D. 400

Correct Answer: C. 200

Question 9

A decrease of 20% followed by an increase of 20% changes an initial value of 100 to:

A. 96

B. 100

C. 104

D. 140

Correct Answer: A. 96

Question 10

Which practice best improves the reliability of a quantitative business model?

A. Rounding every intermediate result

B. Ignoring units to simplify presentation

C. Documenting assumptions and testing sensitivity

D. Extrapolating indefinitely from a short dataset

Correct Answer: C. Documenting assumptions and testing sensitivity