Online Marketing: Strategy, Tools, and Hands-On Practice

The Human Knowledge Project

Chapter 09 — Paid Online Advertising

Introduction

Paid online advertising allows an organization to purchase access to an audience rather than waiting for that audience to discover it organically.

Common forms include:

Advertising can produce traffic quickly, but traffic alone is not success.

A campaign can generate thousands of impressions and clicks while producing little useful value if the audience, message, offer, landing page, or measurement system is poorly designed.

Paid advertising also introduces direct financial risk. Every experiment can consume money.

For that reason, effective advertising begins before the first advertisement is purchased.

The marketer should know:

The purpose of paid advertising is not simply to buy attention.

It is to purchase measurable opportunities to achieve a defined marketing objective.

1. Start With the Objective

An advertising campaign should begin with a specific objective rather than with the advertisement itself.

Possible objectives include:

The objective determines what should be measured.

If the objective is awareness, impressions or reach may be relevant.

If the objective is website traffic, qualified visits and their cost may matter more.

If the objective is enrollment or sales, clicks alone are insufficient. The marketer needs to know whether those clicks produce the intended action.

A useful campaign objective connects advertising activity to a meaningful result.

For example:

"Generate 100 website visits" describes traffic.

"Generate 20 qualified course registrations at an acceptable acquisition cost" describes a business or organizational result.

Define the desired result before spending money. Otherwise, impressive advertising statistics can conceal a campaign that is accomplishing very little.

2. Know What a Conversion Is

A conversion occurs when a person completes an action that the campaign is intended to produce.

Depending on the objective, a conversion might be:

The conversion should be defined before the campaign begins.

This matters because advertising platforms can report many activities that look encouraging without demonstrating that the campaign achieved its real purpose.

For example, an advertisement may receive 1,000 clicks but produce only two registrations.

The clicks demonstrate that people responded to the advertisement. They do not demonstrate that the complete marketing process worked effectively.

The marketer should therefore distinguish between:

advertising response — such as an impression or click

and

desired outcome — the meaningful action the campaign was designed to produce.

This distinction becomes essential when calculating advertising cost and deciding whether a campaign should continue.

3. Major Advertising Models

Online advertising can be purchased and delivered in several ways.

Common models include:

Pay per click (PPC) — the advertiser generally pays when someone clicks the advertisement.

Cost per thousand impressions (CPM) — the advertiser pays based on the number of times advertisements are displayed.

Cost per action or acquisition (CPA) — cost is evaluated in relation to a defined action or conversion.

Different advertising systems may use different pricing, bidding, and optimization methods, but the underlying marketing question remains the same:

What useful result was obtained for the money spent?

A low cost per click is not automatically good.

Suppose Campaign A produces clicks for $0.50 but very few visitors convert, while Campaign B produces clicks for $2.00 but those visitors frequently complete the desired action.

Campaign B may provide much greater value despite having the higher click cost.

Advertising costs should therefore be interpreted in relation to audience quality, conversion behavior, and the campaign objective.

Do not optimize a campaign for the cheapest activity when a more meaningful result can be measured.

4. Search Advertising

Search advertising places paid messages in or near search results when people search for particular words or phrases.

Its major advantage is intent.

Someone searching for:

"online algebra course"

is expressing a different level of immediate interest from someone who happens to see an education advertisement while reading unrelated content.

Search campaigns commonly involve decisions about:

Keyword selection should reflect the language and intent of the intended audience.

Broad searches may produce substantial traffic but include many people who are poor matches.

More specific searches may produce fewer impressions while attracting people whose needs more closely match the offer.

The advertisement and destination should also agree.

If an advertisement promises information about a particular course, sending the visitor to a general homepage forces that person to search again.

A more relevant landing page usually provides a clearer path from the searcher's question to the promised information and desired action.

5. Social-Media Advertising

Social-media advertising places paid messages within platforms where people are consuming content, communicating, or participating in communities.

Unlike many search advertisements, these advertisements often reach people who are not actively searching for the product, service, or resource at that moment.

Targeting may use factors such as:

Available targeting methods vary by platform and can change over time.

The marketer should therefore begin with the audience rather than with the platform's targeting controls.

Ask:

Who are we trying to reach?

Why might this message matter to them?

What evidence suggests this audience is appropriate?

What action should they take next?

Social advertising can be useful for introducing an organization to new audiences, distributing valuable content, promoting events or resources, and reconnecting with people who have already shown interest.

Precise-looking targeting settings do not guarantee a precisely matched audience. Actual campaign results must still be measured.

6. Display, Video, and Sponsored Advertising

Paid advertising is not limited to search engines and social-media feeds.

Other forms include:

These channels can be useful when an organization wants exposure to an audience gathered around a particular subject or source.

For example, a specialized newsletter with a modest readership may be valuable if its subscribers closely match the intended audience.

Evaluate these opportunities by asking:

Do not purchase advertising merely because the audience is large.

A smaller, highly relevant audience may produce better results than a much larger audience with little interest in the offer.

As with every paid channel, the value of the placement should ultimately be judged by evidence rather than by the size or prestige of the publication or platform.

7. Audience Targeting

Advertising becomes wasteful when money is spent reaching people who have little reason to care about the offer.

Audience targeting attempts to reduce that waste.

Targeting decisions may involve:

Targeting should follow the audience research developed earlier in this course.

Do not create an advertising audience merely because a platform provides a targeting option. Each important targeting choice should have a reason.

Targeting can also become too narrow.

An audience defined by too many assumptions may exclude people who could legitimately benefit from the offer and may provide too little data for useful testing.

Begin with a defensible audience hypothesis, measure actual response, and refine the audience as evidence accumulates.

Targeting is not proof that the correct people were reached. It is a testable attempt to reach them more efficiently.

8. Budgeting and Spending Limits

Paid advertising should operate within a predetermined financial limit.

Before launching a campaign, define:

A small initial test can provide information before a larger amount of money is committed.

For example, instead of immediately committing an entire monthly advertising budget to one campaign, a marketer might run a controlled test to learn:

Advertising platforms may offer automated recommendations intended to increase traffic, reach, or spending.

Treat those recommendations as proposals, not as commands.

The platform and the advertiser do not necessarily have identical objectives. The advertiser is responsible for deciding whether additional spending serves the organization's actual goal.

Set financial boundaries before the campaign begins and change them deliberately rather than reactively.

9. Cost per Acquisition

One of the most useful advertising measurements is the cost required to produce a desired result.

If a campaign spends $200 and produces 20 registrations, the advertising cost per registration is $10.

The general idea is:

Cost per acquisition = advertising cost divided by the number of acquisitions

The acquisition must be defined clearly.

Depending on the campaign, it might mean:

Cost per acquisition helps connect advertising spending to results.

It should not be interpreted without context.

A $20 acquisition cost could be excellent for one organization and unsustainable for another. The acceptable amount depends on the economic or organizational value of the result.

The marketer should therefore ask:

The objective is not necessarily to obtain the lowest possible cost.

The objective is to obtain useful results at a cost that makes sense for the organization.

10. Return on Advertising Spend

When advertising directly produces measurable revenue, the marketer may compare that revenue with advertising cost.

A common measurement is return on advertising spend, often abbreviated ROAS.

ROAS = revenue attributed to advertising divided by advertising cost

For example, if $500 in advertising produces $2,000 in attributable revenue, the ROAS is 4.

This can also be expressed as $4 in revenue for every $1 spent on advertising.

ROAS is useful, but it is not the same as profit.

Revenue may still need to cover:

Not every organization or campaign is designed to generate immediate revenue. A nonprofit, educational organization, or awareness campaign may need different measures of value.

Use financial measurements when they fit the objective, but do not force every campaign into a revenue calculation.

The larger principle is to connect advertising cost to the most meaningful outcome that can reasonably be measured.

11. The Advertisement and the Landing Page

An advertisement does not operate by itself.

It creates an expectation that the destination must fulfill.

If an advertisement promises:

the landing page should make that promised subject immediately recognizable.

This relationship is called message match.

Poor message match creates friction.

A visitor who clicks an advertisement for a specific resource and arrives at a general homepage may wonder whether the promised resource exists or where to find it.

A strong landing page should help the visitor quickly understand:

Advertising performance therefore cannot be evaluated only by examining the advertisement.

A weak landing page can make a good advertisement appear ineffective, while a misleading advertisement can produce clicks that never had a realistic chance of converting.

Evaluate the complete path from advertisement to outcome.

12. Writing the Advertisement

An advertisement has very little time and space to communicate.

Its message should quickly help the intended audience understand:

Specific language is usually more useful than vague promotional language.

Compare:

"Transform your future with an amazing learning experience."

with:

"Learn introductory algebra online with step-by-step lessons and practice exercises."

The second statement communicates more concrete information.

Effective advertising copy should also remain truthful.

Avoid:

The purpose is not to obtain every possible click.

An advertisement can also help unsuitable prospects decide not to click. That can save money and improve the quality of the traffic that does arrive.

Write for the audience you actually want to reach.

13. Testing Advertisements

Paid advertising provides an opportunity to test assumptions with measurable evidence.

A useful test changes one important variable while keeping other conditions as similar as practical.

Variables might include:

For example, two advertisements might use the same audience, budget, offer, and landing page but different headlines.

If many elements change at the same time, it becomes harder to determine what caused the difference in results.

Tests also need enough evidence to be meaningful.

A result based on a handful of impressions or clicks may reflect chance rather than a reliable difference.

Do not declare a winner simply because one version briefly performs better.

Examine the metric connected to the actual objective.

A headline that produces more clicks is not necessarily better if those additional visitors rarely complete the desired action.

Record what was tested, what happened, and what you learned.

The purpose of testing is not merely to improve an advertisement. It is to improve your understanding of the audience and the marketing system.

14. Retargeting and Frequency

Some advertising systems allow an organization to show advertisements to people who have previously interacted with its website, content, or other digital properties.

This is commonly called retargeting or remarketing.

A previous visitor may already recognize the organization and therefore respond differently from someone encountering it for the first time.

Retargeting might be used to reach people who:

Retargeting should be used carefully.

Repeatedly showing the same advertisement can become annoying and may damage rather than strengthen the relationship.

Frequency measures how often people are exposed to advertising over a period of time.

Higher frequency is not automatically better.

Marketers should watch for signs that additional exposure is producing diminishing results, unnecessary cost, or audience fatigue.

Retargeting also involves privacy considerations. Organizations should follow applicable laws, platform rules, consent requirements, and their own privacy commitments.

The goal is relevant follow-up, not unlimited tracking or repeated pressure.

15. Platform Metrics and Independent Judgment

Advertising platforms provide dashboards containing measurements such as impressions, clicks, engagement, conversions, and cost.

These measurements are useful, but the platform should not be the only source of evidence.

Whenever practical, compare advertising-platform data with information from:

Differences between systems are common because platforms may use different attribution methods, time periods, identifiers, and definitions.

The marketer should understand what a reported conversion actually means before using it to make spending decisions.

Platform recommendations also require independent judgment.

An advertising system may recommend:

Some recommendations may improve performance. Others may primarily increase activity or spending.

Evaluate each recommendation against the campaign's objective, evidence, budget, and organizational interests.

The advertiser remains responsible for deciding what success means.

16. Start Small and Scale From Evidence

A new advertising campaign begins with uncertainty.

The marketer may have good reasons for choosing an audience, message, offer, and channel, but those choices remain hypotheses until real people respond.

For that reason, begin with a controlled test whenever practical.

A useful initial campaign should have:

When results are weak, do not assume that increasing the budget will solve the problem.

Investigate the system.

The problem may involve:

When results are strong, additional spending should also be deliberate.

A campaign that works at a small scale may perform differently when expanded to a larger audience or budget.

Increase spending in measured steps and continue watching the cost and quality of the resulting conversions.

Paid advertising works best as a controlled learning process: test, measure, learn, adjust, and scale only when the evidence supports doing so.

Hands-On Project — Build a Paid Advertising Test Plan

In this project, you will design a small paid advertising experiment for a real organization, product, service, or educational resource.

You do not need to spend money or launch the advertisement. The purpose is to create a campaign plan that could be launched and measured.

A. Define the Campaign

Write down:

  1. The product, service, course, resource, or organization you will advertise.
  2. The specific audience you want to reach.
  3. The problem, need, or interest that connects this audience to the offer.
  4. The primary campaign objective.
  5. The exact conversion you will measure.

Make the objective measurable.

Instead of:

"Get more people interested."

write something such as:

"Generate 25 course registrations from the campaign."

B. Choose the Advertising Channel

Choose one primary advertising channel.

Examples include:

Explain why this channel is appropriate for the audience and objective.

Do not choose a channel simply because it is popular.

C. Define the Budget

Create a small test budget.

Record:

The purpose is to establish financial boundaries before advertising begins.

D. Create the Advertisement

Draft one advertisement for the campaign.

Include:

Then ask whether the advertisement accurately represents what the visitor will find after clicking.

Remove unsupported claims, unnecessary hype, and anything likely to attract irrelevant clicks.

E. Design a Test

Choose one variable to test.

For example:

Keep the other important campaign elements as consistent as practical.

Write down what you expect to happen and why.

F. Create the Measurement Plan

Identify the measurements you will record.

At minimum, consider:

If revenue is relevant, also consider return on advertising spend.

Finally, write three decision rules:

  1. What evidence would cause you to continue the campaign?
  2. What evidence would cause you to change the campaign?
  3. What evidence would cause you to stop the campaign?

A paid campaign should begin with both a spending limit and a learning objective.

Exercises

Concepts and Understanding

1. Why can paid advertising generate traffic without necessarily producing useful results?

2. Why should a campaign objective be defined before advertising begins?

3. What is a conversion?

4. Explain the difference between an advertising response and a desired outcome.

5. What is the difference between PPC and CPM advertising?

6. Why can search advertising be especially useful for reaching people with immediate intent?

7. Why might a specific search phrase produce more valuable traffic than a broad search phrase?

8. How does social-media advertising differ from search advertising in the way it commonly reaches people?

9. Why can a smaller specialized audience sometimes be more valuable than a much larger general audience?

10. Why should audience targeting begin with audience research rather than with the options offered by an advertising platform?

11. What danger can result from defining an advertising audience too narrowly?

12. Why should a marketer establish a spending limit before launching a campaign?

13. What is cost per acquisition, and how is it calculated?

Application and Analysis

14. Why is ROAS not the same as profit?

15. What is message match, and why does it matter in paid advertising?

16. Give one example of advertising language that is too vague and rewrite it to communicate a more specific benefit or offer.

17. Why might an advertisement that produces fewer clicks actually be more effective than one that produces many clicks?

18. When testing two advertisements, why is it useful to change only one important variable at a time?

19. Why should a marketer avoid declaring a test winner after only a very small number of impressions or clicks?

20. What is retargeting?

21. Why can excessive advertising frequency harm campaign performance or audience trust?

22. Why should advertising-platform conversion data sometimes be compared with independent organizational records?

23. Give two reasons an advertising platform's recommendation to increase spending should be evaluated rather than accepted automatically.

24. A campaign spends $300 and produces 15 registrations. What is the advertising cost per registration?

25. A campaign spends $400 and produces $1,600 in attributable revenue. What is its ROAS, and what important information does ROAS alone not tell you?

26. Design a small paid advertising experiment for a real or hypothetical organization. State the audience, objective, channel, budget, conversion, one variable to test, and the evidence that would cause you to continue, change, or stop the campaign.

Chapter 09 — Answer Key: Concepts and Understanding

1.

Paid advertising can generate impressions or clicks from people who never complete the meaningful action the organization needs. Traffic is activity; useful results depend on the campaign objective and conversion.

2.

The objective determines what success means, what should be measured, and whether the campaign is worth continuing.

3.

A conversion occurs when a person completes the intended action, such as registering, purchasing, subscribing, applying, or submitting a qualified inquiry.

4.

An advertising response is an action such as seeing or clicking an advertisement. A desired outcome is the meaningful organizational result the campaign was designed to produce.

5.

PPC generally charges when someone clicks an advertisement. CPM charges according to the number of times an advertisement is displayed, commonly per thousand impressions.

6.

Search advertising can reach people while they are actively expressing a need, question, or interest through their search.

7.

A specific phrase may attract fewer people but better match the need and intent associated with the offer.

8.

Search advertising commonly responds to expressed search intent. Social-media advertising often reaches people while they are consuming content or participating on a platform rather than actively searching for the offer.

9.

A specialized audience may contain a higher proportion of people whose interests and needs closely match the offer.

10.

Audience research provides evidence about whom the organization actually needs to reach. Platform controls are tools for implementing that strategy, not substitutes for it.

11.

Excessive narrowing may exclude legitimate prospects, depend on unsupported assumptions, and produce too little data for useful testing.

12.

A predetermined limit controls financial risk and prevents campaign decisions from becoming an uncontrolled response to platform activity or short-term results.

13.

Cost per acquisition is the advertising cost required for each defined acquisition. It is calculated by dividing advertising cost by the number of acquisitions.

Chapter 09 — Answer Key: Application and Analysis

14.

ROAS compares attributable revenue with advertising cost. It does not account for all other costs, such as labor, fulfillment, fees, support, and overhead, so it is not a measure of profit.

15.

Message match is the consistency between what an advertisement promises and what the visitor finds after clicking. Strong message match reduces confusion and creates a clearer path toward the intended action.

16.

Answers will vary. A vague statement such as "Discover an amazing learning experience" could be replaced with a specific statement such as "Learn introductory algebra online with step-by-step lessons and practice exercises."

17.

The advertisement with fewer clicks may attract a more relevant audience whose members convert at a higher rate or produce more valuable outcomes.

18.

Changing one important variable makes it easier to determine which change contributed to a difference in results.

19.

Very small samples can produce differences caused by chance rather than by a meaningful performance difference.

20.

Retargeting is advertising directed toward people who have previously interacted with an organization's website, content, or other digital properties.

21.

Excessive frequency can create annoyance, waste money, produce diminishing returns, and reduce audience trust.

22.

Independent records can help determine whether platform-reported conversions correspond to actual registrations, sales, inquiries, or other organizational outcomes.

23.

The platform may benefit from increased advertising activity or spending, and its optimization objective may differ from the advertiser's actual organizational objective.

24.

$20 per registration.

25.

The ROAS is 4, meaning $4 in attributable revenue for every $1 of advertising cost. ROAS alone does not show profit because it excludes other business and operating costs.

26.

Answers will vary. A strong response defines a specific audience and objective, selects a justified channel, establishes a limited budget and measurable conversion, changes one meaningful test variable, and states evidence-based rules for continuing, modifying, or stopping the campaign.