Paid Search and Shopping Advertising

Introduction

Paid search advertising places sponsored messages in response to a user’s search query. Shopping advertising presents product information—such as title, image, price and availability—within search-engine shopping results. Local campaigns connect search demand with physical locations, directions, telephone calls and store visits. Together, these activities form a central part of search engine marketing (SEM), because they allow an organisation to respond to expressed consumer intent.

Unlike many forms of interruption advertising, paid search is usually activated by a user’s declared need. A person searching for “emergency plumber near me” has different commercial value and urgency from a person searching for “how to repair a leaking tap.” Effective campaign management therefore depends on matching search intent, keywords, landing pages, offers and measurement.

Search platforms use auctions to determine whether an advertisement appears, where it appears and what the advertiser pays. The advertiser does not simply purchase a fixed position. Performance depends on bid, relevance, expected user experience, competition, context and the platform’s optimisation systems.

Learning Objectives

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

1. Explain search intent and its importance in campaign planning.
2. Design practical account, campaign and ad-group structures.
3. Distinguish keyword match types and use negative keywords responsibly.
4. Explain quality factors, ad rank and landing-page relevance.
5. Plan search ads, extensions or assets, shopping feeds and local campaigns.
6. Select bidding strategies and calculate key SEM performance metrics.
7. Evaluate campaign performance and make evidence-based optimisation decisions.
8. Identify ethical, privacy, regulatory and operational risks in paid advertising.

1. Search Intent and the Paid Search Value Chain

Search intent is the probable purpose behind a query. Four useful categories are:

The categories overlap. A query such as “dentist appointment Manchester” has local and transactional intent. Intent must therefore be inferred from the complete query, location, device, time, previous behaviour and landing-page context—not from a keyword in isolation.

The paid search value chain can be expressed as:

query → eligibility → auction → impression → click or interaction → landing-page experience → conversion → business value

A campaign may fail at any stage. Relevant keywords are ineffective if the advertisement is disapproved; a high click-through rate is not commercially useful if the landing page is slow or the traffic does not convert.

Keyword research

Keyword research combines quantitative evidence and qualitative judgement. Useful sources include search-query reports, website search data, customer-service questions, competitor observation, keyword-planning tools and sales-team knowledge. The objective is not to collect the largest possible list. It is to identify queries for which the organisation can provide a relevant response at an economically acceptable cost.

A useful research record contains:

• the keyword or query theme;
• likely intent;
• product or service category;
• expected commercial value;
• geographic and language constraints;
• proposed landing page;
• positive and negative keyword considerations;
• measurement event.

Keyword volume is not the same as opportunity. High-volume generic terms may generate expensive, poorly qualified traffic, while lower-volume specific terms can produce stronger conversion rates and margins.

2. Account and Campaign Architecture

A paid search account should reflect meaningful differences in business objective, geography, budget, product range, language, audience, conversion goal or bidding logic. A common hierarchy is:

account → campaign → ad group or asset group → keyword/product target → advertisement or creative → landing page

The exact terminology varies by platform and campaign type, but the management principle is consistent: group items that should share settings and separate items that require different control.

Campaign-level decisions

Campaigns commonly contain budget, location, language, network, schedule, bidding and conversion-goal settings. Separate campaigns may be appropriate when:

• countries or regions use different currencies, languages or legal requirements;
• product margins differ substantially;
• brand and non-brand activity require separate reporting;
• store locations need distinct budgets;
• seasonal products have different schedules;
• lead quality or conversion values differ.

Over-segmentation is also harmful. If a campaign contains too few impressions or conversions, automated bidding may have insufficient evidence and management becomes unnecessarily complex. The best structure is the simplest one that preserves important control and reporting distinctions.

Ad-group organisation

An ad group should contain a tightly related theme and advertisements that can respond naturally to that theme. For example, a retailer might separate “waterproof hiking boots” from “lightweight trail shoes.” Excessively broad ad groups lead to generic advertisements and weak landing-page alignment. Excessively narrow structures can create duplication and make testing difficult.

Brand, generic, competitor and remarketing search activity should usually be analysed separately. Competitor targeting may be legally and ethically sensitive, and its performance should not be combined with brand traffic, which often has much higher expected conversion rates.

3. Match Types and Negative Keywords

Match types control how closely a user’s query must relate to a targeted keyword. Platform behaviour changes over time, so practitioners should consult current documentation; nevertheless, the following conceptual distinctions remain useful.

Match types are not literal guarantees that only the written phrase will appear. Close variants, spelling differences, reordered terms, implied words and semantically related searches may be included depending on the platform. Consequently, the search-query report is essential.

Negative keyword strategy

Negative keywords should exclude traffic that is irrelevant, uneconomic or operationally impossible. Examples include “free,” “jobs,” “repair manual,” or a product type the organisation does not sell. Negatives may be applied at campaign or ad-group level. A campaign-wide negative can accidentally block a valuable query in another product category, so scope must be checked carefully.

A negative-keyword process should:

1. review actual queries rather than relying only on assumptions;
2. identify irrelevant intent and low-value patterns;
3. distinguish a genuinely poor query from a valuable query with poor ad or landing-page treatment;
4. apply the narrowest safe negative;
5. document the reason and monitor unintended exclusions.

Negative keywords are not a substitute for good account structure. If a campaign repeatedly attracts the wrong product category, restructuring and more precise targeting may be better than accumulating hundreds of exclusions.

4. Auctions, Ad Rank and Quality

A search auction determines eligibility, placement and often price. Simplified explanations frequently describe position as a function of bid and quality. In practice, platforms use more complex systems incorporating expected impact of assets, auction context, user device, location, query meaning, competition and policy eligibility.

Quality-related factors commonly include:

• expected click-through rate;
• advertisement relevance to the query;
• landing-page experience;
• historical performance and contextual signals;
• relevance and usefulness of additional information such as assets.

A high bid cannot reliably compensate for a misleading advertisement or a poor landing page. Conversely, strong relevance may allow an advertiser to compete effectively without the highest bid, although no platform guarantees a particular position or cost.

Landing-page quality

A useful landing page should deliver what the advertisement promised. Important characteristics include:

• clear headline and offer;
• mobile-friendly design;
• fast and stable loading;
• transparent pricing, delivery and returns;
• prominent but trustworthy call to action;
• accessible contact and privacy information;
• accurate product availability;
• minimal friction in forms and checkout.

Tracking tags and consent controls should be implemented without degrading usability. A page that converts poorly can make apparently successful keyword activity commercially unprofitable.

Ad assets and extensions

Modern platforms use assets—formerly often called extensions—to provide additional information. Examples include sitelinks, callouts, structured snippets, telephone assets, location assets, image assets, prices, promotions and lead forms. Assets improve usefulness, visibility and navigation, but they do not guarantee more clicks or a particular rank.

Assets should be:

• factually accurate and current;
• consistent with the landing page;
• specific rather than repetitive;
• compliant with editorial and legal rules;
• selected according to the user’s likely need.

5. Search Advertising Design and Testing

A search advertisement should connect the user’s query with a credible value proposition. Strong messages normally include the product or service, a meaningful benefit, evidence or reassurance, and a clear next step. Claims must be supportable; urgency, discounts and comparative claims must not mislead.

Responsive formats allow the platform to combine multiple headlines and descriptions. Advertisers should provide genuinely distinct but compatible assets rather than many superficial variations. Important information should appear in more than one suitable position because not every combination will display every asset.

Testing should have a defined hypothesis. For example: “A delivery-date message will increase qualified conversion rate for high-intent product searches.” The test should identify the audience, period, primary metric, guardrail metrics and decision rule. A click-through-rate increase is not automatically a success if conversion rate or profit falls.

6. Shopping Advertising and Product Feeds

Shopping advertising is feed-driven. Instead of targeting only a written keyword, the platform evaluates product information and matches it with search and shopping intent. A product feed commonly contains:

• unique product identifier;
• title and description;
• brand and manufacturer information;
• product category and product type;
• price and currency;
• availability;
• high-quality image URL;
• link to the product page;
• global trade item number where applicable;
• condition, colour, size and variant data;
• shipping, tax and regulatory information where required.

Feed quality is a major performance factor. Titles should put important identifying information early, while descriptions should accurately communicate features and use cases. Images must represent the product clearly and comply with platform rules. Price and availability must remain synchronised with the website; discrepancies create disapprovals, poor customer experience and possible loss of trust.

Feed governance

A professional feed process includes:

1. source-system ownership and documented field definitions;
2. scheduled updates appropriate to price and stock volatility;
3. validation before submission;
4. diagnostics monitoring;
5. controlled supplemental data for promotions or seasonal labels;
6. change logs and escalation procedures.

Product segmentation can support management by margin, category, price band, seasonality or inventory status. However, segmentation should not create so many campaigns that budgets and learning are fragmented.

Shopping performance should be evaluated with product-level and category-level data. A product may receive many clicks but produce weak profit because of low margin, high returns or shipping cost. Revenue alone is therefore an incomplete decision measure.

7. Local Search Campaigns

Local campaigns seek actions connected with physical locations. Relevant outcomes include calls, requests for directions, store visits, local website visits and in-store purchases. Effective local advertising requires accurate business profiles, consistent address and telephone information, location targeting, opening hours and location-specific landing pages.

Local intent is affected by distance, service area, operating hours, device, urgency and availability. A campaign for a restaurant should not treat a search during closed hours in the same way as a search during a busy lunch period. Location exclusions and radius settings must be reviewed carefully, especially when a business serves only selected areas.

Offline conversion integration is particularly valuable. If the organisation can connect advertising clicks with qualified calls, booked appointments or completed in-store sales, bidding can optimise for business outcomes rather than shallow online actions.

8. Bidding and Budget Management

Manual and automated bidding represent different control philosophies. Manual approaches provide direct bid control but require time and dependable judgement. Automated strategies use signals and historical data to optimise toward a selected outcome, such as clicks, conversions, conversion value, impression share or a target return on advertising spend (ROAS).

Common strategies include:

• maximise clicks, subject to a budget or bid constraint;
• maximise conversions;
• target cost per action (CPA);
• maximise conversion value;
• target ROAS;
• impression-share or visibility objectives.

Automated bidding does not remove the need for strategy. It requires accurate conversion tracking, sufficient data, realistic targets, stable settings and appropriate exclusions. A target set below economically feasible cost may suppress volume or encourage undesirable traffic. Major changes to budget, targeting, conversion definitions or landing pages can also disrupt learning.

Core metrics

Let:

• I = impressions;
• C = clicks;
• V = conversions;
• S = advertising spend;
• R = attributed conversion revenue or value.

Click-through rate is:

Cost per click is:

Conversion rate, when defined as conversions per click, is:

Cost per conversion is:

Return on advertising spend is:

These formulas assume consistent definitions and attribution windows. If conversions are counted per impression, session or qualified lead rather than per click, the denominator must be changed and clearly labelled. ROAS is a ratio, often displayed as a multiple or percentage, and does not equal profit: product cost, fulfilment, salaries and overheads may still need to be deducted.

Worked example

Suppose a campaign records 40,000 impressions, 1,600 clicks, 80 conversions, advertising spend of £2,400 and attributed revenue of £12,000.

CTR = (1,600 ÷ 40,000) × 100 = 4%

CPC = £2,400 ÷ 1,600 = £1.50

CVR = (80 ÷ 1,600) × 100 = 5%

CPA = £2,400 ÷ 80 = £30

ROAS = £12,000 ÷ £2,400 = 5.0, or 500%

The campaign appears efficient on these measures, but management should still examine profit margin, new-versus-returning customers, cancelled orders, attribution bias and customer lifetime value.

9. Measurement, Attribution and Optimisation

Measurement begins with a conversion specification. A conversion should represent an economically meaningful action, such as a completed purchase, qualified lead, booked appointment or store transaction. Micro-actions such as page views or button clicks may be useful diagnostic events but should not automatically drive bidding.

A measurement plan should define event name, trigger, value, currency, source, consent status, deduplication rule, attribution window and responsible owner. Test implementations in the browser and analytics platform, then reconcile platform-reported results with CRM, sales and finance data.

Attribution is an allocation convention, not a direct observation of causality. Last-click attribution may undervalue earlier interactions; data-driven models depend on data quality and model assumptions. Reported conversions may differ across platforms because of time zones, click windows, cross-device identity, consent choices and counting rules. Trends and incremental experiments are often more informative than treating one attributed number as literal truth.

A practical optimisation cycle is:

1. diagnose performance by query, product, device, location, audience and time;
2. identify the largest controllable problem;
3. form a hypothesis;
4. change one meaningful element or use a controlled test;
5. monitor primary and guardrail metrics;
6. document the outcome and retain or reverse the change.

Guardrail metrics might include conversion quality, refund rate, impression share, brand-query coverage, average order value or profit.

10. Illustrative Case: UrbanTrail Outdoor Retailer

Background

UrbanTrail is a fictional mid-sized retailer selling hiking equipment through an online store and six physical branches. Its paid search account contains a single mixed campaign for brand terms, generic footwear terms and local store searches. Product titles are inconsistent, stock information updates only once daily, and the organisation reports online revenue but not store sales.

Problem

Traffic has increased, yet profit has not. Generic searches attract searches for “free hiking gear” and employment opportunities. The same budget is used for high-margin boots and low-margin accessories. Local branches report that customers often arrive after seeing products online, but those interactions are not recorded.

Analysis

The account is structurally unable to distinguish brand demand from incremental generic demand. Search-query analysis identifies irrelevant themes. Product-level data shows that several heavily advertised products have high return rates. The feed also contains titles such as “Boot 300,” which provide little information to shoppers.

Approach and solution

UrbanTrail creates separate brand, generic footwear, shopping and local campaigns. It applies carefully scoped negative keywords for employment, free products and repair instructions. Generic ad groups are organised by product intent, with relevant landing pages for waterproof boots, trail shoes and children’s footwear.

The retailer rewrites feed titles to include brand, model, product type and key attributes; synchronises stock more frequently; and establishes a feed-diagnostic owner. Location assets and branch-specific pages are added. Calls and store-direction actions are tracked, while completed store purchases are uploaded where consent and data governance permit.

Bidding is initially based on reliable purchase and qualified-lead conversions rather than every website interaction. Product segmentation incorporates margin and stock availability. A test compares a delivery reassurance message with a generic quality message, using qualified conversion rate and contribution margin as decision measures.

Outcome

In this illustrative case, the organisation obtains clearer reporting, reduces irrelevant clicks and shifts budget toward products and locations with stronger economic value. The most important improvement is not a single advertisement; it is the connection among query control, feed accuracy, landing-page relevance, conversion quality and financial reporting.

Learning points

1. Account structure should reflect decisions the organisation actually needs to make.
2. Search-query and product-level evidence are more useful than assumptions alone.
3. Revenue and platform conversions should be connected to margin and offline outcomes.
4. Feed governance is an operational responsibility, not merely a media task.
5. Local advertising is strongest when online and in-store measurement are integrated.

11. Ethical, Legal and Governance Considerations

Paid advertising influences consumer choices and uses personal or contextual data. Advertisers should comply with applicable data-protection, consumer-protection, advertising and platform rules. Consent requirements vary by jurisdiction and by the technologies used; legal advice may be necessary for complex implementations.

Important governance principles include:

• do not make unsubstantiated health, financial, environmental or performance claims;
• disclose material conditions, exclusions and recurring charges;
• avoid discriminatory targeting or exclusionary delivery practices;
• protect customer identifiers and restrict access to advertising data;
• respect sensitive categories and avoid inferring vulnerable personal characteristics;
• provide accessible experiences for people with disabilities;
• maintain approval records and an escalation process for policy changes.

Automation and AI can generate keywords, advertisements, product attributes, forecasts and recommendations. These tools can improve efficiency but may reproduce bias, invent claims, misclassify products or obscure why traffic was purchased. Human review, source verification, brand safeguards and transparent change control remain necessary.

12. Contemporary Developments and Professional Practice

Search platforms increasingly combine keyword, audience, product, creative and automated bidding signals. Search-result pages also contain maps, images, videos, answer formats and merchant experiences, reducing the distinction between traditional search, shopping and local discovery. Privacy changes and reduced availability of third-party identifiers increase the value of first-party data, consent quality, server-side measurement where appropriate, and robust aggregate analysis.

Professional practitioners should maintain a testing calendar, naming convention, change log, budget forecast and incident procedure. They should be able to explain not only what happened to a metric, but why it changed, how reliable the evidence is, and what business decision follows.

Chapter Summary

Paid search and shopping advertising convert expressed demand into measurable advertising opportunities. Effective management begins with search intent and continues through account architecture, match types, negative keywords, auction relevance, landing-page experience, advertisement assets, feed governance and local information accuracy. Shopping campaigns depend heavily on product data, while local campaigns require reliable location and offline measurement.

Bidding strategies must be aligned with accurate conversion definitions and economically meaningful goals. CTR, CPC, conversion rate, CPA and ROAS are useful but incomplete; profit, customer quality, returns, lifetime value and incrementality may change the decision. Continuous optimisation requires query and product diagnosis, explicit hypotheses, controlled tests and appropriate guardrails. Ethical claims, privacy, accessibility, fairness and human oversight are essential to responsible practice.

Key Terms

Search intent: The probable purpose or need behind a user’s search query.

Paid search: Advertising displayed in response to search activity.

Shopping advertising: Product-focused advertising using structured product information.

Local campaign: Advertising designed to generate actions connected with physical locations.

Keyword: A term or theme used to help determine search-ad eligibility.

Match type: A rule controlling how closely a query relates to a keyword target.

Negative keyword: A term or theme used to prevent unwanted ad triggering.

Ad group: A group of closely related targets and advertisements.

Ad asset: Additional information such as a sitelink, callout or location detail.

Landing page: The webpage reached after an advertisement interaction.

Quality factor: A platform assessment of relevance and expected user experience.

Ad rank: A platform’s auction-related assessment affecting eligibility and position.

Product feed: Structured data describing products for commerce advertising.

Conversion: A defined action that has business significance.

Attribution: A rule for assigning credit for a conversion to interactions.

CTR: Click-through rate, calculated as clicks divided by impressions.

CPC: Cost per click.

CPA: Cost per conversion or acquisition.

ROAS: Revenue or conversion value divided by advertising spend.

Incrementality: The additional outcome caused by advertising compared with a suitable counterfactual.

Review Questions

Short Answer Questions

1. What is search intent, and why should it influence campaign structure?
2. Distinguish broad, phrase, exact and negative keyword concepts.
3. Identify four important fields in a shopping product feed.
4. Why can a high click-through rate fail to produce business success?
5. What is the difference between a conversion and a micro-action?

Long Answer Questions

1. Explain how an advertiser can design an account structure that balances control with sufficient data for optimisation.
2. Discuss the relationship among advertisement relevance, landing-page experience, quality factors and auction performance.
3. Describe a complete process for creating, validating and governing a shopping feed.
4. Compare manual and automated bidding, including the conditions required for responsible automated bidding.
5. Explain why attribution-platform results may differ from CRM and finance results.

Analytical / Application Questions

1. A retailer receives 25,000 impressions, 1,000 clicks, 50 conversions and spends £1,500. Calculate CTR, CPC, conversion rate and CPA. Interpret one strength and one potential concern.
2. A local service campaign receives many calls but few qualified appointments. Design a diagnostic plan covering query intent, advertisement promises, call quality, landing page and conversion tracking.
3. A shopping campaign produces strong revenue but negative profit because advertised products have low margins and high returns. Recommend changes to segmentation, bidding, feed management and performance measurement.

Multiple Choice Questions with Answers

Question 1

Which query most clearly shows transactional local intent?

A. how to clean hiking boots

B. best hiking boots for beginners

C. buy waterproof hiking boots near Leeds

D. history of hiking footwear

Correct Answer: C. buy waterproof hiking boots near Leeds

Question 2

What is the primary purpose of a negative keyword?

A. To increase the maximum bid

B. To prevent ads from showing for unwanted query themes

C. To improve product-image resolution

D. To create a new landing page

Correct Answer: B. To prevent ads from showing for unwanted query themes

Question 3

Which campaign-level distinction most strongly justifies separate campaigns?

A. Two advertisements use different punctuation

B. Two keywords contain the same number of words

C. Two regions require different currencies and budgets

D. Two headlines have different lengths

Correct Answer: C. Two regions require different currencies and budgets

Question 4

Which item is most directly associated with shopping-feed quality?

A. Accurate price and availability

B. The office seating plan

C. The advertiser’s annual leave policy

D. The number of sales staff

Correct Answer: A. Accurate price and availability

Question 5

A campaign has 10,000 impressions and 500 clicks. What is its CTR?

A. 0.5%

B. 5%

C. 20%

D. 50%

Correct Answer: B. 5%

Question 6

A campaign spends £800 and records 40 conversions. What is its CPA?

A. £20

B. £32

C. £40

D. £800

Correct Answer: A. £20

Question 7

Which statement about ROAS is correct?

A. It is always equal to profit margin

B. It measures revenue or conversion value relative to advertising spend

C. It is calculated as impressions divided by clicks

D. It excludes all attribution assumptions

Correct Answer: B. It measures revenue or conversion value relative to advertising spend

Question 8

Why should brand and generic search campaigns often be reported separately?

A. They always use different currencies

B. Brand traffic often has different intent and baseline performance

C. Generic campaigns cannot use negative keywords

D. Brand campaigns cannot use landing pages

Correct Answer: B. Brand traffic often has different intent and baseline performance

Question 9

Which is the strongest example of a meaningful bidding conversion for a local service business?

A. A page view

B. A five-second visit

C. A qualified appointment booking

D. A browser refresh

Correct Answer: C. A qualified appointment booking

Question 10

Why is human review important when AI generates paid-search assets?

A. AI can never produce text

B. AI-generated content may contain unsupported claims or policy violations

C. Human review guarantees the lowest CPC

D. AI tools automatically eliminate attribution problems

Correct Answer: B. AI-generated content may contain unsupported claims or policy violations