Online Advertising Cards

This deck introduces the building blocks of online advertising, from basic concepts like impressions and clicks to the key metrics marketers use every day. You'll find clear definitions for terms such as CTR, CPC, CPM, CPA, and ROAS, along with explanations of related ideas like conversion rate, frequency caps, and customer lifetime value. It's designed as a friendly glossary that helps you connect the dots between how ads are bought, measured, and optimized.

It's a great fit for anyone new to digital marketing, whether you're a small business owner running your first campaigns, a student preparing for a course or certification, or a working professional who wants a quick refresher on the terminology. If you've ever felt lost when colleagues mention ROAS, CAC, or break-even targets, these cards will give you a solid foundation to follow conversations and ask better questions.

To get the most out of these cards, try studying in short sessions over several days rather than cramming everything at once, since the metrics often build on one another. When you review, pause after each term and try to come up with your own example, like imagining a small online store and plugging in numbers for clicks, conversions, and cost. This kind of active recall, paired with the spaced repetition built into the deck, is the fastest way to turn these definitions into working knowledge you can apply on the job.

Online advertising is the paid distribution of messages across digital channels to drive measurable outcomes such as awareness, leads, or sales. At the most basic level, it is built on a few primitives. An impression is a single ad view opportunity served, or counted as served, to a user or device. A click is a user interaction that sends them from the ad to a destination such as a website, app, or store. From these two primitives, derived ratios become the foundation of performance analysis. Click-through rate (CTR) is the ratio of clicks to impressions, while conversion rate (CVR) is the ratio of conversions to clicks, though the exact denominator should be defined consistently across reports.

A conversion is any tracked action aligned to a business goal, such as a purchase, lead, signup, or install. Because a user may see or click an ad long before converting, advertisers define a conversion window: the period after an ad interaction during which a conversion can still be attributed to that interaction, for example a 7-day click window or a 1-day view window. The customer journey is often modeled as a funnel with three stages, top of funnel (TOFU) for awareness, middle (MOFU) for consideration, and bottom (BOFU) for conversion, and each stage calls for different creative, targeting, and measurement approaches.

When a user clicks an ad, they typically arrive on a landing page designed specifically to convert that traffic. The most important principle of landing page design is message match: the promise and creative of the ad should be consistent with what the user sees on the page. Strong message match reduces bounce rates and improves conversion rate, while mismatched ads and pages erode the value of the click.

Beyond CTR and CVR, advertisers track a family of cost and return metrics. Cost per click (CPC) is spend divided by clicks. Cost per mille (CPM) is the cost per one thousand impressions, calculated as (spend divided by impressions) multiplied by 1000. Cost per acquisition or action (CPA) is spend divided by conversions. Return on ad spend (ROAS) is revenue attributed to ads divided by ad spend. Importantly, ROAS is not the same as return on investment (ROI): ROAS uses revenue and excludes non-ad costs, while ROI accounts for profit or net return after all costs, so a campaign can be ROAS-positive but unprofitable once product costs are included.

Two metrics connect advertising to the broader business: customer lifetime value (LTV) and customer acquisition cost (CAC). LTV is the expected gross margin or revenue from a customer over the entire relationship with the business, while CAC is the total acquisition spend divided by new customers acquired, with both the cost scope and time window defined clearly. For a paid channel to be sustainable, LTV should meaningfully exceed CAC. The break-even ROAS is the ratio of revenue to ad spend required to cover the cost of goods and variable costs; in revenue-based form, it is often expressed as 1 divided by the contribution margin.

Understanding what drives CPA is central to optimization. At a high level, CPA approximates CPC divided by CVR, or equivalently CPM divided by (CTR multiplied by CVR). Lowering CPA therefore depends on either reducing CPC or CPM, or improving CTR and CVR. Two related audience metrics are reach, the number of unique people or devices exposed to an ad, and frequency, the average number of times each person or device sees the ad over a period. Frequency caps limit how often a person sees an ad, which reduces waste, manages creative fatigue, and improves the efficiency of reach.

Audience strategy in paid media typically splits into two complementary approaches. Prospecting targets new audiences who have not yet engaged with the brand, with the goal of building awareness and feeding the top of the funnel. Retargeting, also called remarketing, shows ads to people who previously engaged with the brand's site, app, or content. Retargeting tends to perform best when aimed at high-intent audiences with recent engagement, paired with a strong offer and controlled frequency, because those users are already familiar with the brand and closer to converting.

Once an audience is exposed to ads, the creative itself becomes a limiting factor. In social and other machine-learning-driven ad systems, new campaigns enter a learning phase during which the platform explores delivery to find the best-performing users and contexts. Performance during this phase is unstable, and frequent edits to targeting, creative, or budget can reset the learning process and delay optimization. Even after learning stabilizes, audiences eventually experience creative fatigue: repeated exposure to the same hook, format, or offer causes CTR and CVR to decay over time.

Fighting fatigue requires rotating in new creatives, refreshing hooks and offers, expanding the audience pool, and tightening frequency caps. The disciplined way to test which change actually moves performance is A/B testing, a controlled comparison of a single variable, such as a creative, landing page, or bid strategy, with consistent measurement. Testing multiple variables at once confounds results because any observed lift cannot be attributed to a specific change; isolating one variable at a time preserves the ability to draw clear conclusions.

Attribution is the set of rules that decides which ad interaction gets credit for a conversion. Click-through attribution credits conversions that occur after a user clicks an ad, while view-through attribution credits conversions that occur after an impression even if the user never clicked. Last-click attribution is the simplest model: it credits 100% of a conversion to the most recent eligible click before the conversion. Last-click is easy to implement but can be misleading because it over-credits bottom-funnel touchpoints and under-credits channels that create demand earlier in the journey.

Because attribution models are heuristics rather than ground truth, sophisticated advertisers measure incrementality: the causal lift in outcomes caused by ads, compared with what would have happened without them. Incrementality is typically measured through holdout tests, geo tests, split tests, or randomized experiments that include a control group. These designs isolate the true contribution of advertising and help allocate budget more rationally than rule-based attribution alone.

The technical plumbing that makes attribution work includes UTM parameters, which are appended to URLs to pass campaign, source, and medium data into analytics; tracking pixels, which are JavaScript tags that record events such as pageviews and conversions and tie them back to ad interactions; and server-side conversion tracking, in which conversion events are sent from the advertiser's own server to ad platforms. Server-side tracking is increasingly important because browser restrictions and ad blockers can cause client-side signals to be lost, degrading measurement and optimization.

Search advertising operates on its own set of mechanics. Quality Score is a platform estimate of the relevance of an ad and its keyword, as well as expected performance, and it affects both ad rank and the CPC an advertiser pays. Ad Rank is the auction score used to order ads on the results page; it is typically a function of bid, quality components, and the expected impact of ad extensions. Together, these signals determine not only position but also cost.

Impression share is the share of impressions an advertiser received out of the total they were eligible to receive, and it is a quick diagnostic for lost opportunity. Impression share can be reduced by budget limits, low bids, low quality, tight targeting, or low ad rank. Improving impression share usually means addressing one of those constraints: raising bids or budgets, lifting quality through better ad relevance and landing pages, or relaxing overly narrow targeting.

Search keywords can be matched to user queries at different levels of strictness. Broad match reaches the widest set of queries, phrase match requires the query to include the meaning or phrasing of the keyword, and exact match is the tightest, though specific behaviors vary by platform and continue to evolve. Negative keywords prevent ads from showing on irrelevant queries, and the search term report, the log of actual queries that triggered an ad, is one of the most useful diagnostics in search: it surfaces new converting queries to add as keywords and irrelevant ones to negate. Ad extensions, sometimes called assets, add extra information such as sitelinks, callouts, and structured snippets, and they tend to improve CTR and overall ad utility.

A bid strategy is the set of rules, automated or manual, that determines how much to bid in each auction in pursuit of an objective such as a target CPA, a target ROAS, or maximum clicks. Target CPA (tCPA) is appropriate when the value of each conversion is roughly uniform, because it focuses bidding on hitting a specific cost per action. Target ROAS (tROAS) is preferable when conversion values vary and the value of each outcome matters; it lets the system bid more aggressively for high-value users. Conversion value optimization extends this idea by bidding toward higher-value outcomes using reported conversion values rather than treating all conversions as equal.

Underneath every bid strategy is an ad auction: a real-time selection process that decides which ad is shown and at what price based on bids and quality or utility signals. In a first-price auction, the winner pays exactly what they bid. In a second-price auction, the winner pays just above the next highest bid. Many modern markets use hybrid designs that draw on both, and the exact auction shape matters because it changes how bidders should strategize and how prices converge to market value.

Programmatic advertising is the automated buying and selling of ad inventory through software and auctions, often via real-time bidding (RTB). In an RTB system, a bid request, a package of information about an impression opportunity including context, device, and user signals, is sent to potential buyers, who respond with bids in milliseconds. The infrastructure is typically organized around three roles: the demand-side platform (DSP), which buys on behalf of advertisers; the supply-side platform (SSP), which sells on behalf of publishers; and the ad exchange, which matches bids with inventory. Understanding these roles is essential for diagnosing where money is spent, who has visibility into what, and how to control quality.

Even after an ad is delivered, several quality and safety concerns shape whether it actually works. Viewability measures whether an ad had a real opportunity to be seen, typically defined by standards like a percentage of pixels in view for a minimum duration. Non-viewable impressions waste spend and can inflate reach and frequency while deflating CTR and CVR, which is why viewability is a baseline hygiene metric for any large campaign.

Invalid traffic (IVT), commonly called ad fraud, refers to non-human or deceptive activity that generates fake impressions, clicks, or conversions. Common anti-fraud controls include pre-bid filters that screen inventory before purchase, allowlists and blocklists of approved or excluded domains and apps, third-party verification vendors, anomaly monitoring on traffic patterns, and post-bid clawbacks for traffic that is later deemed fraudulent. Brand safety is the related discipline of ensuring ads do not appear next to content that could harm the advertiser's reputation; an allowlist restricts delivery to approved placements, while a blocklist excludes risky ones.

Targeting itself is also evolving. Contextual targeting matches ads to page or app content rather than to user identity, which makes it more resilient in a privacy-constrained environment. Lookalike or similar audience targeting builds model-based audiences that resemble a seed list such as customers or site visitors, but these audiences can degrade if the seed is small or biased, if tracking signals are lost, if the market saturates, or if the underlying model changes. Industry-wide, cookie deprecation is reducing the ability to track and target across sites, affecting attribution, retargeting, and measurement. On mobile, Apple's App Tracking Transparency (ATT) policy requires permission to track users across apps and sites, reducing available device identifiers, and SKAdNetwork provides a privacy-preserving install attribution framework that returns delayed, aggregated conversion data. A consent management platform (CMP) collects and stores user consent choices and signals them to vendors so that advertising and analytics respect applicable privacy rules.