The RandallAds™ Marketing Operations Glossary provides organizational stakeholders with a consolidated reference disclosure of contemporary direct-marketing operational terminology, attendant formulas where applicable, and cross-network nomenclature reconciliation. Entries are organized by operational category and presented in alphabetical order within each category. The Glossary is updated on a continuous basis as the practice evolves; current entry count is 44.
- Cost Per Acquisition (CPA)
- Formula: CPA = Spend ÷ Conversions
- The average cost paid for each conversion event.
- Note: Conversion definition varies by campaign objective: purchase, sign-up, lead-form completion, app install, page view of a key page. Always verify the configured conversion event before benchmarking against a comparable campaign.
- Cost Per Click (CPC)
- Formula: CPC = Spend ÷ Clicks
- The average cost paid for each click on an advertisement.
- Note: A primary lever in cost-cap, manual-bid, and target-CPA bid-strategy decisions. Distinct from Maximum CPC (the bid ceiling) and Average CPC (the realized cost).
- Cost Per Install (CPI)
- Formula: CPI = Spend ÷ App Installs
- The average cost paid for each mobile-app install.
- Note: Pair with retention-cohort analysis (Day-1, Day-7, Day-30) to evaluate the true install value. CPI alone obscures install quality.
- Cost Per Lead (CPL)
- Formula: CPL = Spend ÷ Leads
- The average cost paid for each captured lead.
- Note: Lead quality is the load-bearing variable that CPL alone does not measure. A low CPL with a poor lead-to-opportunity rate is not in fact a low cost.
- Cost Per Mille (CPM)
- Formula: CPM = (Spend ÷ Impressions) × 1000
- The cost paid per thousand impressions.
- Note: “Mille” is Latin for one thousand. The unit reflects the practical reality that single-impression pricing is too granular for billing purposes. The standard pricing convention for awareness-tier display and video inventory.
- Effective CPM (eCPM)
- Formula: eCPM = (Total Earnings ÷ Impressions) × 1000
- A normalized revenue metric used by publishers to compare yield across non-CPM-priced inventory (CPC, CPA, revenue-share) on a per-thousand-impression basis.
- Click-Through Rate (CTR)
- Formula: CTR = Clicks ÷ Impressions
- The proportion of impressions that resulted in a click on the advertisement.
- Note: Benchmark within network, format, and vertical. Cross-network CTR comparison is rarely informative; a 0.4% CTR is excellent for display and poor for search.
- Conversion Rate (CVR)
- Formula: CVR = Conversions ÷ Clicks
- The proportion of clicks that resulted in a conversion event.
- Note: Some platforms compute CVR against impressions rather than clicks, or against a custom event funnel. Confirm the denominator before comparing.
- Return on Ad Spend (ROAS)
- Formula: ROAS = Revenue ÷ Spend
- Revenue generated per dollar of advertising spend.
- Note: Typically expressed as a ratio (e.g., 4.0× ROAS = $4 revenue per $1 spent). Distinct from ROI; ROAS does not subtract the spend.
- Return on Investment (ROI)
- Formula: ROI = (Revenue − Spend) ÷ Spend
- Net return per dollar of advertising spend, as a ratio.
- Note: Frequently confused with ROAS. ROI subtracts spend before dividing; ROAS does not. A 4.0× ROAS corresponds to a 3.0× ROI.
- Frequency
- Formula: Frequency = Impressions ÷ Reach
- The average number of times each unique user saw an advertisement during a measurement window.
- Note: High frequency may indicate audience fatigue (declining CTR, rising negative-feedback signals); low frequency may indicate insufficient reinforcement to drive recall.
- Frequency Cap
- A limit on the number of times an advertisement is permitted to serve to a unique user within a defined time window.
- Note: Common formats: 3 impressions per 24 hours, 10 impressions per 7 days. Helps mitigate audience fatigue and concentrate reach budget on incremental users.
- Impressions
- Total count of advertisement views, including multiple views by the same user.
- Note: Networks differ on impression-counting standards: served impressions (delivered to the page), measured impressions (with measurable viewability signal), and viewable impressions (meeting the IAB viewability threshold of 50%-pixels-for-1-second for display, 50%-pixels-for-2-seconds for video).
- Reach
- The number of unique users who saw an advertisement at least once during a measurement window.
- Note: Distinct from Impressions, which counts total ad views including repeat views by the same user.
- Attribution Window
- The period of time after an advertisement interaction during which a subsequent conversion will be credited to that interaction.
- Note: Common windows: 1-day-click, 7-day-click, 28-day-click, 1-day-view, 7-day-view. Longer windows inflate attributed conversions; shorter windows undercount them. Always disclose the window when reporting conversion counts.
- First-Click Attribution
- An attribution model that credits the entire conversion to the first advertisement interaction in the user's path.
- Note: The mirror image of last-click. Over-credits awareness-stage touchpoints.
- Last-Click Attribution
- An attribution model that credits the entire conversion to the most recent advertisement interaction preceding the conversion event.
- Note: The default in most measurement platforms. Tends to over-credit lower-funnel touchpoints (branded search, retargeting) and under-credit upper-funnel awareness activity.
- Multi-Touch Attribution (MTA)
- An attribution model that distributes credit for a conversion across all preceding advertisement interactions, typically by a weighting function such as linear, time-decay, or position-based.
- Note: More accurate in principle, more operationally complex in practice. Walled-garden networks restrict the cross-network signal needed for true MTA, so most MTA implementations approximate within a single network's data.
- View-Through Conversion
- A conversion attributed to an advertisement impression that was viewed but not clicked, where the conversion occurred within the attribution window.
- Note: Often controversial. View-through credit inflates apparent campaign performance and is platform-self-attested. Use with healthy skepticism.
- Ad Rank
- Formula: Ad Rank ≈ Bid × Quality Score (simplified)
- A composite score that determines which advertisements serve and in which positions for a given auction.
- Note: Real ad-rank formulas incorporate format effects, predicted impact of extensions, expected click-through-rate adjustments, and other factors that platforms only partially disclose.
- First-Price Auction
- An auction format in which the winning bidder pays a price equal to the bidder's own winning bid.
- Note: The current standard for most programmatic display inventory. Bidders must internalize the cost discipline that second-price auctions provided automatically.
- Quality Score
- A platform-assigned score (commonly 1–10) reflecting expected advertisement relevance, click-through rate, and landing-page experience. Used in ad-rank determination.
- Note: Improving Quality Score reduces the effective cost-per-click required to maintain a given ad-rank position. Originally a Google Ads concept; analogous mechanisms exist on other networks under different names.
- Second-Price Auction
- An auction format in which the winning bidder pays a price equal to the second-highest bid plus a small increment, rather than the bidder's own bid.
- Note: Historically the Google Ads search auction format. Most major display exchanges transitioned to first-price auctions during 2018–2019 in coordination with the broader header-bidding ecosystem.
- Custom Audience
- A targeting audience defined by uploaded first-party identifier lists (email, phone, mobile advertising identifier) or platform-tracked behaviors (page views, app events, video engagements).
- Note: Match-rate against the platform's user base varies substantially by identifier type and freshness. Email match-rates are generally highest; phone match-rates trail.
- Daypart
- A specific time-of-day window within which advertisements are eligible to serve.
- Note: Used to align ad delivery with business hours, observed audience-activity windows, or budget-pacing constraints. A common configuration restricts serving to weekday business hours plus weekend mornings.
- Lookalike Audience
- A targeting audience constructed by a platform's modeling system to identify users with characteristics similar to a specified seed audience.
- Note: Quality depends heavily on seed-audience size, recency, and homogeneity. A seed audience that is too small (typically under 1,000 users) or too heterogeneous will produce a weak model.
- Retargeting
- Targeting users who have previously interacted with the advertiser's properties (website visit, app open, video view, abandoned cart, product page view).
- Note: Effective for lower-funnel conversion lift but susceptible to over-frequency complaints. Pair with frequency-cap discipline.
- Accelerated Pacing
- A budget-pacing strategy that spends as quickly as available inventory permits until the budget is exhausted.
- Note: Use sparingly. Frequently exhausts the daily budget during low-quality early-day inventory and underspends remaining higher-quality windows.
- Daily Budget
- A budget allocation expressed as the maximum spend permitted in a single calendar day.
- Note: Most platforms permit up to 2× the daily budget on any given day, provided the trailing 7-day spend remains within 7× the daily budget. Pacing-target dashboards should account for this within-period flexibility.
- Even Pacing
- A budget-pacing strategy that distributes spend evenly across the campaign's flight period.
- Note: The default for most campaigns. May leave opportunity unspent during high-availability inventory moments; trade-off is predictability of daily delivery.
- Lifetime Budget
- A budget allocation expressed as the maximum total spend permitted over the campaign's full flight period.
- Note: Distributed across the flight by the platform's pacing logic. Better suited to campaigns with a defined start and end than to always-on activity.
- Google Click Identifier (GCLID)
- An auto-tagged identifier appended to destination URLs for clicks originating from Google Ads, used for offline conversion import and cross-platform attribution.
- Note: Functionally analogous identifiers exist for other networks: FBCLID for Meta, MSCLKID for Microsoft Advertising, TTCLID for TikTok Ads, GBRAID and WBRAID for Google Ads iOS/web app conversions.
- Tracking Pixel
- A small (typically 1×1-pixel transparent) image or script embedded in a webpage or email that fires a request to a measurement endpoint when loaded, enabling impression or interaction logging.
- Note: Increasingly supplanted by server-side conversion APIs as browser-based pixel signal degrades under tracking-prevention measures.
- UTM Parameters
- Standardized URL query-string parameters (utm_source, utm_medium, utm_campaign, utm_term, utm_content) used to identify the source, medium, and context of a click in downstream analytics.
- Note: "UTM" originally stood for Urchin Tracking Module, a reference to Urchin Software, a pre-Google Analytics product Google acquired in 2005. The convention persists; the etymology is largely forgotten.
- Private Marketplace (PMP)
- An auction-based deal type in which a publisher offers a defined inventory pool to a curated set of invited buyers under negotiated minimum-bid floors.
- Note: Bridges the gap between open-exchange programmatic and direct-buy. Higher minimums than open exchange; less rigid than insertion-order direct buys.
- Programmatic
- Automated, auction-driven advertising-inventory purchasing, in contrast to direct-buy inventory contracted by manual insertion order.
- Note: Encompasses open-exchange, private marketplace (PMP), preferred-deal, and programmatic-guaranteed deal types.
- Real-Time Bidding (RTB)
- An auction protocol in which display advertising impressions are auctioned individually as they become available, with the winning bid determined within the page-load window (typically under 100 milliseconds).
- Note: The dominant protocol for open-exchange display inventory. Bid requests carry impression-context signal (URL, user identifiers, device, geo, prior-page-view history) to which bidders respond programmatically.
- A/B Test
- A controlled comparison between two variants (A and B) where users are randomly assigned to one variant and the difference in outcomes is measured.
- Note: Distinguished from multivariate testing, which evaluates multiple-element combinations simultaneously. A/B tests yield cleaner causal claims; multivariate tests yield more variant comparisons per impression.
- Minimum Detectable Effect (MDE)
- The smallest effect size a given test design is statistically equipped to detect at a specified significance and power level, given the sample size and baseline conversion rate.
- Note: If MDE is larger than the operationally meaningful effect size, the test is underpowered: it will likely fail to detect a real but small effect. Increase sample size, accept lower significance, or accept lower power.
- Statistical Significance
- A measurement of the probability that an observed difference between test variants is not due to random chance, conventionally expressed as a p-value or confidence level.
- Note: Common thresholds (p < 0.05, 95% confidence) are conventions, not laws. Statistically significant differences may still be too small to be operationally meaningful; pair significance with effect-size disclosure.
- Above the Fold (ATF)
- Placement in the portion of a webpage that is visible without scrolling.
- Note: The term derives from newspaper layout convention referring to content visible above the physical fold of a folded broadsheet. Increasingly imprecise on responsive layouts and variable-height browser windows; viewability metrics have largely supplanted the term in serious measurement contexts.
- Carousel
- A multi-card advertising format in which the user can swipe or scroll horizontally through a series of related card units within a single ad placement.
- Note: Typical implementations support 2–10 cards. Useful for showcasing product variants, sequential narrative, or feature-by-feature breakdowns.
- In-Feed
- Native advertising placement integrated into the visual flow of a content feed (social-platform feed, news-aggregator feed, e-commerce product listing).
- Note: Performance is highly dependent on creative parity with surrounding organic content. Disclosure conventions vary by jurisdiction and platform.
- In-Stream Video
- Video advertising that plays before (pre-roll), during (mid-roll), or after (post-roll) host video content.
- Note: Subject to skippability conventions that vary by network and host content length. Skippable formats typically permit user dismissal after 5 seconds; non-skippable formats are capped at 15 or 30 seconds depending on platform.
Disclosure of Continuing Vernacular Evolution
The contemporary direct-marketing operational vernacular evolves on a continuous basis as platform-substrate conventions, attribution-disclosure frameworks, and regulatory-compliance regimes are subjected to periodic adjudication. The Glossary is maintained by RandallCloud Marketing Operations Practice on a best-effort cadence; individual entries reflect industry conventions current as of the most recent editorial review. Users requiring engagement-grade vernacular reconciliation are directed to contact RandallCloud Commercial Sales for a bespoke advisory engagement.