The Digital Marketing & AI Glossary (2026): 421 Terms Organized by Topic

Marketing

Understanding people's needs, creating relevant value for them, and delivering it in a way that achieves business goals.

Example: A bakery chooses its products, prices, and sales channels based on its customers.

Marketing strategy

Connected choices that define who we serve, what value we offer, and how we will compete and achieve our goals.

Example: A store serves newlyweds with ready-made bedroom bundles.

Marketing plan

Turning strategy into specific activities, timelines, responsibilities, budgets, and metrics.

Example: A three-month plan for launching a store.

Tactic

A specific action taken to execute part of a strategy.

Example: A video ad that explains how to use the product.

Business model

How a business creates value, delivers it, and generates revenue from it.

Example: A monthly subscription for coffee delivery.

Product-market fit (PMF) (PMF)

Real, sustained demand for a product that solves an important problem for a specific audience.

Example: Customers make repeat purchases and recommend the service to others.

Go-to-market (GTM) (GTM)

Coordinating the audience, offer, pricing, sales process, and channels to launch a product or enter a market.

Example: Launching an accounting service tailored for restaurants.

Marketing mix (4Ps) (4Ps)

Product, price, place, and promotion decisions working together to deliver the offer.

Example: A small pack at an accessible price sold through neighborhood grocery stores.

Business-to-business (B2B) (B2B)

Selling products or services to organizations that use them in their operations.

Example: A company markets clinic management software to medical practices.

Business-to-consumer (B2C) (B2C)

Selling products or services directly to individuals for personal use.

Example: A store sells clothing to families.

Direct-to-consumer (DTC / D2C) (DTC / D2C)

A brand selling its products directly to consumers without relying primarily on independent retail distributors.

Example: A furniture manufacturer sells directly through its own online store.

Channel

A medium used to reach people, communicate with them, or complete a sale.

Example: Search engines, Instagram, or email.

Owned / paid / earned media

Grouping exposure into assets you control, placements you buy, and coverage or mentions others give you.

Example: Your website, your paid ad, and a customer recommendation.

Organic marketing

Attracting attention without paying directly for each impression or click.

Example: An article that reaches people through search engines.

Performance marketing

Managing marketing around measurable outcomes such as leads, sales, and acquisition cost.

Example: Optimizing campaigns based on the cost per qualified lead.

Brand marketing

Activities that build brand recognition, associations, and preference among the audience.

Example: A campaign introducing people to what the brand stands for.

Demand generation / capture

Demand generation builds awareness and interest in a need or solution; demand capture converts people already searching for it.

Example: An educational video followed by a search ad when the buyer looks for a solution.

Omnichannel

Connecting channels so the customer keeps their context when moving from one touchpoint to another.

Example: A customer starts an order on the website and completes or follows up on it in the store branch.

Market research

Gathering information about customers, competitors, and demand to support better business decisions.

Example: Conducting interviews before launching a new service.

Primary / secondary research

Primary research collects new data directly; secondary research uses existing published reports and data.

Example: Interviewing customers and reviewing an industry report.

Qualitative / quantitative research

Qualitative research explores reasons and meanings; quantitative research measures numbers and patterns.

Example: Interviews to understand why an issue happens and a survey to measure how common it is.

Segmentation

Dividing customers into groups that differ in needs or behaviors that matter for marketing.

Example: Gift buyers versus daily-use buyers.

Target audience

The specific group of people we choose to direct an offer or message toward.

Example: Store owners who want to increase their sales.

Ideal customer profile (ICP) (ICP)

A description of the customer or organization whose needs, budget, and circumstances best fit the offer.

Example: A store with a proven product and the capacity to fulfill orders reliably.

Buyer persona

A practical summary of a typical buyer's motivations, pain points, objections, and decision-making process.

Example: A store owner who worries about wasting their ad budget.

Jobs to be done (JTBD) (JTBD)

Understanding the progress a person is trying to make in a specific situation when they choose a solution.

Example: Buying a ready-made meal to save time after work.

Pain point

An obstacle, cost, or frustration that pushes a customer to look for a solution.

Example: Slow responses to customer inquiries.

Motivation

The functional, psychological, or social reason that makes an outcome matter to the customer.

Example: A business owner wants to feel in control of their operations.

Objection

A concern or reason that causes a customer to hesitate or delay a buying decision.

Example: Will this product fit the dimensions of my room?

Voice of customer (VoC) (VoC)

The actual words customers use, revealing what matters to them and how they describe their problems.

Example: Customers repeatedly saying, "I do not know where to start."

Customer insight

An evidence-backed interpretation that reveals an underlying reason driving customer behavior.

Example: People delay buying not because of the price, but because installation feels unclear.

Customer interview

A structured conversation focused on understanding past experiences and decisions rather than asking for hypothetical opinions.

Example: Ask the customer about the last time they made a purchase.

Survey

Standardized questions used to collect comparable answers from a group of people.

Example: Asking buyers why they chose the store.

Competitor analysis

Comparing competitors' offers, messaging, pricing, and customer experiences to spot gaps and opportunities.

Example: Comparing five competing furniture stores.

Ad library / swipe file

A categorized collection of ads and examples saved to study creative angles and formats as a reference.

Example: Tagging an ad by its hook style and funnel stage.

Social listening

Monitoring what people say online about a topic, category, or brand to understand problems and trends.

Example: Reading recurring questions in post comments.

Benchmark

A reference metric or standard used to compare performance while accounting for context differences.

Example: Comparing cost per acquisition across two similar periods.

Brand

The set of meanings, expectations, and impressions people associate with a business or product.

Example: A store name that people associate with reliable service.

Branding

Shaping a consistent experience, identity, and message to build a clear impression in the audience's mind.

Example: Aligning the brand promise, customer service style, and visual design.

Brand strategy

Choices that define the brand's audience, positioning, promise, personality, and consistent expression.

Example: A brand focused on practical everyday comfort for families.

Brand positioning

The intended place a brand occupies in the customer's mind relative to alternatives.

Example: Easy-care, comfortable furniture built for small apartments.

Differentiation

A meaningful difference that customers value enough to justify choosing your offer over alternatives.

Example: Faster installation backed by a clear warranty.

Value proposition

A clear statement of who you serve, the main benefit you deliver, and why your offer is worth choosing.

Example: A complete, coordinated bedroom setup ordered in a single purchase.

Unique selling proposition (USP) (USP)

A specific, distinct reason to buy from you rather than relevant competitors.

Example: Custom dimensions delivered within a guaranteed timeframe.

Brand promise

The core expectation a brand commits to delivering across the customer experience.

Example: A clear, predictable buying process from order to delivery.

Reason to believe (RTB) (RTB)

Concrete proof or evidence that makes a brand's claim credible and believable.

Example: Showing a documented durability test for the fabric.

Brand purpose / values

Purpose explains why the brand exists; values are the principles that guide its decisions and behavior.

Example: Refusing to use a misleading claim even if it increases clicks.

Brand personality

Human traits that guide how a brand looks, sounds, and interacts with people.

Example: Clear, calm, and practical.

Brand voice / tone

Voice is the brand's consistent expression style, while tone adapts to the specific situation and emotion.

Example: Keeping a warm voice while using a serious, respectful tone when handling a complaint.

Messaging framework

A structured map of the core promise, key messages, and proof points organized by audience and need.

Example: One core message supported by three key benefits and proof for each benefit.

Brand story

A connected narrative that explains the problem, vision, and value in a way that matters to the audience.

Example: Why the founder started solving the custom sizing problem.

Visual identity

The system of visual elements that help people recognize and distinguish the brand.

Example: Coordinated colors, fonts, imagery, and logo usage.

Logo / logotype

A visual mark or stylized lettering of the business name used for identification.

Example: The store's icon mark and custom wordmark.

Brand guidelines

Rules and examples that keep language, design, and execution consistent across touchpoints.

Example: Instructions on when to use each color and how to write headlines.

Design system / tokens

Reusable components, rules, and shared values used to build consistent interfaces and designs.

Example: Button colors and spacing saved as shared design tokens.

Distinctive brand assets

Sensory or visual cues that help people recognize a brand even without seeing its name.

Example: A signature color, packaging shape, or recurring phrase.

Brand awareness / recall

Awareness means recognizing the brand; recall means bringing it to mind when thinking of the category or need.

Example: A person names the store immediately when they need a mattress.

Brand equity

The added commercial value that comes from people's familiarity, expectations, and trust in a brand.

Example: Preferring a trusted brand over similar alternatives.

Mental / physical availability

Mental availability is how easily the brand comes to mind in buying situations; physical availability is how easy it is to find and buy.

Example: Remembering the store and then easily finding the right size in stock.

Category entry points (CEPs) (CEPs)

Situations, triggers, or needs that prompt people to think about buying a product category.

Example: Getting married, moving into a new home, or renovating a bedroom.

Rebranding

A major change in a brand's positioning, identity, or market presentation.

Example: Expanding a brand from a single product to complete room solutions.

Offer

The complete package of product, price, terms, support, and bonuses that a customer evaluates.

Example: A mattress packaged with installation, warranty, and return terms.

Feature / benefit

A feature describes what the product has or does; a benefit explains how that improves the customer's life.

Example: Washable fabric (feature) means easier everyday cleaning (benefit).

Customer journey / touchpoint

The customer journey is the full sequence of experiences before and after buying; a touchpoint is any single interaction with the brand.

Example: Watching a video, asking a question on chat, and then completing a purchase.

Marketing funnel

A model that organizes how people move from discovering a solution to evaluating it and taking action.

Example: Watching content, visiting the product page, and placing an order.

Top of funnel (TOF) (TOF)

The stage focused on introducing a relevant audience to the problem, solution, or brand.

Example: A video explaining the signs of sleeping on the wrong mattress.

Middle of funnel (MOF) (MOF)

The stage focused on helping interested prospects understand differences and evaluate the solution.

Example: Comparing materials and usage methods.

Bottom of funnel (BOF) (BOF)

The stage focused on removing final obstacles for people who are close to buying.

Example: Clarifying delivery times, return policies, and payment options.

Awareness levels

Classifying how much a prospect knows about their problem, available solutions, and your specific offer so you can tailor the message.

Example: A person knows their back hurts in the morning but does not know which mattress type solves it.

Lead

A person or organization that has shown interest and can be followed up with based on permission and context.

Example: Someone who submitted their phone number to request information.

Lead generation

Attracting interested people and turning their interest into contact details or conversations for follow-up.

Example: A form to request a consultation.

Lead magnet

A useful resource or benefit offered to encourage someone to share their contact details with clear permission.

Example: A size selection guide offered in exchange for an email address.

Lead qualification

Checking how well a prospect fits the offer based on their need, budget, authority, and timing.

Example: Confirming the prospect's city, business type, and budget.

Conversion

Completing the specific action chosen to measure success in a given context.

Example: Completing a purchase, booking an appointment, or submitting a form.

Micro / macro conversion

A micro conversion is a supporting step along the way; a macro conversion is the primary outcome tied to business goals.

Example: Adding an item to the cart versus completing a purchase.

Call to action (CTA) (CTA)

A phrase or button that clearly tells the person what next step to take.

Example: "Choose your size" or "Book your appointment."

Social proof

Evidence from other people's experiences or actions that helps a buyer gauge trust.

Example: Verified customer reviews or documented case studies.

Risk reversal

Terms or guarantees that reduce the buyer's fear of losing money or time if the choice does not work out.

Example: A clear, straightforward return policy.

Scarcity / urgency

Scarcity relates to limited supply or capacity; urgency relates to a time-bound reason to act now.

Example: A genuinely limited stock batch or a real deadline.

Bundle

Combining complementary products or services into a single package offer.

Example: Selling a pillowcase, fitted sheet, and duvet cover together as a set.

Upsell / cross-sell

An upsell suggests a higher-tier version of the chosen item; a cross-sell suggests a related add-on item.

Example: Suggesting a thicker mattress model (upsell) or adding a matching pillow (cross-sell).

Content strategy

Deciding the role of content, who it serves, what topics it covers, and how it will be distributed and measured.

Example: Educating buyers on how to choose the right product before asking them to order.

Content pillar

A recurring core theme that organizes content ideas around the brand's expertise and audience needs.

Example: Fabric care as a recurring theme for a home furnishings store.

Content calendar

A schedule that outlines what will be published, when, on which channel, and who owns each piece.

Example: A weekly schedule covering videos, social posts, and emails.

Copywriting

Writing designed to help the right audience understand value and take a specific action.

Example: An ad headline that connects a customer problem directly to the solution.

Content writing

Creating text that delivers useful knowledge, explanations, or guidance to an audience.

Example: An article explaining how to measure and choose the right size.

Hook

An opening that earns the right audience's attention and gives them a reason to keep watching or reading.

Example: "Before you buy a mattress, check this measurement."

Angle

A specific lens or entry point used to frame a problem or benefit for a particular audience.

Example: Focusing on saving time matching items rather than focusing on luxury.

Concept

The central creative idea that connects the message to how it is executed visually or narratively.

Example: A side-by-side split-screen comparison showing two ways to use the product.

Headline / primary text

The headline highlights a concise main point, while the primary text explains or supports the message.

Example: A headline about sleep comfort paired with body text explaining the material.

Script

The written sequence of spoken lines, scenes, and visual cues used to produce a video or audio piece.

Example: An opening hook followed by the problem, a product demo, and a call to action.

Storyboard

A frame-by-frame visual outline showing shot composition and the flow of a video before filming.

Example: Six sketched frames mapping out the shots of a short ad.

Attention-interest-desire-action (AIDA) (AIDA)

A classic copywriting framework that orders a message from grabbing attention to building interest and desire, then asking for action.

Example: A hook, followed by a key benefit, proof, and a call to action.

PAS (Problem-Agitate-Solve) (Problem-Agitate-Solve)

A framework that states a specific problem, highlights its real-life consequences, and presents the solution.

Example: Slow replies lose orders; a structured follow-up system fixes it.

Storytelling

Conveying an idea through a relatable situation, character, challenge, and meaningful change.

Example: A story about choosing the wrong size at first and then learning a better way to measure.

Proof / substantiation

Verifiable data, demonstrations, or documentation used to back up a promise or comparison.

Example: Showing a documented measurement instead of claiming to be "the absolute best."

Advertorial

Promotional content written in the style of an educational article or story, with clear disclosure that it is an ad.

Example: A sponsored article explaining a common sleep issue and how a specific mattress addresses it.

Content repurposing

Turning a core piece of content into multiple formats tailored to different channels and contexts.

Example: Turning a long lesson into short video clips and a downloadable checklist.

Localization / transcreation

Adapting language, examples, cultural references, and emotional meaning to a local market rather than translating word for word.

Example: Explaining a marketing concept using an example from a local store.

User-generated content (UGC) (UGC)

Content created by customers or users about an experience, topic, or product.

Example: A customer filming how they use their new purchase at home.

Creative brief

A short document defining the goal, target audience, core message, proof points, deliverables, and constraints for a creative project.

Example: A brief for a 30-second product comparison video.

Art direction

Defining the visual logic, mood, and style that best support the message.

Example: Using warm, natural home photography to convey everyday comfort.

Visual hierarchy

Arranging visual elements by size, contrast, and placement so the viewer sees the most important point first, then supporting details.

Example: Guiding the eye from the headline to the product and then to the order button.

Typography

Choosing fonts, sizes, weights, and spacing to support readability and brand identity.

Example: Selecting a clean, legible Arabic typeface for headlines and body copy.

Color palette / contrast

A defined set of brand colors combined with sufficient visual contrast to separate elements and keep text readable.

Example: Dark text placed on a light background.

Layout / composition

Organizing space, alignment, and elements within a graphic design or video frame.

Example: Placing the product on one side with clean space for the headline on the other.

Aspect ratio

The proportional relationship between the width and height of an image or video.

Example: 9:16 for vertical stories and reels, and 1:1 for a square feed post.

Formula: Aspect Ratio = Width ÷ Greatest Common Divisor (W, H) : Height ÷ Greatest Common Divisor (W, H) | Worked Example: A vertical Reel or TikTok video measuring 1080px wide by 1920px high has a greatest common divisor of 120: (1080 ÷ 120) : (1920 ÷ 120) = 9:16.

Resolution

The number of pixels that make up the width and height of an image or video.

Example: 1080 by 1920 pixels.

Formula: Total Pixel Count = Horizontal Pixels (Width) × Vertical Pixels (Height) | Worked Example: A Full HD vertical ad at 1080 × 1920 contains 1080 × 1920 = 2,073,600 pixels (approx. 2.07 megapixels), whereas a 720 × 1280 export contains 921,600 pixels (both share a 9:16 aspect ratio).

Safe zone

The inner area of a frame where text and important visuals remain visible without being covered by platform buttons or cropped at the edges.

Example: Placing subtitles high enough so they are not hidden behind app interface icons.

Negative space

Intentional empty space around elements that gives the layout breathing room and directs focus.

Example: Leaving a calm, uncluttered background beside the product.

Key visual

The master visual composition that sets the look and feel from which all campaign formats are adapted.

Example: A hero product scene adapted consistently across banners, social posts, and landing pages.

Creative asset / variation

A creative asset is a production file or piece; a variation is an adapted version that changes a specific element such as the hook, headline, or format.

Example: One core video edited with three different opening hooks.

Creative diversity

Testing genuinely different messaging angles, formats, visual styles, and customer situations across your ads.

Example: Running a product demo video, a comparison graphic, and an objection-handling story.

Iteration

Refining an existing creative piece based on performance data or audience feedback to improve results.

Example: Making the proof scene clearer after noticing viewers dropped off at that second.

Creative fatigue

A decline in ad performance as the target audience sees the same creative too many times and stops responding.

Example: Cost per purchase rises as frequency climbs while other conditions remain unchanged.

Pattern interrupt

An unexpected visual or audio cue that breaks the viewer's scrolling habit and supports the message.

Example: Opening a video immediately with a striking side-by-side comparison shot.

Product demonstration

Showing how the product works or the result it creates in action rather than just talking about it.

Example: Filming how the mattress foam responds under pressure.

B-roll / shot list

B-roll is supplemental footage shown while the main voiceover plays; a shot list is the planned checklist of clips to film.

Example: Showing close-up shots of fabric texture and packaging while explaining material quality.

Captions / subtitles

On-screen text that lets viewers follow the spoken dialogue or meaning visually, even with the sound off.

Example: Clear Arabic text displayed in sync with the spoken video.

Organic reach

The number of unique accounts or people who saw a piece of content without paid promotion.

Example: A post reaching both followers and non-followers through shares and platform recommendations.

Engagement

Audience actions on content such as comments, saves, shares, likes, and clicks, depending on how the platform defines them.

Example: Saving a buying guide post to reference later.

Social engagement rate

Total engagements divided by a chosen base metric such as reach, impressions, or follower count.

Example: 100 engagements divided by 2,000 reach equals a 5% engagement rate.

Formula: Social Engagement Rate (%) = (Total Engagements ÷ Chosen Denominator [Reach, Impressions, or Followers]) × 100 | Worked Example: A post reaches 2,000 unique accounts and earns 45 likes, 15 comments, 12 shares, and 28 saves (100 total engagements): (100 ÷ 2,000) × 100 = 5.0% Engagement Rate by Reach.

Watch time / retention

Watch time is the duration viewers spend watching; retention shows what percentage of viewers stay through each part of the video.

Example: Losing half of the viewers before the product is even shown.

Formula: Average Watch Time = Total Watch Time ÷ Total Video Plays | Retention at Second t (%) = (Viewers at Second t ÷ Initial Plays) × 100 | Worked Example: A 30-second video receives 1,000 plays and accumulates 9,500 seconds of watch time: Average Watch Time = 9,500 ÷ 1,000 = 9.5 seconds. If 480 viewers remain at the 5-second mark, retention at 5s is (480 ÷ 1,000) × 100 = 48%.

Completion rate

The percentage of video plays that reach the end, based on the platform's definition of a view.

Example: 300 full plays out of 1,000 video starts equals a 30% completion rate.

Formula: Completion Rate (%) = (Completed Views to End ÷ Total Video Starts) × 100 | Worked Example: Out of 1,000 video starts on a 20-second product demo, 300 viewers watch all the way to the end: (300 ÷ 1,000) × 100 = 30.0% Completion Rate.

Hook rate / hold rate

Practical metrics used to evaluate how well a video stops the scroll and keeps viewers watching, calculated using specific time thresholds and denominators defined in your reporting.

Example: Three-second video views divided by total impressions.

Formula: Hook Rate (%) = (3-Second Video Views ÷ Impressions) × 100 | Hold Rate (%) = (ThruPlays [15s+] ÷ 3-Second Video Views) × 100 | Worked Example: An ad logs 10,000 impressions, 3,200 three-second views, and 960 ThruPlays (15s+): Hook Rate = (3,200 ÷ 10,000) × 100 = 32.0%; Hold Rate = (960 ÷ 3,200) × 100 = 30.0%.

Content distribution

Getting content in front of the target audience through owned channels, social platforms, partnerships, and republishing.

Example: Publishing a lesson and then sharing it across the community group and email list.

Community management

Facilitating discussions, moderating replies, encouraging participation, and building relationships around a shared interest.

Example: Hosting a weekly discussion thread on applying a marketing lesson.

Influencer / creator marketing

Partnering with individuals who have audience trust or strong content creation skills to communicate a brand message.

Example: A creator testing and reviewing a product in a disclosed partnership.

Partnership ads / Spark Ads

Platform ad formats that let a brand run paid ads through or linked to a creator's handle and post with their authorized permission.

Example: Putting paid ad spend behind a creator's video with their authorization.

Affiliate marketing

Rewarding a partner with a commission for driving a qualified action, such as a sale or lead, tracked through agreed links or codes.

Example: Paying a commission on confirmed orders placed through a partner's tracking link.

Social commerce

Using social media platforms to discover, interact with, and purchase products directly.

Example: Browsing a product catalog or storefront inside a social app.

Social search

Using a social platform's search bar to find product reviews, real experiences, or how-to explanations.

Example: Searching TikTok or Instagram for "how to arrange a small bedroom."

Share of voice (SOV) (SOV)

A brand's share of total impressions, ad spend, search visibility, or mentions within a defined competitive set.

Example: The percentage of category mentions a brand receives compared to four named competitors.

Formula: Share of Voice (SOV %) = (Your Brand Metric [Mentions, Impressions, or Spend] ÷ Total Category Metric Across Defined Competitors) × 100 | Worked Example: Across five competing furniture stores with 40,000 total category mentions in a month, your brand accounts for 6,000 mentions: (6,000 ÷ 40,000) × 100 = 15.0% Share of Voice.

Paid media / PPC

Paid media means buying ad distribution; pay-per-click (PPC) is a billing model where you pay for each click.

Example: Search ads billed per click.

Campaign / ad set / ad group

Account hierarchy levels used to organize objectives, budgets, targeting, or keywords, and the ads themselves.

Example: Structuring an account into a campaign, then an ad set or ad group, and then individual ads.

Objective / optimization event

The campaign objective sets the overall outcome goal, while the optimization event tells the ad algorithm which specific action to maximize.

Example: Selecting the Sales objective and optimizing for the Purchase event.

Ad auction

The automated system that decides which ads appear and in what order by combining the advertiser's bid with ad quality and estimated action rates.

Example: Multiple advertisers competing for an impression when a user opens their feed or runs a search.

Placement

The specific location where an ad appears within a platform or ad network.

Example: Stories, main feed, Reels, or search results.

Impressions

The total number of times an ad or piece of content is displayed, including multiple views by the same person.

Example: One person seeing the same ad three times counts as three impressions.

Reach

The number of unique people or accounts exposed to the content within a given reporting period.

Example: 100 distinct people saw the ad.

Frequency

Impressions divided by reach within the same period and scope.

Example: 3,000 impressions divided by 1,000 reach equals an average frequency of 3.

Formula: Frequency = Total Impressions ÷ Unique Reach | Worked Example: A campaign serves 3,000 total impressions to 1,000 unique people over 7 days: Frequency = 3,000 ÷ 1,000 = 3.0 average views per person.

Click-through rate (CTR) (CTR)

The number of specified clicks divided by impressions, multiplied by 100.

Example: 20 clicks from 1,000 impressions equals a 2% CTR.

Formula: Click-Through Rate (CTR %) = (Total Clicks ÷ Total Impressions) × 100 | Worked Example: An ad receives 20 link clicks from 1,000 impressions: CTR = (20 ÷ 1,000) × 100 = 2.0%.

Cost per click (CPC) (CPC)

Ad spend divided by the number of specified clicks.

Example: $100 divided by 200 clicks equals $0.50 per click.

Formula: Cost per Click (CPC) = Total Ad Spend ÷ Total Clicks | Also: CPC = (CPM ÷ 1,000) ÷ CTR (decimal) | Worked Example: Spending $100 to generate 200 link clicks yields: CPC = $100 ÷ 200 = $0.50 per click.

Cost per mille (CPM) (CPM)

Ad spend divided by impressions, then multiplied by 1,000.

Example: $50 spent for 10,000 impressions equals a $5 CPM.

Formula: Cost per Mille (CPM) = (Total Ad Spend ÷ Total Impressions) × 1,000 | Worked Example: Spending $50 on a campaign that delivers 10,000 impressions results in: CPM = ($50 ÷ 10,000) × 1,000 = $5.00 per 1,000 impressions.

Cost per lead (CPL) (CPL)

Specified ad spend divided by the number of leads generated.

Example: $200 divided by 20 inquiries equals $10 per lead.

Formula: Cost per Lead (CPL) = Total Campaign Spend ÷ Total Leads | Cost per Qualified Lead (CPQL) = Total Spend ÷ Qualified Leads | Worked Example: A campaign spends $200 and captures 20 form leads: CPL = $200 ÷ 20 = $10.00. If only 8 of those leads meet qualification criteria, CPQL = $200 ÷ 8 = $25.00.

Cost per action / acquisition (CPA) (CPA)

Total cost divided by the conversions specified in the report.

Example: $300 divided by 10 purchases equals a $30 CPA.

Formula: Cost per Acquisition / Action (CPA) = Total Channel Spend ÷ Attributed Conversions | Worked Example: Spending $300 on a campaign that generates 10 completed purchases yields: CPA = $300 ÷ 10 = $30.00 per purchase.

Cost per view (CPV) (CPV)

Ad spend divided by qualified video views as defined by the platform.

Example: Comparing the cost per qualified view between two video ads.

Formula: Cost per View (CPV) = Total Video Ad Spend ÷ Qualified Video Views | Worked Example: A video campaign spends $120 and generates 4,000 qualified ThruPlay views: CPV = $120 ÷ 4,000 = $0.03 per view.

Conversion rate (CVR) (CVR)

The number of conversions divided by specified conversion opportunities, multiplied by 100.

Example: 20 orders divided by 1,000 sessions equals a 2% conversion rate.

Formula: Conversion Rate (CVR %) = (Total Conversions ÷ Total Sessions [or Clicks / Users]) × 100 | Worked Example: An online store records 20 completed orders from 1,000 sessions: CVR = (20 ÷ 1,000) × 100 = 2.0%.

Return on ad spend (ROAS) (ROAS)

Attributed ad revenue divided by ad spend.

Example: $3,000 in revenue divided by $1,000 in ad spend equals 3x ROAS.

Formula: Return on Ad Spend (ROAS) = Attributed Ad Revenue ÷ Total Ad Spend | Worked Example: A campaign generates $3,000 in attributed order revenue from $1,000 in ad spend: ROAS = $3,000 ÷ $1,000 = 3.0x (or 300%).

Budget / bid strategy

The budget sets spending limits, while the bid strategy determines how the system competes in auctions to reach the goal.

Example: Setting a daily budget paired with conversion optimization bidding.

Pacing

Tracking the speed of ad spend relative to time and the planned budget.

Example: Spending half of the monthly budget halfway through the month.

Formula: Budget Pacing (%) = ((Spend to Date ÷ Total Monthly Budget) ÷ (Days Elapsed ÷ Total Days in Month)) × 100 | Worked Example: On day 15 of a 30-day month (50% time elapsed), a team has spent $6,000 of a $10,000 budget (60% spent): Pacing = (60% ÷ 50%) × 100 = 120% (overspending pace by 20%).

Prospecting

Directing ad delivery to discover people who are not yet existing customers or known warm leads.

Example: Running a campaign designed to reach new buyers.

Retargeting / remarketing

Re-engaging people who have a known prior interaction with the brand.

Example: Showing ads to people who visited a product page.

Audience exclusion

Preventing a specific segment from receiving ads when platform settings allow it.

Example: Excluding recent buyers from a new-customer acquisition offer.

Scaling

Increasing spend or reach while monitoring result quality and profitability.

Example: Gradually raising ad spend while checking incremental profit.

Meta Ads Manager

The interface for creating, managing, and measuring Meta ads.

Example: Setting up a campaign across Facebook and Instagram.

Meta Business Portfolio

A workspace for organizing business assets, people, and access permissions.

Example: Managing pages, ad accounts, and product catalogs in one place.

Meta Pixel

A browser-based tool that sends website activity events to help with measurement and ad delivery.

Example: Recording a product view or purchase event.

Conversions API (CAPI) (CAPI)

A connection that sends event data from business systems to Meta to support measurement and optimization.

Example: Sending a confirmed purchase event from the server.

Standard / custom event

A standard event is a predefined action recognized by the platform, while a custom event is defined for a specific business need.

Example: Using the standard Purchase event versus a custom quote-calculator event.

Custom audience

An audience built from a source such as platform engagement, website visits, or an eligible customer list.

Example: Creating an audience of people who watched a video or visited the website.

Lookalike audience

An audience identified by the platform based on shared characteristics with a source audience.

Example: Building a lookalike audience from high-value buyers.

Broad targeting

Minimizing manual audience restrictions and giving the delivery system room to find suitable people.

Example: Setting country and age range without narrow interest lists.

Advantage+ (Advantage+)

An umbrella for Meta automation features across audiences, placements, creative, and campaigns.

Example: Automated delivery across eligible placements.

Advantage+ audience

An audience setting that uses signals and suggestions alongside specific controls to expand discovery.

Example: Providing an audience suggestion while applying available account controls.

Advantage+ placements

Allowing the system to allocate ads across eligible placements to achieve the campaign goal.

Example: Delivering ads across both Stories and Feed placements.

Campaign budget / ad set budget (CBO / ABO) (CBO / ABO)

Campaign budget distributes spend across ad sets, while ad set budget assigns a separate budget to each ad set.

Example: Using a shared campaign budget or setting individual test budgets per ad set.

Learning phase

A period when the delivery system explores patterns to optimize ad set performance.

Example: Seeing fluctuating results shortly after launching an ad set.

Learning limited

A status indicating that the ad set has not gathered enough optimization events for the system to exit learning.

Example: Generating too few purchases relative to how the budget is split.

Instant form / click-to-message

An instant form collects lead details inside the platform, while a click-to-message ad opens a chat conversation.

Example: Submitting an in-app interest form or starting a WhatsApp chat.

Catalog ads

Ads that pull data from a product feed to automatically show relevant items.

Example: Displaying items directly from the store's product catalog.

Existing post / post ID

Using an existing social post in an ad via its post ID when settings allow.

Example: Running an ad using a post that already has comments and likes.

Andromeda (Andromeda)

A Meta system for retrieving ad candidates within its broader recommendation architecture.

Example: Understanding its role to see how automated ad retrieval works at scale.

Generative Ads Model (GEM) (GEM)

An advertising foundation model developed by Meta to improve ad recommendation systems.

Example: Understanding that ad delivery relies on multiple large recommendation models.

Creative-led targeting

Designing the message and visuals to clearly signal who the offer serves, helping guide delivery through engagement signals.

Example: Calling out a specific store-owner problem in the opening hook.

Search campaign

A campaign that shows ads when user searches match targeting eligibility and campaign settings.

Example: Showing an ad when someone searches to buy bedding.

Keyword / search term

A keyword is the targeting setting chosen by the advertiser, while a search term is the exact query typed by the user.

Example: Targeting the keyword 'mattresses' and matching the search term 'best mattress for back pain'.

Search intent

The underlying goal behind a query, such as learning, comparing options, or buying.

Example: Searching for 'types of mattresses' versus 'buy king mattress online'.

Broad / phrase / exact match

Match types that control how closely a user query must relate to the targeted keyword based on meaning and Google's rules.

Example: Choosing the appropriate match type for a target keyword.

Negative keyword

A keyword setting that excludes unsuitable searches according to its negative match type rules.

Example: Excluding 'jobs' from a commercial service campaign.

Responsive search ad (RSA) (RSA)

A search ad format where you provide multiple headlines and descriptions that the system combines.

Example: Supplying several headlines highlighting different product benefits.

Ad asset / extension

An element that supplements an ad, such as sitelinks, call buttons, or structured information.

Example: Adding links to size guides and shipping details.

Quality Score

A diagnostic score for a keyword based on expected CTR, ad relevance, and landing page experience.

Example: Checking why a keyword has below-average ad relevance.

Formula: Quality Score (1 to 10) = Diagnostic Rating of (Expected CTR + Ad Relevance + Landing Page Experience) | Worked Example: A keyword with Above Average Expected CTR, Above Average Ad Relevance, and Average Landing Page Experience typically scores 8/10 or 9/10 in Google Ads diagnostics.

Ad Rank

A score that determines ad eligibility and placement using multiple factors including bid, ad quality, and context.

Example: A highly relevant ad outranking a competitor with a higher bid.

Formula: Ad Rank = f(Max CPC Bid, Auction-Time Ad & LP Quality, Ad Rank Thresholds, Search Context, Expected Asset Impact) | Worked Example: In a simplified auction model, Advertiser A bidding $1.50 with high auction-time quality (score weight 9 -> rank 13.5) outranks Advertiser B bidding $2.00 with weak relevance (score weight 5 -> rank 10.0).

Impression share

Impressions received divided by the estimated number of eligible impressions.

Example: Appearing in 60% of eligible search opportunities.

Formula: Impression Share (%) = (Actual Impressions Received ÷ Total Estimated Eligible Impressions) × 100 | Worked Example: Your search campaign was eligible for an estimated 10,000 searches this week and appeared 6,000 times: Impression Share = (6,000 ÷ 10,000) × 100 = 60.0% (with 40% lost to budget or rank).

Smart Bidding

Automated bid strategies that use machine learning to optimize for conversions or conversion value.

Example: Optimizing bids around total sales value.

Target CPA (tCPA) (tCPA)

An average cost-per-conversion target used to guide automated bidding where available.

Example: Aiming for an average of $20 per conversion.

Formula: Maximum Profitable tCPA = (Average Order Value × Pre-Ad Contribution Margin %) - Target Net Contribution per Order | Worked Example: With a $100 AOV and 45% pre-ad contribution margin ($45 gross contribution), aiming to keep $25 per order after ads sets tCPA = $45 - $25 = $20.00.

Target ROAS (tROAS) (tROAS)

A target ratio of conversion value to ad spend that guides automated bidding.

Example: A 300% target equals 3x return on ad spend.

Formula: Target ROAS (tROAS %) = (Target Conversion Value ÷ Ad Spend) × 100 | Worked Example: To generate $300 in conversion value for every $100 of ad spend, configure tROAS = ($300 ÷ $100) × 100 = 300%.

Maximize conversions / value

Bid strategies that aim to get the highest conversion count or conversion value within your budget and settings.

Example: Maximizing total bookings or total order value.

Performance Max (PMax) (PMax)

An automated campaign type that runs across Google's ad inventory to achieve conversion goals.

Example: A single campaign serving across Search, YouTube, Display, and Shopping inventory.

Asset group / audience signal

An asset group organizes creative and feed assets in PMax, while an audience signal helps guide initial system learning.

Example: Grouping images and copy for one product line paired with relevant audience signals.

Demand Gen

A visual campaign type across eligible Google surfaces designed to generate interest and conversions.

Example: Running video and image ads to reach prospective buyers across YouTube and Discover.

Google Shopping / product feed

Ads powered by structured product data such as title, price, image, and availability.

Example: A product card appearing in search results with its photo and price.

Google Merchant Center

A platform for managing product feed data and eligibility across Google surfaces.

Example: Uploading a product feed and resolving availability or pricing errors.

YouTube / Display ads

Video or visual banner ads served across eligible inventory outside standard search results.

Example: Running an explainer video on YouTube or a banner ad across partner sites.

Brand / non-brand search

Brand search includes the business or brand name, while non-brand search expresses a need or category without naming a brand.

Example: Searching for the store's name versus searching for 'best cotton bed sheets'.

Offline conversion import

Sending downstream outcomes from a CRM or sales system back to the ad platform to connect them with ad interactions.

Example: Sending a closed-won deal event back to Google Ads after an initial lead form.

Enhanced conversions

Using cryptographically hashed first-party customer data to improve conversion matching in Google.

Example: Sending a hashed customer email address in line with consent and platform policies.

AI Max for Search

A suite of features that expands search matching and adapts creative assets inside Search campaigns.

Example: Expanding query matching and tailoring ad copy dynamically.

Search engine optimization (SEO) (SEO)

Improving a website and its content so search engines can understand and surface it in unpaid search results.

Example: Publishing a helpful page that directly answers what a searcher needs.

Search engine results page (SERP) (SERP)

The page displayed by a search engine showing organic results, ads, and other features in response to a query.

Example: A results page showing links, images, map listings, and direct answers.

On-page / off-page / technical SEO

Three areas of SEO covering page content, external authority signals, and site crawlability and usability.

Example: Improving page copy, earning external links, and fixing crawl errors.

Keyword research

Discovering what your audience searches for and how those queries connect to their needs and intent.

Example: Gathering common questions people ask when choosing bed sizes.

Long-tail keyword

A specific search query that usually has lower individual search volume than broad head terms.

Example: Searching for a cotton duvet cover in a specific mattress size.

Crawling / indexing

Crawling is discovering and fetching pages, while indexing is analyzing and storing them for potential display in search results.

Example: A search engine crawled the page but chose not to index it.

Title tag / meta description

Page metadata elements that describe a page's topic and may be used by search engines in result snippets.

Example: Writing a clear title tag and concise description that match the page content.

Heading structure (H1-H6) (H1-H6)

Ordered headings that organize and show the hierarchy of a page's sections.

Example: Using one main H1 heading followed by H2 and H3 subheadings for sections.

Internal link / backlink

An internal link connects pages on the same website, while a backlink points to your site from another website.

Example: Linking from a size guide article to a relevant product page.

Anchor text

The clickable words in a hyperlink that describe the destination page.

Example: Using 'bedding size guide' instead of 'click here'.

Canonical URL

A signal that specifies the preferred URL version when duplicate or near-identical pages exist.

Example: Pointing multiple filtered URLs of a product page to one main canonical URL.

Redirect / 404

A redirect sends visitors and crawlers to a different URL, while a 404 status indicates that the requested page was not found.

Example: Redirecting a discontinued product URL to its direct replacement.

robots.txt / noindex (robots.txt / noindex)

A robots.txt file guides crawler access, while a noindex directive asks search engines not to index a page when they read it.

Example: Applying noindex to an internal utility page that should not appear in search results.

XML sitemap

A file that lists important website URLs to help search engines discover them.

Example: An XML sitemap containing canonical product and category URLs.

Structured data / schema markup

Standardized code that describes a page's entities and properties for automated systems.

Example: Product markup specifying price, currency, and stock status.

Rich result

An enhanced search listing with extra visual or informational features for eligible pages.

Example: A product search result showing price, availability, and review ratings.

E-E-A-T (E-E-A-T)

A framework for evaluating content quality through Experience, Expertise, Authoritativeness, and Trustworthiness.

Example: A detailed guide written by a practitioner with documented hands-on experience.

Local SEO

Optimizing a business's online presence and information to appear for location-related searches.

Example: Showing up when someone searches for a nearby bakery.

Google Business Profile (GBP) (GBP)

A profile for managing an eligible business's information across Google Search and Maps.

Example: Updating business hours, address, and photos.

Name-address-phone consistency (NAP) (NAP)

Keeping a business's name, address, and phone number consistent across key directories and platforms.

Example: Listing the exact same phone number and address on the website and Google Business Profile.

Google Search Console (GSC) (GSC)

A Google tool for monitoring a website's search visibility, query performance, and indexing issues.

Example: Reviewing which search queries drive clicks and impressions to the site.

Digital public relations (digital PR) (digital PR)

Earning relevant media coverage, relationships, and brand mentions across digital publications.

Example: Publishing original industry research that trade publications cite.

Domain authority / rating (DA / DR) (DA / DR)

Comparative scores created by third-party SEO tools to estimate domain link strength using their own models.

Example: Comparing the relative link strength of two domains inside the same tool.

AI search

Search experiences that use AI models to interpret questions, synthesize information, or generate answers.

Example: Receiving a synthesized answer accompanied by source links.

AI Overviews

AI-generated summaries that appear in some Google search results alongside supporting links.

Example: Seeing an overview summarizing a multi-part question at the top of search results.

AI Mode

A Google search experience designed for exploring complex questions and comparisons with AI answers and links.

Example: Comparing product options and then asking a follow-up question.

Generative engine optimization (GEO) (GEO)

A common term for making brand information clear, credible, and easy for AI answer systems to use and cite.

Example: Publishing accurate product guides backed by evidence and consistent brand facts.

Answer engine optimization (AEO) (AEO)

Structuring content to clearly answer specific questions across direct-answer and AI search experiences.

Example: Adding a section that directly answers a frequent buyer question.

AI citation / brand mention

A citation links or attributes information to a specific source, while a brand mention names the brand with or without a link.

Example: An AI response naming a store and linking to a source article for a product claim.

Query fan-out

Issuing multiple related sub-queries behind the scenes to gather information from different angles of a question.

Example: Breaking a product comparison question into sub-searches for price, materials, and warranty.

Zero-click search

A search session that ends without the user clicking through to an external website.

Example: Reading a direct answer on the search results page without clicking a link.

Formula: Zero-Click Search Rate (%) = (Searches Ending Without an Outbound Website Click ÷ Total Searches) × 100 | Worked Example: If 620 out of 1,000 informational queries are answered directly inside the search interface without an external click: Zero-Click Rate = (620 ÷ 1,000) × 100 = 62.0%.

AI visibility / share of answers

Custom metrics that track how often a brand is mentioned or cited across a defined sample of prompts.

Example: Appearing in 12 out of 50 tested buyer questions.

Formula: AI Share of Answers (%) = (Prompts Mentioning or Citing Your Brand ÷ Total Tested Category Prompts) × 100 | Worked Example: Your brand is cited or recommended in 12 out of 50 standardized category buying questions: AI Share of Answers = (12 ÷ 50) × 100 = 24.0%.

llms.txt (llms.txt)

A proposed text file format that summarizes a website's key links and documentation for language model tools.

Example: Providing a concise markdown index of product documentation.

Agentic commerce

Using AI agents to help discover, compare, or complete shopping steps with user permission.

Example: An AI agent shortlisting products that match a buyer's budget and delivery rules.

Landing page

A page a visitor arrives on, typically built to serve a specific message or conversion action.

Example: A dedicated page designed for ordering a bedding bundle.

Homepage / product detail page (PDP) (PDP)

The homepage introduces the brand and main navigation paths, while a product detail page explains a specific item.

Example: Browsing store categories on the homepage and then checking size and fabric specs on a product page.

Conversion rate optimization (CRO) (CRO)

Improving the offer and buying experience through research and testing to increase valuable conversions.

Example: Clarifying shipping costs and delivery times to reduce buyer hesitation.

Formula: Revenue per Visitor (RPV) = Conversion Rate (CVR) × Average Order Value (AOV) = Total Revenue ÷ Total Visitors | Worked Example: A store with a 2.0% CVR (0.02) and $80 AOV earns RPV = 0.02 × $80 = $1.60 per visitor. Lifting CVR to 2.5% while holding AOV at $80 raises RPV to $2.00 (+25% revenue from the same traffic).

User experience / interface (UX / UI) (UX / UI)

User experience covers how easily someone completes a task, while user interface refers to the visual and interactive elements they use.

Example: Making size selection straightforward with a clear, easy-to-tap button.

Above the fold / hero section

Above the fold is what appears on screen before scrolling, and the hero section presents the primary opening message.

Example: Showing a clear headline, product image, value proposition, and call to action at the top of the page.

Message match

Consistency in promise, offer, and tone between an ad and its destination page.

Example: An ad for a bundle leading directly to a page featuring that exact bundle and its terms.

Friction

Extra effort, confusion, or unnecessary steps that make completing an action harder.

Example: Asking for unnecessary personal details on a lead form.

Trust signal

Verifiable information that reduces doubt about the offer or the business.

Example: Displaying a physical address, clear return policies, and working contact details.

Heatmap / session recording

Tools that visualize aggregate click and scroll patterns or replay how visitors interact during a session.

Example: Spotting that visitors overlook an important button or tap on a non-clickable image.

Mobile-first / responsive design

Mobile-first design starts with the mobile experience, while responsive design adapts layouts across different screen sizes.

Example: Designing buttons and forms that are easy to read and tap on a phone screen.

Accessibility

Making content and interfaces usable for people with diverse abilities and assistive technologies.

Example: Providing readable color contrast, descriptive alt text, and full keyboard navigation.

Core Web Vitals

Real-world user experience metrics measuring loading speed, interactivity, and visual stability.

Example: Monitoring LCP, INP, and CLS scores across mobile and desktop users.

Formula: Core Web Vitals Pass Criteria (75th Percentile of Field Loads): LCP <= 2.5s AND INP <= 200ms AND CLS <= 0.10 | Worked Example: If 78% of mobile page loads render the hero image in <= 2.3s (LCP), respond to taps in <= 140ms (INP), and shift layout by <= 0.04 (CLS), the URL passes all three Core Web Vitals.

Largest Contentful Paint (LCP) (LCP)

The time it takes for the largest visible content element in the viewport to render.

Example: A slow-loading hero image delaying the main visual display.

Formula: LCP (Seconds) = Timestamp(Largest Visible Content Element Rendered) - Timestamp(Navigation Start) [Good <= 2.5s] | Worked Example: A product page with a 0.4s TTFB, 0.3s image discovery delay, 0.9s WebP download, and 0.2s render completes LCP at 0.4 + 0.3 + 0.9 + 0.2 = 1.8 seconds (Good).

Interaction to Next Paint (INP) (INP)

A metric measuring page responsiveness by tracking the time from a user interaction to the next visual update.

Example: A noticeable delay before the screen updates after tapping a button.

Formula: Interaction Latency (INP ms) = Input Delay + Processing Time + Presentation Delay [Good <= 200ms] | Worked Example: When a shopper taps "Select Size", a 25ms main-thread wait + 60ms React state update + 35ms browser paint yields an interaction latency of 25 + 60 + 35 = 120ms (Good).

Cumulative Layout Shift (CLS) (CLS)

A metric that measures unexpected shifts in the position of visible elements while a page loads.

Example: A checkout button shifting downward just as an image finishes loading.

Formula: Layout Shift Score = Impact Fraction × Distance Fraction [Good Cumulative Session Window Score <= 0.10] | Worked Example: An un-dimensioned banner loads late, affecting 50% of the viewport (0.50) and pushing the buy button down by 14% of the screen height (0.14): Shift Score = 0.50 × 0.14 = 0.07.

Checkout / cart abandonment

Checkout is the stage of completing an order, while cart abandonment occurs when items are added to the cart without completing a purchase within the measured window.

Example: A shopper adding a product to the cart and leaving before paying.

Formula: Cart Abandonment Rate (%) = (1 - (Completed Purchases ÷ Shopping Carts Created)) × 100 | Worked Example: Out of 400 shoppers who add an item to their cart, 100 complete payment: Cart Abandonment Rate = (1 - (100 ÷ 400)) × 100 = 75.0%.

Google Analytics 4 (GA4) (GA4)

A measurement system that tracks website and app usage through events and reporting.

Example: Tracking session sources and purchase behavior across a store.

User / session / page view

A user is a measured visitor identity, a session is a period of interaction, and a page view is a single page load that can repeat.

Example: One user returning for a new session and viewing three pages.

Event / event parameter

An event records a specific action, while event parameters provide additional details about that action.

Example: A 'purchase' event accompanied by order value and currency parameters.

GA4 key event

An event marked in GA4 as an action important to business success.

Example: Marking a qualified inquiry submission as a key event.

Google Ads conversion action

A defined customer action measured as a conversion in Google Ads, which can be used for bidding depending on goal settings.

Example: Tracking a completed purchase or confirmed booking as a primary conversion action.

Engaged session

A GA4 session that lasts longer than the configured duration threshold, includes a key event, or has at least two page or screen views.

Example: A short session that includes a purchase counts as an engaged session.

Formula: Engaged Session = (Session Duration >= 10 Seconds) OR (Key Events >= 1) OR (Page / Screen Views >= 2) | Worked Example: A visitor lands on a page, clicks one internal link to a second page within 6 seconds, and leaves: because Page Views = 2, GA4 classifies the session as Engaged.

GA4 engagement / bounce rate

Engagement rate is engaged sessions divided by total sessions, and bounce rate in GA4 is its inverse (100% minus engagement rate).

Example: A 70% engagement rate equals a 30% bounce rate.

Formula: GA4 Engagement Rate (%) = (Engaged Sessions ÷ Total Sessions) × 100 | GA4 Bounce Rate (%) = 100% - Engagement Rate (%) | Worked Example: Out of 1,000 total sessions, 700 meet the Engaged Session criteria: Engagement Rate = (700 ÷ 1,000) × 100 = 70.0%, and Bounce Rate = 100% - 70.0% = 30.0%.

Google Tag Manager (GTM) (GTM)

A tag management system for deploying and controlling measurement and marketing tags and their firing rules.

Example: Firing a conversion tag when a user submits a form.

Data layer

A structured JavaScript object on a website that passes reliable data to tag management and analytics tools.

Example: Passing order value and transaction ID into the data layer on the confirmation page.

Tracking tag / pixel

A snippet of code or network request that sends measurement signals when specified conditions occur.

Example: Recording a page visit or purchase event from the browser.

Urchin tracking module parameters (UTM) (UTM)

URL parameters added to links to identify the traffic source, medium, campaign, and creative content.

Example: Appending utm_source, utm_medium, and utm_campaign to an ad destination URL.

Source / medium / channel

Source identifies where traffic came from, medium describes the type of referral, and channel groups similar traffic sources together.

Example: Recording 'google' as source, 'cpc' as medium, and 'Paid Search' as the channel.

Attribution

Rules that assign credit for a conversion to known marketing touchpoints.

Example: Giving an ad credit for an order placed after a click.

Attribution window

The lookback period after an ad interaction during which a conversion can be credited to that interaction.

Example: Counting a purchase that happens three days after an ad click.

Click-through / view-through conversion

A click-through conversion is credited after a click, while a view-through conversion is credited after an impression without a qualifying click.

Example: Someone seeing an ad without clicking it and completing a purchase later.

Last-click / data-driven attribution

Last-click gives all credit to the final qualifying click, while data-driven attribution distributes credit across touchpoints using a statistical model.

Example: Comparing how credit shifts across channels under both attribution models.

Server-side tracking

Processing or sending measurement events through a server rather than relying on the browser alone.

Example: Sending a purchase event directly from the order management system.

Event deduplication

Identifying duplicate instances of the same conversion event so it is counted only once.

Example: Matching a browser event and a server event using a shared event ID.

First-party / third-party data

First-party data is collected directly from your relationship with your audience, while third-party data comes from an external source without that direct relationship.

Example: Order history from your own store versus audience segments bought from an external broker.

Zero-party data

A term for data that a customer intentionally shares about their preferences, needs, or intentions.

Example: A shopper selecting their skin type in a product quiz.

Cookie

A small piece of data stored by a web browser that can be used to remember state or an identifier.

Example: Keeping items saved in a shopping cart across page loads.

Consent / consent management platform (CMP) (CMP)

Consent is the user's permission choice, while a CMP presents the notice, records the choice, and passes consent signals to tags.

Example: Allowing visitors to accept or reject analytics and advertising cookies.

Consent Mode

A mechanism that adjusts how Google tags behave based on user consent status signals.

Example: Updating ad and analytics storage permissions after a user interacts with a consent banner.

Personally identifiable information (PII) (PII)

Data that directly identifies an individual or can link activity to a specific person depending on context.

Example: An email address, phone number, or linkable personal identifier.

Hashing / anonymization

Hashing converts an input value into a fixed cryptographic string, while anonymization aims to prevent data from being linked back to an individual.

Example: Hashing an email address using SHA-256 for consented conversion matching.

Modeled conversions

Statistical estimates generated by a platform to account for conversions that could not be directly observed.

Example: Estimating conversions when cookie restrictions prevent direct link tracking.

Metric / dimension

A metric is a quantitative number, while a dimension is a descriptive attribute used to group or break down metrics.

Example: Revenue is a metric, and city is a dimension.

Key performance indicator (KPI) (KPI)

A selected metric that directly tracks progress toward an important business goal.

Example: Tracking the number of profitable new customers acquired.

Dashboard / reporting

A dashboard displays key metrics visually, while reporting explains the results, their context, and the next action to take.

Example: Sharing weekly performance numbers alongside a summary of what changed and what to adjust.

Leading / lagging indicator

A leading indicator signals likely future performance, while a lagging indicator summarizes an outcome that has already happened.

Example: Qualified demos booked today as a leading indicator, and closed revenue at quarter-end as a lagging indicator.

Vanity metric

A number that looks impressive on the surface but does not show meaningful progress toward a business goal.

Example: Accumulating high video views from an audience that never buys.

Cohort

A group that shares a starting event or period so its behavior can be tracked over time.

Example: Customers who made their first purchase in January.

Retention / churn rate

Retention measures continuing customers; churn measures those who stop within a defined rule and period.

Example: A subscription renewal versus a cancellation.

Formula: Retention Rate (%) = (Active Cohort Customers at End ÷ Cohort Customers at Start) × 100 | Churn Rate (%) = (Lost Cohort Customers ÷ Cohort Customers at Start) × 100 | Worked Example: A subscription service starts the quarter with 500 customers; 425 renew at quarter end and 75 cancel: Retention Rate = (425 ÷ 500) × 100 = 85.0%; Churn Rate = (75 ÷ 500) × 100 = 15.0%.

A/B test / split test

A concurrent comparison with proper random assignment to estimate the effect of a specific difference.

Example: Changing the promise phrasing while keeping the rest of the experience fixed.

Hypothesis

A specific prediction linking a change to a cause and a measurable outcome.

Example: Clarifying delivery terms will increase order completion.

Control / treatment

The control group stays without the tested change; the treatment group receives the change.

Example: Visitors seeing the current page versus the new version.

Sample size / statistical power

Sample size is the number of observations; statistical power is the probability of detecting a specific effect if it exists.

Example: Estimating enough visits to detect a realistic difference.

Formula: Approx. Sample Size per Variant (n) ≈ 16 × [p × (1 - p)] ÷ (MDE_absolute)² [for 80% Power, Alpha = 0.05] | Worked Example: With a 4.0% baseline conversion rate (p = 0.04) and a goal to detect a +0.8 percentage-point lift (to 4.8%, so MDE_absolute = 0.008): n ≈ 16 × (0.04 × 0.96) ÷ (0.008)² = 0.6144 ÷ 0.000064 = 9,600 visitors per variant (19,200 total).

Statistical significance

An assessment of how compatible the data is with a null-effect hypothesis under a defined model and decision threshold.

Example: A test producing a p-value below the chosen threshold.

Formula: Two-Proportion Z-Score = (CVR_Treatment - CVR_Control) ÷ Standard Error [Statistically Significant at 95% when |Z| >= 1.96, p-value < 0.05] | Worked Example: Control converts 200 / 10,000 (2.00%) and Treatment converts 245 / 10,000 (2.45%): the difference is +0.45 percentage points with Standard Error ≈ 0.209%, giving Z = 0.45 ÷ 0.209 = 2.15 (p ≈ 0.031 < 0.05).

Confidence interval

A range describing the uncertainty around an estimate using a specific statistical procedure.

Example: Presenting an estimated effect with a range instead of a single number.

Formula: 95% Confidence Interval = Point Estimate (Observed Lift) ± (1.96 × Standard Error) | Worked Example: An experiment shows a +0.60 percentage-point conversion lift with a Standard Error of 0.20 percentage points: the 95% Confidence Interval is 0.60% ± (1.96 × 0.20%) = [+0.21 pp to +0.99 pp].

Relative lift / percentage points

Relative lift compares the change to the original value; percentage points subtract two percentages directly.

Example: From 2% to 3% is a 1 percentage point increase and a 50% relative lift.

Formula: Percentage Points (pp) = Rate_New - Rate_Old | Relative Lift (%) = ((Rate_New - Rate_Old) ÷ Rate_Old) × 100 | Worked Example: Lifting conversion rate from 2.0% to 3.0% is an absolute increase of 3.0% - 2.0% = +1.0 percentage point (pp), and a relative lift of ((3.0 - 2.0) ÷ 2.0) × 100 = +50.0%.

Incrementality / holdout test

Incrementality is additional causal impact; a holdout test estimates it by comparing an exposed group to an unexposed group.

Example: A randomized experiment measuring additional orders.

Formula: Incremental Revenue = Revenue_Treatment - Revenue_Scaled_Holdout | Incremental ROAS (iROAS) = Incremental Revenue ÷ Ad Spend | Worked Example: A retargeting campaign spends $5,000; the exposed group generates $25,000 while an equivalent randomized holdout group buys $15,000 organically: Incremental Revenue = $25,000 - $15,000 = $10,000, so true iROAS = $10,000 ÷ $5,000 = 2.0x (even if platform attribution claimed 4.0x).

Correlation / causation

Correlation is concurrent change; causation means changing one factor produced a change in another.

Example: Spend and sales rising together is not enough to prove causation.

Selection / survivorship bias

A skewed conclusion caused by who entered the sample or who remained visible in it.

Example: Studying only winning ads and ignoring failed ones.

Marketing mix modeling (MMM) (MMM)

Statistical models using aggregated data to estimate the contributions of marketing and other factors.

Example: Studying spend, sales, and seasonality over time.

Formula: Revenue_t = Baseline_t + Σ [Beta_i × HillSaturation(Adstock(Spend_i,t))] + Seasonality_t + Error_t | Worked Example: If statistical regression shows Baseline = $120k/month and Paid Social contributes $45k on $15k spend after controlling for Eid seasonality, the MMM channel contribution is 3.0x.

Ecommerce

Buying and selling products or services through digital channels.

Example: A store receiving orders and payments online.

Stock keeping unit (SKU) (SKU)

An internal code that identifies an item or a specific combination of its options.

Example: A blue cover in a specific size has its own separate code.

Product variant

A version of a product that differs by an attribute such as size or color.

Example: The same design offered in two colors and three sizes.

Gross / net sales

Gross sales are before discounts and returns; net sales are after the deductions defined in the reporting system.

Example: A sale of 100 with a discount of 10 gives 90 before other deductions.

Formula: Gross Sales = Units Sold × Unit Selling Price | Net Sales = Gross Sales - Discounts - Returns | Worked Example: A store sells 10 items at $100 each ($1,000 Gross Sales), applies $100 in promo codes, and processes one $90 return: Net Sales = $1,000 - $100 - $90 = $810.

Revenue / profit

Revenue is the recognized value of sales; profit is what remains after specified costs.

Example: Sales of 100 do not equal a profit of 100.

Formula: Net Profit = Recognized Net Revenue - COGS - Variable Fulfillment & Ad Costs - Fixed Operating Overhead | Worked Example: On $100 of net revenue with $38 COGS, $12 shipping/payment fees, $25 ad spend, and $15 fixed overhead allocation: Net Profit = $100 - $38 - $12 - $25 - $15 = $10 (10% net margin).

Cost of goods sold (COGS) (COGS)

The cost of purchasing or producing the units sold during the period.

Example: Fabric and manufacturing cost for the sold item.

Formula: COGS = Beginning Inventory + Purchases / Production During Period - Ending Inventory | Worked Example: A store starts the month with $20,000 in inventory, manufactures $15,000 more, and ends the month with $23,000 in unsold stock: COGS = $20,000 + $15,000 - $23,000 = $12,000.

Gross profit / gross margin

Gross profit is net sales minus COGS; gross margin divides gross profit by net sales.

Example: Sales of 100 and cost of 60 give 40 gross profit and a 40% margin.

Formula: Gross Profit ($) = Net Sales - COGS | Gross Margin (%) = ((Net Sales - COGS) ÷ Net Sales) × 100 | Worked Example: An order with $100 in Net Sales and $60 in COGS yields Gross Profit = $100 - $60 = $40, and Gross Margin = ($40 ÷ $100) × 100 = 40.0%.

Contribution margin

Revenue minus specified variable costs; the margin ratio divides that remainder by revenue.

Example: Revenue of 100 minus 65 in variable costs equals 35.

Formula: Contribution Margin ($) = Net Revenue - Variable Costs | Contribution Margin (%) = (Contribution Margin ($) ÷ Net Revenue) × 100 | Worked Example: On $100 Net Revenue with $42 COGS, $10 fulfillment/shipping, and $3 payment fees ($55 total variable non-ad cost): Pre-Ad Contribution Margin = $100 - $55 = $45 (45.0%). If CPA is $20, Post-Ad Contribution = $45 - $20 = $25 (25.0%).

Average order value (AOV) (AOV)

Order sales value as defined by the report divided by the number of orders.

Example: 1,000 in sales divided by 20 orders equals 50.

Formula: Average Order Value (AOV) = Total Qualifying Order Sales ÷ Total Number of Orders | Worked Example: A store generating $1,000 in sales across 20 orders has an AOV = $1,000 ÷ 20 = $50.00 per order.

Customer acquisition cost (CAC) (CAC)

Specified acquisition costs divided by the number of new customers acquired.

Example: 1,000 in acquisition cost divided by 20 new customers equals 50.

Formula: Customer Acquisition Cost (CAC) = Total Acquisition Costs ÷ Total New Customers Acquired | Worked Example: Spending $1,000 on acquisition to win 20 first-time buyers yields CAC = $1,000 ÷ 20 = $50.00 per new customer.

Blended CAC / paid CAC

Blended CAC combines channels within a stated scope; paid CAC isolates the paid acquisition scope.

Example: Total acquisition cost divided by all new customers.

Formula: Blended CAC = Total Marketing & Acquisition Spend ÷ All New Customers | Paid CAC = Paid Media Spend ÷ Paid-Attributed New Customers | Worked Example: A brand spends $10,000 on ads plus $2,000 on creative ($12,000 total) and acquires 300 total new customers (200 via paid ads, 100 organic): Paid CAC (ad spend only) = $10,000 ÷ 200 = $50.00; Fully-loaded Blended CAC = $12,000 ÷ 300 = $40.00.

Customer lifetime value (CLV / LTV) (CLV / LTV)

Realized or projected value from a customer across a specific period or the entire relationship.

Example: Estimating the contribution of a returning customer.

Formula: Contribution LTV = Average Order Value (AOV) × Purchase Frequency × Pre-Ad Contribution Margin (%) × Customer Lifespan | Worked Example: Over 12 months, an average customer places 2.5 orders at an $80 AOV ($200 gross revenue) with a 45% pre-ad contribution margin: 12-Month Contribution LTV = $80 × 2.5 × 0.45 = $90.00.

LTV:CAC ratio / payback period

The ratio compares customer value to acquisition cost; payback period is the time needed to recover CAC from customer contribution.

Example: Repeat purchases recover acquisition cost within a few months.

Formula: LTV:CAC Ratio = Contribution LTV ÷ CAC | CAC Payback Period (Months) = CAC ÷ Monthly Contribution Margin per Customer | Worked Example: With a 12-Month Contribution LTV of $90 and a CAC of $30: LTV:CAC Ratio = $90 ÷ $30 = 3.0:1. If the first order generates $22.50 in pre-ad contribution and month-2 repeat orders add $7.50, CAC Payback is achieved in Month 2 ($22.50 + $7.50 = $30.00).

Break-even ROAS

The return on ad spend needed to cover specified costs; equals 1 divided by pre-ad contribution margin under matching assumptions.

Example: A 40% pre-ad margin gives a 2.5x break-even ROAS.

Formula: Break-Even ROAS = 1 ÷ Pre-Ad Contribution Margin (Decimal) [or 100% ÷ Pre-Ad Contribution Margin %] | Worked Example: If a product has a 40% pre-ad contribution margin (0.40 after COGS, shipping, and payment fees): Break-Even ROAS = 1 ÷ 0.40 = 2.50x (or 250%).

Marketing efficiency ratio (MER) (MER)

Total business revenue divided by specified marketing spend for the period.

Example: 50,000 in revenue divided by 10,000 in spend equals 5x.

Formula: Marketing Efficiency Ratio (MER) = Total Store / Business Revenue ÷ Total Marketing Spend [Also expressed as Marketing % = Spend ÷ Revenue] | Worked Example: A brand generates $50,000 in total store revenue while spending $10,000 across Meta, Google, and TikTok ads: MER = $50,000 ÷ $10,000 = 5.0x (meaning ad spend is 20% of total revenue).

Return on investment (ROI) (ROI)

Net return associated with an investment divided by its cost within a defined scope and period.

Example: 200 in net profit divided by 1,000 in cost equals 20%.

Formula: Return on Investment (ROI %) = ((Net Profit from Investment - Cost of Investment) ÷ Cost of Investment) × 100 | Worked Example: A campaign costs $1,000 in total investment and generates $3,000 in sales; after deducting $1,800 in COGS and fulfillment, net gain before investment is $1,200, leaving $200 net profit: ROI = ($200 ÷ $1,000) × 100 = +20.0%.

Marginal ROAS

Additional revenue divided by additional spend when investment is increased.

Example: Adding 1,000 in spend matched by 2,000 in additional revenue equals 2x.

Formula: Marginal ROAS (mROAS) = Change in Revenue (Revenue_New - Revenue_Old) ÷ Change in Spend (Spend_New - Spend_Old) | Worked Example: At $5,000 spend, revenue is $20,000 (4.0x average ROAS). Scaling spend by +$1,000 to $6,000 raises revenue by +$2,000 to $22,000 (3.67x average ROAS): Marginal ROAS on the extra $1,000 = $2,000 ÷ $1,000 = 2.0x.

New customer ROAS

First-order revenue from new customers divided by the specified spend in the report.

Example: Separating first-order value from returning customer orders.

Formula: New Customer ROAS (ncROAS) = First-Order Revenue from New Customers ÷ Total Acquisition Ad Spend | Worked Example: Total store revenue from ads is $15,000 on $5,000 spend (3.0x total ROAS), of which $10,500 comes from first-time buyers and $4,500 from returning customers: ncROAS = $10,500 ÷ $5,000 = 2.1x.

Repeat purchase rate

The percentage of customers in a defined group who make an additional purchase within the measurement window.

Example: 30 out of 100 customers bought a second time within 90 days.

Formula: Repeat Purchase Rate (%) = (Cohort Customers with >= 2 Orders within Window ÷ Total Customers in Cohort) × 100 | Worked Example: Out of 100 first-time buyers acquired in January, 30 place a second order within 90 days: 90-Day Repeat Purchase Rate = (30 ÷ 100) × 100 = 30.0%.

Refund / return / cancellation

A refund returns money; a return sends back a product; a cancellation stops an order.

Example: A paid order that was canceled and refunded.

Cash on delivery (COD) (COD)

Collecting order payment at the time of delivery through the merchant or carrier service.

Example: Phone confirmation before dispatching the order.

Formula: Net COD Realization Rate (%) = (COD Orders Delivered & Cash Collected ÷ Total COD Orders Placed) × 100 | Effective CPA = Ad Spend ÷ Delivered Orders | Worked Example: A campaign spends $600 and generates 30 COD orders on the website ($20 checkout CPA), but 6 cancel on phone confirmation and 4 refuse courier delivery (20 delivered & collected): Net COD Realization = (20 ÷ 30) × 100 = 66.7%, and true Delivered CPA = $600 ÷ 20 = $30.00.

Fulfillment / shipping subsidy

Fulfillment covers processing and packing an order; shipping subsidy is the portion of delivery cost paid by the merchant instead of the customer.

Example: The merchant pays part of the delivery fee.

Customer relationship management (CRM) (CRM)

A system and practice for organizing customer data, interactions, opportunities, and follow-ups.

Example: Logging every inquiry, its status, and the assigned owner.

Pipeline / lifecycle stage

A pipeline tracks deal progress; a lifecycle stage describes a person's status in their relationship with the business.

Example: Inquiry, then qualified, then proposal, then customer.

Marketing / sales qualified lead (MQL / SQL) (MQL / SQL)

Agreed classifications for fit and readiness for follow-up or sales engagement.

Example: A suitable inquiry that confirmed its need and timeline.

Lead scoring

Ranking leads using agreed indicators of fit and interest.

Example: Points assigned for matching need and recent activity.

Lead nurturing

Helpful, progressive follow-up that assists an interested prospect in evaluating until ready.

Example: Sending an explanation and comparison, then answering an objection.

Lifecycle marketing

Adapting communication to a person's stage from discovery through repeat purchase and referral.

Example: A welcome message, then onboarding support, then a reorder prompt.

Email campaign / automated flow

A campaign is a scheduled send; an automated flow is a sequence triggered and branched by rules and actions.

Example: A seasonal newsletter versus a welcome series.

Trigger / condition / action

A trigger starts the automation; a condition selects the path; an action executes a step.

Example: A new order, then checking the city, then assigning to a staff member.

Welcome / abandoned checkout / post-purchase flow

Automated messages tailored to sign-up, unfinished checkout, or completed orders.

Example: Usage instructions sent after delivery.

Reactivation / win-back

Appropriate outreach to a customer or lead who has stopped engaging or buying.

Example: A message to a customer who has not purchased in a long time.

Email deliverability

The ability of messages to reach recipients' inboxes rather than being rejected or sent to spam.

Example: Improving sender reputation, list hygiene, and authentication.

SPF / DKIM / DMARC (SPF / DKIM / DMARC)

Mechanisms for verifying sending sources, cryptographic signing, and policies for handling authentication failures.

Example: Configuring the company domain before launching email campaigns.

Open rate

The percentage of delivered messages recorded as opened by the tracking system.

Example: 200 unique opens out of 1,000 delivered messages.

Formula: Email Open Rate (%) = (Unique Tracked Opens ÷ Total Emails Delivered [Sent - Bounced]) × 100 | Worked Example: A newsletter is sent to 1,050 addresses, 50 bounce (1,000 delivered), and 200 unique recipients open it: Open Rate = (200 ÷ 1,000) × 100 = 20.0%.

Email click rate / CTOR

Click rate is unique clicks divided by delivered messages; CTOR is unique clicks divided by unique opens as defined by the tool.

Example: 50 clicks divided by 1,000 delivered equals 5%.

Formula: Email Click Rate (%) = (Unique Clicks ÷ Emails Delivered) × 100 | Click-to-Open Rate (CTOR %) = (Unique Clicks ÷ Unique Tracked Opens) × 100 | Worked Example: From 1,000 delivered emails with 200 tracked opens and 50 unique link clicks: Email Click Rate = (50 ÷ 1,000) × 100 = 5.0%, and CTOR = (50 ÷ 200) × 100 = 25.0%.

Hard / soft bounce

A hard bounce is a permanent failure unlikely to resolve; a soft bounce may come from a temporary issue.

Example: A nonexistent address versus a full inbox.

Opt-in / double opt-in / unsubscribe

Opt-in is permission to receive messages; double opt-in adds a confirmation step; unsubscribe stops the subscription.

Example: Confirming an email address through a link.

WhatsApp template / session message

A template is a format subject to platform approval and rules; session messages are governed by the service window and terms.

Example: A follow-up template for a customer who requested information.

Customer data platform (CDP) (CDP)

A system that unifies customer data from multiple sources to build usable profiles.

Example: Linking visits, orders, and support tickets according to identity and consent.

Artificial intelligence (AI) (AI)

Systems that perform tasks requiring capabilities such as recognition, prediction, generation, and specific decision-making.

Example: Classifying customer inquiries.

Machine learning (ML) (ML)

Methods that learn patterns from data to use in prediction or decision-making.

Example: Predicting purchase probability from past signals.

Generative AI (GenAI) (GenAI)

Systems that produce new content such as text, images, audio, and video.

Example: Drafting ad copy variations from a brief.

Large language model (LLM) (LLM)

A model that learns language patterns and uses them to generate or process text and perform related tasks.

Example: Summarizing interviews or categorizing objections.

Foundation model

A broad general-purpose model trained at scale that can be used and adapted for many tasks.

Example: A single model supporting writing, summarization, and analysis.

Multimodal model

A model that processes or produces more than one data type, such as text, image, and audio.

Example: Reading an ad image and explaining its elements.

Model / AI application

The model is the generation or reasoning engine; the application adds an interface, tools, and context around it.

Example: A chat app that pairs a model with web search.

Training / inference

Training adjusts model parameters from data; inference uses the model to produce an output.

Example: Asking a ready-made model a question is inference.

Prompt / prompt engineering

The request given to a model and how it is structured to clarify the task, context, and output criteria.

Example: Specifying the audience, goal, length, and style examples.

Context engineering

Selecting and organizing the information, instructions, and tools the model needs during a task.

Example: Attaching the brand guide, offer details, and trusted sources.

System / developer / user instructions

Instruction layers that govern general behavior, application rules, and user requests according to the platform hierarchy.

Example: Application rules followed by a request to write an ad.

Zero-shot / few-shot prompting

Zero-shot describes the task only; few-shot adds examples that illustrate the desired output.

Example: Providing three approved Arabic copy examples before requesting a new version.

Prompt template / prompt chaining

A template is reused with variables; chaining connects the output of one step to the input of another.

Example: Research, then brief, then draft, then review.

Token

A piece of text representation counted by the model, which may be a word, part of a word, or a symbol.

Example: Prompt and response length affect token usage.

Context window

The limit of context and tokens related to a request that a model can process by design.

Example: A long document may exceed available capacity.

Memory / conversation history

Conversation history is what was said earlier in the chat; memory is a mechanism for saving and retrieving details across tasks.

Example: Saving a preferred writing style.

Hallucination / confabulation

An output that sounds plausible but is false or unsupported by the context and sources.

Example: Inventing a study or a product specification.

Grounding

Constraining the model's work to information, tools, and sources tied to the task.

Example: Writing a reply strictly from a verified knowledge base.

Retrieval-augmented generation (RAG) (RAG)

Fetching relevant information and adding it to the context before generating an answer.

Example: Searching the product database and then replying with its specifications.

Fine-tuning

Updating a model's parameters using examples or data to align its behavior or task performance.

Example: Customizing a consistent classification style.

Reasoning model / effort

A model or setting that allocates extra processing to solve tasks requiring multi-step planning.

Example: Analyzing conflicts in a plan with multiple constraints.

Temperature

A setting that adjusts the randomness of output selection in models that support it.

Example: Lowering variation when extracting fixed fields.

Structured output / JSON schema

Producing data in a defined format that is easy to parse and validate programmatically.

Example: Returning product name, price, and currency in separate fields.

Evaluation (eval) / golden dataset (eval)

Tests and criteria paired with reference examples to measure the quality of an AI system.

Example: A sample set of responses scored for accuracy, tone, and safety.

Human-in-the-loop (HITL) (HITL)

Inserting human verification or decision-making at key points in an AI workflow.

Example: Approving ad claims before publication.

Automation / workflow

Automation executes work automatically; a workflow links steps and conditions to achieve an outcome.

Example: Logging an inquiry, assigning it, and sending an alert.

Chatbot / AI assistant

Conversational interfaces for answering or assisting, which may use rules or models and tools.

Example: An assistant that answers product questions.

AI agent

A system that uses a model to dynamically choose steps and tools to achieve a goal within granted permissions.

Example: Searching for sources, building a brief, and verifying it.

Agentic workflow

Coordinating steps, models, and tools with varying degrees of dynamic decision-making.

Example: Gathering data and passing it to an analysis agent and a reviewer.

Tool use / function calling

A model requesting the execution of a specific function through a structured interface and then using its result.

Example: Querying live inventory instead of guessing availability.

Application programming interface (API) (API)

Communication rules that allow software programs to request data or actions from other programs.

Example: Connecting a CRM to the order management system.

Webhook

One system sending an automated request to another when an agreed event occurs.

Example: When a new purchase happens, a notification hits the automation tool.

Connector / integration

An integration connects systems; a connector is a prebuilt component that simplifies that link.

Example: Reading a file from Drive inside an assistant.

Model Context Protocol (MCP) (MCP)

A standard for connecting AI applications to external tools, data sources, and context.

Example: An assistant connecting to a file source through an MCP server.

Agent skill / SKILL.md

A package of instructions, resources, and optionally scripts that helps an agent execute a specific type of work.

Example: A skill for auditing copy against the brand guide.

Knowledge base

Organized, trusted content that people or systems reference to answer questions and execute work.

Example: Specifications, options, and policies with a last-updated date.

Embedding / vector store

An embedding converts content into numbers representing its characteristics; a vector store enables searching by them.

Example: Searching for questions with similar meaning.

Semantic / hybrid search

Semantic search looks for meaning similarity; hybrid search combines it with keyword matching or other signals.

Example: Finding a refund policy despite different phrasing.

Chunking / retrieval

Chunking splits materials into smaller units; retrieval selects the units relevant to a query.

Example: Fetching the shipping paragraph from a long manual.

Orchestration / routing

Orchestration manages sequence and resources; routing selects the most suitable tool or path.

Example: Sending a pricing question to a data tool and a style question to an editor.

Multi-agent system

Multiple agents with different roles or capabilities coordinating to complete work.

Example: A researcher, then a writer, then a reviewer.

Browser / computer-use agent

An agent that executes steps through an application interface much like a human user does.

Example: Opening a page and entering data within granted permission.

No-code / low-code automation

Building workflows through visual interfaces with little or no manual programming.

Example: Connecting a form, CRM, and email in a visual editor.

Coding agent / vibe coding

A coding agent reads, edits, and tests code; vibe coding is a style of guiding software creation through natural language.

Example: Building a website prototype through iterative prompts.

Guardrails / permissions

Constraints, checks, and access limits that define what a system can do and how.

Example: Reading customer records without permission to send messages.

Prompt injection

Untrusted content that attempts to divert a model or its tools away from the authorized task.

Example: A web page instructing an agent to expose a private file.

Sandbox

An environment that restricts program access and execution to limit the impact of errors or harmful content.

Example: Testing code without access to production data.

Observability / trace

Logging and reviewing system steps, tool calls, costs, and errors to understand what happened.

Example: Inspecting the source retrieved before a wrong answer was produced.

Latency / rate limit / cost budget

Operational constraints that govern system speed, request volume, and allowed spend.

Example: Capping the number of search calls per task.

Text-to-image / text-to-video

Creating visual media from a written description of the scene, style, and constraints.

Example: A product scene on a table with specific lighting.

Image-to-image / image-to-video

Using a source image to guide a new image or animate a scene into video.

Example: Animating a product shot while keeping its composition.

Reference image / style reference

An image that guides identity, composition, or overall look depending on the generation tool.

Example: One reference for lighting and another for the required shape.

Negative prompt

A description of elements you want to avoid in tools that support exclusion prompts or constraints.

Example: Avoiding extra fingers or unreadable text.

Seed / reproducibility

A random starting value that can help reproduce an output when other conditions stay fixed.

Example: Re-running the same setup for comparison.

Inpainting / outpainting

Inpainting generates inside a selected region; outpainting adds content beyond the image borders.

Example: Fixing a background or extending a vertical scene.

Upscaling

Increasing image or video dimensions using processing that may add estimated details.

Example: Preparing a small image for a larger placement.

Character / product consistency

Maintaining identity and key attributes across multiple images or shots.

Example: Keeping the exact same package and color across all scenes.

Temporal consistency / visual artifact

Temporal consistency keeps elements stable across frames; a visual artifact is an unintended distortion in the output.

Example: Keeping a hand or logo from morphing between shots.

Keyframe / motion prompt

A keyframe sets an important visual state; a motion prompt describes scene and camera transitions.

Example: A static shot followed by a slow zoom-in.

Text-to-speech (TTS) / speech-to-text (STT) (TTS)

TTS generates spoken audio from text; STT transcribes speech into text.

Example: Generating a voiceover or transcribing an Arabic interview.

Voice cloning / synthetic avatar

Technologies that mimic a human voice or appearance to produce synthetic media.

Example: A virtual presenter explaining a lesson.

Lip sync / AI dubbing

Aligning mouth movement with audio and adapting dialogue into another language.

Example: An Arabic version of an educational video.

Synthetic media / deepfake

Synthetic media is generated or modified content; a deepfake convincingly imitates a real person or event.

Example: A face or voice that appears realistic.

Content provenance / Content Credentials

Information about a file's origin and edits that can be bound to it using standards such as C2PA.

Example: A record showing some of the tools used to create an image.

AI bias

Unbalanced or unfair patterns resulting from training data, model design, or how the tool is used.

Example: Excluding local dialects or showing a stereotyped portrayal of a group.

Usage rights / licensing

The terms that govern how source materials and generated outputs may be used commercially and shared.

Example: Verifying the rights for music tracks and reference images.

AI output QA / fact-checking

Reviewing text, visuals, sources, claims, and usage rights before publication.

Example: Checking Arabic spelling, hands, packaging, and price.

Anchoring

A later evaluation being influenced by a reference number or piece of information shown first.

Example: Presenting price tiers where one option serves as a comparison reference.

Framing effect

A change in a person's response when the exact same information is presented in different ways.

Example: Explaining a service cost per day or per month while clearly showing the total.

Loss aversion

A tendency to weigh a potential loss more heavily than an equivalent gain under certain conditions.

Example: Highlighting the time wasted because of the problem.

Choice overload / paralysis

Choosing can become harder when alternatives are too numerous and comparison is complex.

Example: Grouping dozens of items into three clear bundles.

Mere exposure effect

A tendency toward familiarity or preference after repeated exposure under some conditions.

Example: Keeping brand elements consistent across content.

Post-purchase dissonance

Doubt or uneasiness after a decision about whether the choice was right.

Example: A post-purchase message explaining how to use the product and confirming what the buyer will receive.

Peak-end rule

A tendency for memory of an experience to be shaped heavily by its most intense moment and its ending rather than the average.

Example: Improving issue resolution and the final delivery handoff.

Zeigarnik effect

A name for the historical claim that unfinished tasks are remembered better; later evidence does not support a consistent general advantage.

Example: Testing an opening question without automatically crediting its success to this effect.

Charm pricing

Using price endings such as 9 to shift price perception in certain contexts.

Example: Pricing at 99 instead of 100 and testing its impact.

Cognitive load

The mental effort required to understand information or complete a task.

Example: Reducing form fields and explaining options simply.

Programmatic advertising

Buying and selling ad inventory using automated systems, rules, and data.

Example: Purchasing digital placements through a specialized platform.

Demand-side / supply-side platform (DSP / SSP) (DSP / SSP)

A DSP helps advertisers buy inventory; an SSP helps publishers offer and sell their inventory.

Example: An advertiser buying impressions across multiple publishers.

Real-time bidding (RTB) (RTB)

An automated auction for a single impression opportunity executed in real time.

Example: Submitting a bid when an eligible impression becomes available.

Retail media

Ad placements and services offered by retailers or marketplaces based on their shopping context and permitted data.

Example: A sponsored product inside a major marketplace's search results.

Connected TV / over-the-top (CTV / OTT) (CTV / OTT)

CTV refers to internet-connected television devices; OTT refers to video content delivered over the internet outside traditional broadcast.

Example: An ad inside a streaming service on a smart TV.

Native advertising

An ad formatted to match the surrounding editorial or feed experience while clearly disclosing its paid nature.

Example: A sponsored article recommendation on a news site.

Contextual targeting

Selecting ad placements based on page or scene topic rather than relying solely on user profiles.

Example: Showing a kitchenware ad next to a recipe.

Brand safety / suitability

Brand safety avoids harmful content; brand suitability considers what fits a specific brand's values and audience.

Example: Excluding contexts inappropriate for a family product.

Ad fraud / invalid traffic (IVT) (IVT)

Deceptive or invalid activity that generates impressions or clicks not coming from a real eligible audience.

Example: Automated clicks draining ad spend.

Frequency cap

A setting that limits how many times an ad is shown to a person or identifier within a period where supported.

Example: Setting a weekly impression limit for a reach campaign.

Data clean room

A controlled environment enabling collaborative analysis under strict access and aggregation rules without unrestricted raw data sharing.

Example: Measuring audience overlap between partners under an agreement.