Reference
The vocabulary of getting found.
Search and AI answers have their own language, and most of it is used to gatekeep rather than explain. Here is the plain version. Every term below is defined to stand on its own, in words a business owner can act on.
63 terms, including 3 we named ourselves. Written and reviewed by a Raveneye Global specialist.
#
- 16 CFR Part 465
- The FTC Trade Regulation Rule on the Use of Consumer Reviews and Testimonials, effective October 21, 2024, which makes specific fake-review practices a direct rule violation carrying civil penalties.
A
- Activity bias
- The upward distortion in observational ad estimates caused by the fact that users active enough to be served ads are, at that moment, already more likely to search, click, and buy.
- Advertising Value Equivalency (AVE)
- A method of valuing earned media by measuring the ad space or airtime the coverage occupied and reporting what an equivalent paid advertisement would have cost. Formally rejected by the Barcelona Principles.
- Agenda-setting
- The theory that media influence which topics an audience considers important by choosing what to cover, shaping outcomes through selection rather than persuasion.
- Agentic commerce
- Commerce in which an AI agent, acting on a buyer's stated intent, performs discovery, comparison, and increasingly checkout on the buyer's behalf, rather than a human clicking through a storefront directly.
- AI answer engine
- A search interface such as ChatGPT, Perplexity, Google AI Overviews, Gemini or Copilot that returns a synthesized written answer naming a few sources, instead of, or above, a ranked list of links.
- AI Overview
- A synthesized summary Google generates and displays above the organic search results for a query, drawing on multiple sources and often naming a small number of specific businesses or providers.
- AI-washing
- Describing a product as smarter, more autonomous, or more capable with artificial intelligence than the seller can substantiate. It is policed as a deceptive-advertising problem, independent of whether the technology works.
- Anchoring
- The tendency for an initial reference point, such as the first rating or average a buyer sees, to exert disproportionate pull on later judgments, even when it is known to be irrelevant.
- Answer engine optimization (AEO)
- The practice of structuring a page's entity signals, extractable answers, and structured data so a search engine pulls it directly into an answer instead of only ranking it in a list of links. AEO predates large-language-model answers, growing out of featured-snippet and voice-search work, and shares many tactics with generative engine optimization (GEO).
- Astroturfing
- Manufacturing the appearance of organic, independent endorsement or sentiment, for example through fabricated reviews or personas. The opinion-manipulation counterpart to interface dark patterns, and squarely within FTC deception doctrine.
- Attention economy
- The idea that in an information-rich world attention is the scarce resource being competed for and allocated, and therefore the resource a market forms to buy and sell. Stated formally by Herbert Simon in 1971 and named for the network age by Goldhaber in 1997.
- Autonomy level
- A graded measure of how much an AI agent may decide and act on its own. The Cloud Security Alliance framework runs from L0 (no autonomy) to L5 (fully autonomous), with human-approval controls at the lower levels.
C
- Capacity gap
- A shortfall in the resources, skills, time, or infrastructure required to adopt a capability, as distinct from a lack of desire to adopt it. The OECD evidence frames small-firm digital adoption as constrained by capacity, not attitude.
- Consent Mode v2
- Google's mechanism for passing a structured signal of a visitor's consent choices to Google tags before they collect or use data. Required, per Google's documentation, for measuring EEA visitors with Google tags.
- Consideration set
- The small group of options a buyer actually weighs before choosing. When an engine returns one synthesized answer, the consideration set can shrink to the few businesses named inside it.
- Core Web Vitals
- Google's published metrics for loading speed, interactivity, and visual stability. A booking page that clears these thresholds reaches a working form faster on a phone.
- Counterfactual
- The road not taken, recorded. In ad measurement, the ad a control user would have received had they been in the treated group, logged so their behavior can serve as an honest baseline.
- Credence good
- A purchase whose quality the buyer cannot reliably judge even after consuming it, because assessing it requires expertise the buyer lacks. Most medical-aesthetic procedures qualify, which is why credentials and institutions stand in for verification.
D
- Dark pattern
- A deceptive interface design intended to manipulate a user into an action they would not otherwise take, for example confirmshaming, forced continuity, obstructed cancellation, hidden costs, or preselected options. Named as a category by Harry Brignull in 2010.
- Disintermediation
- The removal of a middleman between buyer and seller, so the two transact directly instead of through an intermediary that charges for the introduction. A contractor who ranks in local search, for example, reaches the homeowner directly instead of paying a lead marketplace for that same introduction.
E
- E-E-A-T
- Google's acronym for the qualities of a trustworthy page, used in its Search Quality Rater Guidelines: Experience, Expertise, Authoritativeness, and Trust. It became E-E-A-T in December 2022, when Experience was added to the earlier E-A-T.
- Entity
- A distinct real-world thing (a specific business, person, or place) modeled as a node in a graph, as opposed to a string of characters. Disambiguating a query to the right entity is the core move of modern search.
- Entity consistency
- A business resolving to one unambiguous, identically described identity, meaning the same name, address, and phone number, across its website, directories, and profiles, so a search or AI system can confidently recognize it as a single real-world entity. Weak entity consistency is a common reason an otherwise well-ranked business fails to get named in the answer an AI engine gives.
- Entity resolution
- The process by which engines decide that the business, person, or service on one page is the same real-world entity as the one referenced elsewhere. Aided by sameAs, but adjudicated probabilistically, not asserted by a single field.
F
- Filter bubble
- The self-reinforcing narrowing of the information and viewpoints a person sees, produced when a platform's personalization algorithm keeps showing content similar to what that person already clicked or agreed with. Internet activist Eli Pariser named the concept in his 2011 book The Filter Bubble.
G
- Gatekeeping
- Control over which information reaches an audience before they see it, held by whoever decides what passes through the channel between a source and its public, and how dependent that audience is on the gate. In search, the engine itself acts as gatekeeper of the buyer's attention, choosing which businesses it shows or names.
- Generalized second-price auction (GSP)
- The auction behind most keyword advertising. Advertisers are ranked by bid (weighted by ad quality), the top bidder wins the top slot, and each winner pays a price based on the bid of the competitor ranked below them, not their own bid.
- Generative Engine Optimization (GEO)
- The academically-coined discipline of getting content retrieved and cited inside an LLM-generated answer. The founding benchmark found credible citations, quotations, and specific statistics to be the strongest content levers.
- Geo experiment
- A causal measurement design that randomizes non-overlapping geographic regions into ad and no-ad conditions, measuring lift at the level of markets rather than individuals and requiring no user-level tracking.
H
- Hallucination
- In natural-language generation, the production of fluent, confident text that is false or unsupported by the source. Treated by the research literature as a structural property of the generative mechanism, not a rare bug.
- Human-in-the-loop
- A design pattern in which a person reviews or approves an AI system's consequential actions before they take effect, concentrated at the points where a wrong answer or action would be costly or hard to reverse.
I
- Incremental ROAS (iROAS)
- Return on ad spend computed only on the revenue the ads causally produced, established by a holdout, rather than all attributed revenue. Almost always lower than the platform-reported figure.
- Incrementality
- The additional outcomes advertising actually caused, above the baseline a business would have earned with no ad. It is the difference between the treated and control groups, not the total the platform attributes.
- Informational cascade
- A sequence in which each person, observing the visible choices of those before them, rationally follows the crowd and ignores their own weak private signal, so later choosers copy rather than add information.
- Insider review
- A review written by an officer, manager, employee, or their immediate relative. It must be disclosed as a material connection and cannot be presented as an ordinary customer review.
- Interaction to Next Paint (INP)
- A measure of how quickly a page visibly responds after a user taps or clicks. The passing threshold is 200 milliseconds or less, and it tends to be the hardest Core Web Vital for typical pages to pass.
K
- Knowledge graph
- A structured map of real-world entities, such as people, places, and businesses, and the facts and relationships connecting them, that a search system reasons over instead of matching text. Google's Knowledge Graph is the best-known example of one in production.
L
- Largest Contentful Paint (LCP)
- The time it takes for the largest visible element of a page to render. A core measure of how quickly a page feels loaded; the passing threshold is 2.5 seconds or less at the 75th percentile of real users.
- Last-click attribution
- A reporting rule that assigns a conversion to the final ad click preceding it. It counts clicks that came before sales and cannot distinguish persuasion from the ad merely being present for a committed buyer.
- Local pack
- The set of local business results, typically a map with three listings, that Google shows above the organic results for a location-relevant search. Also called the map pack, it is ranked largely by proximity, relevance, and prominence rather than national brand scale, and for many local searches it is the only result a searcher sees.
M
- Machine-legibility
- The extent to which a business's data is structured, crawlable, and interpretable by the search and AI systems buyers use to find and compare suppliers, a capability requirement layered on top of an already uneven digital divide.
- Machine-Readiness Score Raveneye Global
- A single measured number from 0 to 100 that reads how findable a business is across classic search, the local map pack, AI answers, and its reputation. It is the metric RavenEye reports movement against, rather than claiming a result.
- Marketing Efficiency Ratio (MER)
- Total revenue divided by total marketing spend, a blended account-level figure that cannot be gamed by shifting attribution credit between channels because it assigns no channel-level credit.
- Marketing mix modeling (MMM)
- A statistical method that regresses aggregate outcomes on marketing activity over time, with carryover and saturation transforms, to infer contribution without following individual users.
- Media ecology
- The study of media as environments rather than channels; it examines how the structure of a media surround shapes the thinking, feeling, and behavior of the people inside it. Named by Neil Postman in 1968.
- Multi-touch attribution (MTA)
- An observational method that distributes conversion credit across the several marketing touches a converting customer encountered. Because it reads the paths of people who already converted, it does not establish what the ads caused.
N
- NAP consistency
- Whether a business's name, address and phone number read identically across every directory and citation source, a core signal engines use to resolve which business is which.
- Natural experiment
- A situation where something close to random assignment happens in the real world (such as Yelp rounding a 3.24 and a 3.26 to different half-stars), letting researchers estimate cause rather than mere correlation.
P
- Peeking
- Repeatedly checking a running test and stopping the moment it looks significant. It inflates false positives because every look is another chance for noise to cross the line.
- Primary category
- The single Google Business Profile field that most precisely tells the engine what kind of business you are. It is one of the few local ranking signals a business owner can set directly, which makes it disproportionately powerful for showing up in the right local searches.
R
- Retrieval-augmented generation (RAG)
- A pattern where an engine retrieves documents and generates an answer conditioned on them, often with citations. It narrows the room for fabrication but does not guarantee the answer faithfully represents the retrieved sources.
- Review gating
- A request workflow that screens customers by sentiment first, then sends only the satisfied ones to a public review platform while diverting the dissatisfied to a private channel. Under 16 CFR 465 it is a prohibited form of review suppression.
- Review recency
- How recently the reviews on a business were written, as distinct from how many it has in total. Buyers and search algorithms alike tend to treat a business with recent reviews as more current and more trustworthy than one whose reviews have gone stale.
- Review velocity
- The rate at which a business earns new reviews over time, for example five reviews a month rather than five a year. A steady or rising velocity signals an active, currently operating business, independent of its average rating or total review count.
S
- Satisficing
- Herbert Simon's term for acting on the first adequate option rather than the optimal one, because exhaustively evaluating alternatives is itself costly. A synthesized answer is a near-ideal satisficing device.
- Search Surface Optimization Raveneye Global
- The method of engineering a business into every surface a buyer's question now reaches, from Google and the map pack to the answers AI engines give in place of a list of links. It is the discipline the Machine-Readiness Score measures.
- Speed to lead
- The elapsed time between a prospect making contact and the business making its first human response. The lead-response literature finds it has an outsized, nonlinear effect on whether the lead is qualified.
T
- The Visibility Corpus Raveneye Global
- A living map of where a market's attention actually sits across every digital surface, read per industry to show a business where the best return on visibility work lies.
V
- Vanity metric
- A number that is real and easy to move in a flattering direction, but that does not correspond to the outcome the business actually cares about. It rises with activity while saying nothing about results.
Z
- Zero-click search
- A search that ends without a click to any external website, because the answer is satisfied on the results page. Its rise is well evidenced and is separate from any change in search volume.
Knowing the words is step one. Knowing your number is step two.
A specialist-reviewed read of how findable your business is across search and AI answers, in the language above. No guaranteed number, and no obligation.