The words behind
reviews, SEO, and AI search.

An aggregate rating is the averaged score across every review of a product, and the schema.org property of the same name that lets search engines render that average as a star snippet beside a listing, summarising many ratings into one citable number.

First-party data is information a business collects directly from its own customers through their interactions with it, such as purchases, account details, survey responses, and product reviews, gathered with consent and owned by the business rather than bought from an outside source.

An incentivized review is one a customer writes after being offered a reward (a discount, loyalty points, a free sample, or an entry to a draw), where the reward must be given regardless of what the review says and must be clearly disclosed, never conditioned on the review being positive.

A photo review is a customer review that includes one or more images the buyer took of the product they received, attached to their written rating so other shoppers can see the item in real use rather than only in the brand's own studio photography.

Questions and answers is a product-page section where a prospective buyer asks something about a product and the brand or a past buyer answers in public, building a running thread of buyer-led questions and specific responses attached directly to the listing.

Review gating is the practice of pre-screening customers and routing only the satisfied ones to post a public review while steering unhappy ones to a private feedback channel, so the published ratings look better than the full customer experience actually was.

Review syndication is the practice of sharing the same customer reviews across multiple products, sites, or retail partners, so that a review collected in one place (for example on a brand site) also appears wherever the product is sold, such as a stockist or marketplace listing.

Review velocity is the rate at which a product gathers new reviews over time, measured as the count of fresh reviews per week or month, which signals to shoppers and to AI systems that the product is actively bought, recently validated, and not relying on stale praise.

A review widget is the on-page component that displays customer reviews on a product or collection page, typically showing the star rating, review count, and individual review text, and is delivered either as JavaScript that injects the content after load or as server-rendered HTML.

Sentiment analysis is the automatic classification of review text as positive, negative, or neutral, often paired with theme extraction that groups recurring topics like shipping, sizing, or support so a store can read the mood of hundreds of reviews without reading each one.

Social proof is the tendency for people to trust what others have already chosen, so on a store it shows up as reviews, ratings, sold counts, and customer photos that signal a product is a safe, common choice and make a hesitant shopper more willing to buy.

A star rating is the 1-to-5 summary of customer sentiment for a product, usually shown as an average of every individual review score so a shopper can read overall quality at a glance without opening each review.

User-generated content is content created by customers rather than the brand: reviews, ratings, photos, videos, and questions and answers, published on a product or store page where shoppers and search engines can read it as first-hand evidence of the product in real use.

A verified buyer is a reviewer whose purchase of the product has been confirmed against order records, so the review carries a label (such as "Verified Buyer" or "Verified Purchase") signalling that the person writing it actually bought the item being reviewed.

Alt text is a short written description of an image, set in the HTML alt attribute, that screen readers announce to visually impaired users and that search engines read to understand what a picture shows and when to surface it in image search.

A backlink is a link from another website to a page on yours, and it acts as a core trust signal in search: each one is read as a vote of confidence, telling search engines that another site found your page worth pointing to.

A canonical URL is the version of a page you tell search engines to treat as the master copy when several URLs serve duplicate or near-duplicate content, declared with a rel=canonical link in the head so ranking signals consolidate onto one address instead of being split.

Core Web Vitals are Google’s three field metrics for page experience: Largest Contentful Paint (LCP) for loading speed, Interaction to Next Paint (INP) for responsiveness, and Cumulative Layout Shift (CLS) for visual stability, each measured on real user visits.

Crawlability is how easily a search engine crawler can reach a page and read its content, governed by whether the page is linked, allowed by robots rules, served without errors, and rendered so its text exists in the HTML the crawler actually parses.

E-E-A-T is the Google framework standing for Experience, Expertise, Authoritativeness, and Trustworthiness, a set of quality signals raters and ranking systems use to judge whether a page and its creator are credible enough to deserve visibility, especially on topics that affect money or health.

A featured snippet is a boxed direct answer Google displays above the standard results, pulled verbatim from a single ranking page, that responds to a query in a sentence, list, or table before the searcher clicks anything.

A Google seller rating is a merchant-level star score, shown out of five, that Google compiles from approved review sources and displays next to your business across Search ads, Shopping, and the Google Customer Reviews badge, rating the store as a whole rather than any single product.

Internal linking is the practice of placing links between pages on your own site, which spreads ranking signal from strong pages to weaker ones, helps search crawlers discover and understand related content, and guides readers from one relevant page to the next.

JSON-LD (JavaScript Object Notation for Linked Data) is the Google-preferred format for structured data: a script block of linked-data JSON, placed in a page, that describes the entity on that page to search engines without touching the visible HTML.

A meta description is the short summary tag in a page's HTML that a search engine may display beneath its result; it is not a direct ranking factor, but a clear, relevant description can improve click-through rate, and Google often rewrites it to better match the searcher's query.

A rich snippet is an enhanced search result that displays extra detail beyond the standard title, URL, and description, such as star ratings, review counts, price, or availability, drawn from structured data Google reads on the page.

Schema markup is structured vocabulary from schema.org that you add to a web page to label its content for machines, telling search engines and AI systems exactly what each element means so they can render rich results and recognise the page as a specific entity.

Structured data is machine-readable information added to a web page, usually following the shared vocabulary at schema.org, that describes what the page is about (a product, a recipe, a review) in a fixed format so search engines and AI systems can read it without guessing.

A title tag is the HTML element that sets a page title, shown by search engines as the clickable blue headline of a result and used in the browser tab, making it one of the strongest on-page signals for what a page is about.

An XML sitemap is a file that lists the URLs on a site you want search engines to know about, so crawlers can discover and prioritise pages for indexing rather than relying only on following internal links to find them.

An AI hallucination is when a language model states something false as if it were true, presenting fabricated facts, citations, or details with the same confident tone it uses for correct answers, because the model predicts plausible text rather than retrieving verified information.

Answer engine optimization is the practice of shaping a brand and its product information so it gets quoted directly inside AI-generated answers, such as ChatGPT, Perplexity, and Google AI Overviews, rather than only ranking as a blue link a person has to click through to read.

Generative engine optimization is the practice of shaping content so it is selected, quoted, and cited inside the answers produced by generative AI systems such as ChatGPT, Gemini, and Google AI Overviews, rather than only ranking as a blue link in a results page.

Google AI Overviews are the AI-generated summaries Google places above the traditional blue links, synthesising an answer from multiple ranking web pages and citing a handful of them, so a query can be resolved on the results page without the user clicking through.

Grounding is the practice of tying an AI-generated answer to verifiable source material, so the model draws from retrieved documents rather than its own memory, which is what lets the answer carry citations back to the pages it actually relied on.

A knowledge graph is a structured database of entities (people, companies, products, places) and the relationships between them, which Google and other systems use to recognize a thing as a known entity rather than a loose string of words on a page.

A large language model is a neural network trained on vast amounts of text to predict the next piece of language, and it is the model type behind systems like ChatGPT, Claude, and Gemini, generating fluent answers one token at a time rather than retrieving stored facts.

llms.txt is a proposed plain-text file placed at a site root that points AI crawlers and language models to the pages and facts a site considers most important, written in Markdown so a model can read it without parsing the full HTML of every page.

A prompt is the question or instruction a person types or speaks to an AI system, such as a chatbot or an AI search assistant, which the model reads to decide what to retrieve, generate, and cite in its answer.

Retrieval-augmented generation is a technique where a language model first retrieves relevant documents from an external source, then generates its answer grounded in that retrieved text, so the response reflects specific, citable material rather than the model recalling facts from its training weights alone.

Semantic search retrieves results by meaning rather than exact keyword matches, converting the query and candidate documents into embeddings (numeric vectors) and ranking by how close those vectors sit in that space, so a page can match a question even when it shares no words with it.

A zero-click search is a query that ends without the user clicking any result, because the answer is delivered directly on the results page through a featured snippet, knowledge panel, or AI Overview rather than on a destination site.

A/B testing is a method of comparing two versions of a page or element by splitting live traffic between them at random, then measuring which version converts better so that the winner is chosen from real behaviour rather than opinion.

Cross-selling is the practice of recommending complementary products alongside the one a shopper is already buying, such as a case with a phone or filters with a coffee maker, to raise order value by adding items that genuinely fit the purchase.

A free shipping threshold is the minimum order value a customer must reach to qualify for free delivery, a rule stores use to nudge each order higher by giving shoppers a clear, low-friction reason to add one more item rather than pay for shipping.

Scarcity is a conversion tactic that signals limited stock or limited time, such as low-stock counts, countdown timers, or short-run availability, to prompt a shopper to act sooner rather than postpone the decision and drift away.

A trust badge is a small graphic placed near a checkout or add-to-cart action to reassure shoppers about security, payment safety, or guarantees, such as an SSL seal, accepted-card icons, a money-back promise, or a verified-merchant mark.

An upsell is an offer that moves a shopper from the item they are considering to a higher-value version of the same thing: a larger size, a premium tier, a longer subscription, or a model with more features, so the order is worth more without selling them something unrelated.

Add-to-cart rate is the share of sessions in which a shopper adds at least one product to the cart, calculated as the number of sessions with an add to cart divided by total sessions over the same period.

Average order value is the average amount a customer spends in a single order, calculated as total revenue divided by the number of orders over the same period.

Bounce rate is the share of visitors who arrive on a page and leave without triggering a second interaction, such as clicking through, scrolling past a tracked point, or moving to another page, expressed as a percentage of total sessions to that page.

Cart abandonment rate is the share of started checkouts that are never completed, calculated as one minus the number of finished purchases divided by the number of carts created over the same period, so a higher figure means more lost intent at the final step.

Click-through rate is the share of people who click a link after seeing it, calculated as clicks divided by impressions and expressed as a percentage; it measures how often a search result, ad, or email link earns the click rather than just the view.

Conversion rate is the percentage of visitors who complete the desired action, usually a purchase, calculated as conversions divided by sessions over the same period, then multiplied by 100; it is the central measure of how well a store turns traffic into outcomes.

Customer acquisition cost is the average amount a business spends to win one new customer, calculated by dividing total acquisition spend (ads, agencies, tools, and the team time behind them) over a period by the number of new customers gained in that same period.

Customer lifetime value is the total profit or revenue a single customer is expected to generate across the entire relationship, from first purchase to last, which tells you how much you can afford to spend acquiring that customer and still come out ahead.

Gross merchandise value is the total monetary value of all goods sold through a store or marketplace over a period, measured at sale price before deducting platform fees, payment costs, discounts applied after the sale, refunds, returns, or the cost of the goods themselves.

Net Promoter Score is a loyalty metric derived from a single survey question, how likely are you to recommend us on a 0 to 10 scale, calculated by subtracting the percentage of detractors (0 to 6) from the percentage of promoters (9 to 10).

Repeat purchase rate is the share of customers who buy more than once, calculated as the number of returning customers divided by total customers over a period, expressed as a percentage; it measures how reliably a store turns a first order into a habit.

Return on ad spend is the gross revenue earned for every unit of currency spent on advertising, calculated as revenue attributed to ads divided by ad spend, so a ROAS of 4 means 4 in revenue for every 1 spent.