AI Assistant is here! - Make the leap to conversational eCommerce

eCommerce Glossary

eCommerce concepts explained plainly for eCommerce Managers, marketers, and developers: from conversion rate and A/B testing to fuzzy search and headless commerce.

A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z

A

A/B testing

An experimental method that compares two versions of an element by showing each to a different group of users to measure which performs better. Commonly used in eCommerce to test changes to search ranking, layouts, or recommendations before a full rollout.

AI-assisted merchandising

The use of artificial intelligence to decide which products to show, in what order, and to whom, without relying on manual rules for every scenario.

Algorithmic ranking

The process by which a search engine orders results based on a combination of textual relevance, business signals, and user behavior. Unlike manual sorting or fixed rules, algorithmic ranking adjusts dynamically, weighing multiple factors simultaneously to surface the most relevant result for each query.

Analytics dashboard

A visual interface that brings together key search performance metrics in one place: queries, clicks, conversions, zero-result terms, and trends.

Assisted selling / guided selling

A shopping experience where technology (chatbots, quizzes, AI assistants) guides the user toward the right product, replicating the role of an in-store sales assistant.

Attribute-based filtering

A mechanism that allows users to narrow down results by selecting specific values from catalog attributes, such as color, brand, material, or price range.

Autocomplete

Real-time suggestions for terms, products, and categories that complete a user's search query before they finish typing it.

Autosuggest

A search feature that surfaces relevant products, categories, and query suggestions as the user types, before they finish writing.

B

Behavioral signals

Data generated by user interactions, searches, clicks, time on page, purchases, that the system uses to personalize and reorder results.

Behavioral targeting

Adapting what a shopper sees, products, search results, recommendations, banners, based on what they've already done: pages visited, searches made, products clicked, items bought.

Boolean search

Search queries built with logical operators (AND, OR, NOT) that tell the engine exactly what to include, exclude, or treat as interchangeable.

Boosting rules

Configurable instructions that adjust the position of products in search results based on business criteria such as stock level, margin, seasonality, or newness. They allow merchandising teams to align search rankings with commercial priorities without overriding the relevance algorithm entirely.

Browse abandonment

When a visitor browses products or runs searches but leaves without adding anything to the cart.

Bundle recommendations

Automatic suggestions that group complementary products into a pack or set, based on co-purchase data and catalog affinity.

C

Callbacks

Functions that execute automatically when a specific event occurs in the search engine, such as a completed search, a result click, or a filter change, allowing behavior to be customized.

Cart abandonment

When a shopper adds items to their online cart but leaves without completing the purchase.

Category merchandising

Manual or algorithmic control over the order, visibility, and presentation of products within a store's category pages.

Contextual search

The ability of the search engine to interpret user intent beyond exact keywords, using contextual signals to return more relevant results.

Conversational commerce

Buying through conversation via chat, WhatsApp, or voice assistants instead of navigating a traditional website.

Cross-selling

Suggesting complementary products based on what a customer is already buying, like recommending a phone case when someone buys a phone.

CSS

The code that controls how a website looks: colors, fonts, spacing, and layout, separate from the content itself.

CTR (internal search) / Search CTR

The percentage of search results that users actually click on, measuring how relevant and appealing those results are to what people were looking for.

Custom Results

Manually curated search results that override the default algorithmic output, giving full control over what appears for specific queries.

D

Data layer

A JavaScript object that acts as a bridge between store events, such as searches, clicks, and purchases, and analytics and advertising tools like Google Tag Manager.

Dynamic filtering

Filtering that recalculates the available options in real time based on the user’s selections, displaying only attribute combinations that lead to actual results.

Dynamic re-ranking

Automatic, real-time reordering of search results based on behavioral signals, context, and business rules, so each user sees the products most likely to convert at the top.

E

Edge caching

A technique that stores search responses on geographically distributed servers to serve frequent queries in milliseconds, without consulting the origin server.

Embedding

A numerical representation of a word or product that captures its meaning, allowing a search engine to compare concepts by semantic proximity rather than exact wording. "Running shoes" and "athletic footwear" are treated as close matches even though they share no words in common.

Empty search / null search

A term that covers two situations: when the user hasn't typed anything yet (empty query), and when what they've typed doesn't match any product (zero-results query).

Engagement rate

The percentage of sessions in which a user interacts meaningfully with the site, beyond simply landing on a page. In eCommerce search, it tracks whether users are clicking results, applying filters, or exploring products rather than bouncing after seeing the first page.

Entity extraction

A natural language processing (NLP) technique that automatically identifies and classifies entities with distinct meaning within a text, such as brands, categories, colors, sizes, and materials.

Exact match

Search behavior where results only come back when the user's query is a literal match with catalog data, without interpreting variations or related terms.

Excluded Results

A mechanism that prevents certain products from appearing in search results, regardless of their relevance or the search query.

F

Faceted navigation / faceted search

A dynamic filtering system that lets users narrow down a set of results by selecting product attribute values such as size, color, price, or brand.

Featured products

Products displayed in prominent positions within search results or category pages, selected either manually or through merchandising rules.

Filters

User interface controls that allow users to refine search results or category listings by selecting attribute values such as price, brand, size, or color.

First-party data

Information a business collects directly from its own customers, through purchases, site behavior, or sign-ups, rather than buying it from a third party.

Frequently Bought Together

An automated recommendation feature that displays products other customers have purchased together with the product the user is currently viewing, based on real co-purchase data.

Freshness scoring

A ranking mechanism that boosts or demotes search results based on how recently a product or piece of content was added or updated. Useful when recency matters, such as new arrivals or time-sensitive promotions.

Fuzzy search / typo tolerance

A search technique that returns relevant results even when a query contains typos or misspellings. A shopper searching for "sunnglases" still finds sunglasses.

G

Guided search

A search experience that helps users narrow down their intent through questions, progressive filters, or contextual suggestions, rather than expecting them to formulate a precise query from the start.

H

Headless commerce

An ecommerce architecture that decouples the customer-facing frontend from the backend systems handling business logic and transactions. The two layers communicate via APIs, giving teams flexibility to build custom experiences without being constrained by a monolithic platform.

I

Indexing

The process by which a search engine reads, analyzes, and stores catalog or website content so it can retrieve relevant results when a user searches. Any updates to the catalog need to be indexed before they appear in results.

Installation script

A snippet of JavaScript added to an online store to connect it to an external service and activate its features. In most cases it only needs to be added once, either directly in the store's theme or via a tag manager.

Intent detection / query intent detection

The automatic identification of what kind of result a user expects from a query: a product, a category, information, or a specific action.

Inverted index

A data structure that maps each word to every document or product in which it appears, allowing a search engine to locate relevant results in milliseconds. Instead of scanning the entire catalog for each query, the search engine simply looks up the term and retrieves the pre-built list of matches.

J

Journey analytics

The analysis of a user's full path from first interaction to conversion, mapping touchpoints and decisions at each stage. Helps teams understand where shoppers drop off and what drives them forward.

M

Merchandising rules

Configurable instructions that determine which products appear first in search results or category pages based on manually defined business criteria. Unlike algorithmic ranking, ranking rules give merchandising teams direct control over specific positions regardless of relevance signals.

Metadata

Structured information that describes the content of a page without being visible to the user, but which determines how search engines and external platforms interpret and display it.

N

NLP (natural language processing)

A set of techniques that allow a system to interpret text written in human language. In search, it identifies the intent behind a query and extracts its meaning beyond the exact words used.

No-results searches / zero-result searches

A query that the search engine is unable to answer with any product from the catalog, leaving the user facing a blank page.

O

Out-of-stock management

The decisions and settings that control how a site search handles products with no available inventory. Common approaches include demoting or hiding out-of-stock items, or keeping them visible with a "notify me" option.

P

People Also Viewed

A collaborative recommendation that surfaces products other shoppers viewed alongside the item currently being viewed. Often displayed as "customers who viewed this also viewed" widgets on product pages.

Personalized ranking

A search results order tailored to each individual user based on their interaction history, preferences, and behavioral signals. Two users searching for the same term may see a different order of results depending on how they have previously browsed, clicked, and purchased.

Product attributes

The structured data fields that describe a product (size, color, material, price, brand) and that the system uses to index it, filter it, and surface it in the right results.

Product categories

The hierarchical structure that groups products by common criteria like type, brand, material, and use case, and defines how users browse and discover the cata.

Product Discovery

The way a shopper finds products in an online store, whether they're actively searching or just browsing around and stumbling on things. It covers the internal search box, recommendations, category pages, and any AI assistant.

Product feed / data feed

A structured file containing all the data from a store's catalog: product names, prices, stock levels, images, and attributes, in a format that external tools and platforms can read and process. Search engines, comparison shopping sites, and ad platforms rely on a product feed to stay in sync with the catalog.

Product quiz / Quiz Maker

A product finder quiz is a short sequence of questions that guides users to the products that best match their needs, without requiring them to search or browse.

Product recommendations

Automatic product suggestions based on a visitor's behavior, broader purchase patterns, or relationships between catalog items. When done well, they increase average order value, reduce drop-off on product pages, and help shoppers discover less visible items.

R

Related searches

Search suggestions that appear on the results page and propose other queries to the user based on behavioral patterns or semantic similarity.

Relevance score / relevance ranking

A score that determines the order in which results appear when a user runs a search. The higher a product's relevance score, the higher it ranks, based on factors like how well it matches the query, its popularity, and any manual boosting rules in place.

Request

Any call made by your store or its users to Doofinder's servers to retrieve data, whether for search results, recommendations, indexing, or any other active function. API requests are typically how usage is measured and billed in search platforms.

Reranking

A process that reorders an initial set of search results by applying additional criteria, so the top results better reflect what is most relevant in context. Reranking typically happens after the initial retrieval step and can factor in signals like user behavior, business rules, or personalization.

S

Search conversion rate

The percentage of search sessions that end in a purchase. It measures how effectively the search engine turns intent into revenue. Users who search are already telling the store what they want, so a low search conversion rate points directly to a relevance or experience problem, not a traffic problem.

Search intent

The classification of what a user is trying to accomplish with a query: navigational, informational, transactional, or commercial investigation. Understanding intent helps a search engine surface the most relevant results rather than just the most literal match.

Search personalization

The ability of a search engine to tailor results to each user based on their behavior, rather than returning the same ranking for everyone. Personalization can draw on in-session signals like clicks and searches, or on longer-term data such as purchase history.

Search query / search term

The text a user types into a search box, which the engine matches against its index to return results. A query can be a single word, a phrase, or even a misspelled term. Understanding how queries are interpreted is key to diagnosing why certain searches return unexpected or poor results.

Search redirect

A rule that sends a user directly to a specific URL when they type a particular search term, bypassing the results page entirely. Useful for routing branded searches, promotional campaigns, or navigational queries to the right destination.

Search results page (SRP)

The page that displays products matching what a user typed into a store's internal search. It is one of the highest-intent pages in any eCommerce site, since users who search are actively looking for something specific.

Searchandising

The practice of applying business criteria, margin, stock level, seasonality, promotions, to influence which products appear first in eCommerce search results. It combines the logic of traditional merchandising with the mechanics of site search, giving teams control over ranking without compromising relevance.

Semantic search

Technology that interprets the meaning behind what the user types, not just the exact words, to return relevant results even when they don't literally match the catalog.

Site search / internal search

The search box on an online store and the engine behind it: the system that takes a user's query and returns relevant results from the store's own catalog.

Sitewide search

Search engine that operates across the entire website or online store content, as opposed to searches scoped to a specific section or category.

Stemming

The process by which a search engine reduces a word to its root in order to match all its variants. A search for "running" also returns results containing "run", "runs", and "runner".

Synonyms

Different words that mean the same thing: they connect the way customers search with the terms in the catalog, preventing zero-result pages caused by vocabulary mismatches.

T

Taxonomy

The hierarchical structure that organizes catalog products into categories, subcategories, and attributes so users can find and filter them.

Tokenization

The process of breaking text into its smallest meaningful units, words, fragments, or characters, so a search engine can analyze and match it.

Truncations

A search technique that expands a query to every word sharing its root, so searching 'run' also returns 'running', 'runner', and 'runs'.

Typeahead

A mechanism that surfaces search suggestions in real time as the user types, before they have finished forming their query. It reduces search effort by anticipating intent and guiding users toward terms the catalog actually contains.

U

URL parameters

The parameters that appear after the ? symbol in a URL and pass additional information to the server. For example, a URL like example.com/search?q=shoes&color=red uses query strings to filter or customize what the page returns, without pointing to a different page.

User intent

The underlying goal behind a search query, which may differ from the literal words a user types. Interpreting intent accurately is key to returning results that feel relevant rather than just technically matched.

V

Vector

A set of numerical values that represents the meaning of a term or product. Two vectors that are close together indicate related concepts, which is how a search engine connects "comfortable" with "ergonomic" even though the words look nothing alike.

Visual search

Technology that allows users to search for products using an image as input instead of text, identifying objects, colors, shapes, and patterns through computer vision.

Z

Zero-results management / no-results handling

A set of strategies and settings that prevent a site search from returning an empty results page when no exact matches are found. Common approaches include falling back to fuzzy matching or surfacing results for related terms.

Zero-results rate

The percentage of site searches that return no results to the user. It measures how often the search engine fails to find a match for a query, whether due to vocabulary mismatches, indexing gaps, or catalog limitations.