Core Web Vitals
Core Web Vitals are three metrics Google uses to measure a page’s experience: how fast the main content loads (LCP), how responsive the page is to clicks and taps (INP) and visual stability (CLS). They belong to technical SEO alongside topics such as Crawl budget, while E-E-A-T and Content cluster concern the quality and organisation of content.
Area: AI
|
Type: Concept
|
Stage: Optimisation
|
Intent: Understand
According to web.dev a page offers a good experience when the main content appears within 2.5 seconds, the response to an interaction arrives within 200 milliseconds and layout shift does not exceed 0.1. The thresholds are measured at the 75th percentile of visits. Mobile and desktop are assessed separately.
LCP (Largest Contentful Paint) measures when the largest element in the first screen appears. It is usually an image or a block of text. INP (Interaction to Next Paint) measures the delay between a click, a tap or a key press and the page’s visible response. CLS (Cumulative Layout Shift) measures how much elements move unexpectedly, like a button that drops just as you are about to tap it.
What counts for Google is the data collected from real visits in Chrome. You can read it in Search Console and PageSpeed Insights. Lab tests such as Lighthouse help find the causes, and they can give values different from those of users because they use a single device and a single connection.
Google uses Core Web Vitals among the signals of its ranking systems and points out that search still shows the most relevant content even when the page experience is poor. They matter most, then, when several pages answer the same question well, and for people who have to fill in a form or complete a purchase on a phone. To appear in Google’s AI Search answers the same SEO good practices apply, including a good page experience, also for the searches the system runs when it breaks down the user’s question (Query fan-out).
The most frequent causes are heavy or late-loading images in the first screen, which worsen LCP, third-party scripts such as chat and marketing tools, which slow down INP, and space not reserved for images, banners and fonts, which pushes CLS up. On a site like the one built for Metalsistem Lombardia, with more than 100 pages organised by application, sector, service and product, it is worth measuring by page type, because pages built on the same template tend to have the same problems.
In the SEO 2.0 & Digital PR service these checks are part of technical SEO. They are repeated after every major release of the site.
When a site is designed or rebuilt, because the thresholds can be written into the contract as a delivery requirement, and after every major release. They also become a priority when Search Console flags URLs that need improvement or are poor among the pages that bring the most enquiries.
The most common is chasing the Lighthouse score and neglecting real-user data, which is what Google looks at. It is also common to check the site only from a computer with a fast connection, to add third-party scripts after launch without checking the values again and to remove useful content to gain a few milliseconds.
They are three measures of a page’s experience. They show how long it takes to display the main content (LCP), how quickly it responds to a click or a tap (INP) and how stable it stays while loading (CLS).
When a site is being designed, to set the thresholds as a delivery requirement, and when the pages that matter most for enquiries and sales show up as slow on mobile in Search Console data.
With data from real visits, in Search Console and PageSpeed Insights, at the 75th percentile of visits. According to web.dev the threshold for a good experience is LCP within 2.5 seconds, INP within 200 milliseconds and CLS no higher than 0.1.
It explains why the challenge goes beyond collecting leads: they need to reach the right team with the right timing and priority.
To connect the term to the LU3G system, explore Marketing Automation and CRM for businesses | LU3G, MQL – Marketing Qualified Lead: what it is and how it works | LU3G, SQL – Sales Qualified Lead: what it is and how it works | LU3G.
Data lineage is the reconstruction of a piece of data’s path: where it originates, what transformations it goes through, which systems it passes through and where it is used. It is useful when a company needs to trust its reports, automations, integrations, CRM, data warehouse or AI systems.
In the DECODE system the topic is connected to Data pipeline, Data warehouse and Data layer, to clarify relationships, differences and practical applications.
RAG governance is the set of rules, controls and metrics that lets a company use Retrieval-Augmented Generation reliably. It is needed when an AI agent or assistant has to retrieve information from documents, knowledge bases, the CRM or internal data before generating an answer.
In the DECODE system the topic is connected to RAG, AI governance and Vector database, to clarify relationships, differences and practical applications.
Customer journey orchestration means coordinating data, touchpoints, CRM and automation to make the customer journey consistent. It goes beyond sending emails or building workflows: it means deciding what should happen when someone interacts with the company, which data drives the choice and which system needs updating.
In the DECODE system the topic is connected to Customer journey, Marketing automation and CRM automation, to clarify relationships, differences and practical applications.
Server-side tracking is an approach to digital measurement in which events and conversion data can pass through a controlled server environment before reaching analytics, advertising or CRM platforms.
In the DECODE system the topic is connected to Server-side tagging, Event tracking and Consent Mode v2, to clarify relationships, differences and practical applications.
AI Search Readiness measures how ready a website is to be read, understood and potentially assessed by AI-based search systems. It isn't the same thing as GEO: GEO works on optimisation for generative engines, while readiness checks whether the minimum conditions are in place to do it well.
In the DECODE system the topic is connected to AI Overview, GEO and Entity SEO, to clarify relationships, differences and practical applications.
Human-in-the-loop, often shortened to HITL, describes a process in which a person steps in at defined points in the work carried out by an artificial intelligence system. Its value lies in deciding where human judgement genuinely reduces risk and ambiguity, rather than in checking every output by hand. At LU3G it is connected to Zero-shot prompting, AI governance and ai agent to design automations that know when to proceed and when to ask for confirmation.
AI orchestration is the coordination of models, tools, data, rules and operational steps within a single AI flow. In practice, it determines which component steps in, with which inputs, checks, fallbacks and logging, so that a multi-step process never hinges on a single unverified answer. In the DECODE system, the concept is read together with AI tokens, Few-shot prompting, ai agent, because a useful decision comes from the relationships between strategy, implementation and measurement.
API integration is best read as a practical lever within cross-functional digital strategy: knowing what it indicates is not enough, you need to see which decision it makes clearer. At LU3G we connect it to Approval flow, Audit and Action to turn the term into a decision map.
In digital work, Workflow automation becomes interesting when it stops being a definition and starts shaping priorities. The connection with Calendar booking, Chat handoff and Whitepaper helps you see where to act and what to measure.
AI agent is useful when it helps interpret the core term ‘agent’ within the digital journey. Its value comes from the connection with Generative AI, AI Overview and Vector database, because a single entry explains little while a network of concepts guides action.
AI governance is the set of rules, responsibilities and controls an organisation uses to decide how it designs, buys, uses and monitors artificial intelligence systems. In practice, it requires an inventory of use cases, decision owners, risk criteria, data checks and procedures for incidents, model changes and suppliers. In the DECODE system, the concept is read together with Human-in-the-loop, MCP – Model Context Protocol, ai agent, because a useful decision comes from the relationships between strategy, implementation and measurement.
Automation is useful when it helps interpret the core idea ‘automation’ along the digital journey. Its value comes from the link with Churn rate, Contact management and Automation rule, because a single entry explains little while a network of concepts guides action.
Redirect 301 is a permanent HTTP redirect that signals a resource has moved for good to a new URL. In practice, it should point to the equivalent destination, without chains or loops; redirecting everything to the homepage confuses users and search engines and does not automatically preserve relevance. In the DECODE system, the concept is read together with Anchor text, 302 redirect, technical seo, because a useful decision comes from the relationships between strategy, implementation and measurement.
Visual hierarchy is the visual organisation that communicates priority and relationships through size, contrast, position, spacing and typography. In practice, it should guide scanning and action without hiding content; emphasis spread everywhere, inconsistent components and insufficient contrast cancel out the hierarchy. In the DECODE system, the concept is read together with TTFB, Headless CMS, core web vitals, because a useful decision comes from the relationships between strategy, implementation and measurement.
TTFB is the time between the request for a resource and the arrival of the first byte of the response. In practice, it includes network, negotiation and server work; caching, databases and the backend all play a part, but a low TTFB does not guarantee fast rendering or interaction. In the DECODE system the concept is read together with Heuristic evaluation, Visual hierarchy, core web vitals, because a useful decision comes from the relationships between strategy, implementation and measurement.
Share of voice is a brand's share of presence compared with competitors across a defined set of channels, keywords, media or conversations. In practice, source and scope must be stated; more visibility does not automatically mean preference, demand or market share. In the DECODE system, the concept is read alongside Outbound marketing, Brand salience, marketing, because a useful decision comes from the relationships between strategy, implementation and measurement.
A Display campaign is an advertising campaign that serves visual creatives across websites, apps and partner inventory to reach selected audiences and contexts. In practice, you need to tell reach, frequency, viewability and assisted conversions apart; placements, exclusions and creative quality reduce waste and unhelpful traffic. In the DECODE system, the concept is read together with CPL, PPC, google ads, because a useful decision comes from the relationships between strategy, implementation and measurement.