Page speed
The broader performance of loading, rendering, interacting with, and navigating a website.
Core Web Vitals · Website Performance · Technical SEO · Frontend Optimization
Improve how quickly your website loads, responds to user interactions, and maintains visual stability across mobile and desktop devices.
I diagnose and resolve performance problems involving servers, rendering, JavaScript, CSS, images, fonts, third-party scripts, caching, CDNs, databases, frameworks, and website templates.
Service overview
Page speed optimization is the process of reducing the time, data transfer, browser work, server processing, and visual disruption required to display and use a website.
It can involve improvements across several technical layers. Origin server work may cover hosting, server response, application processing, database queries, API responses, and redirects. Network and delivery work may cover CDN, compression, HTTP caching, browser caching, DNS, TLS, and geographic delivery. Frontend work may cover HTML, CSS, JavaScript, images, fonts, components, rendering, and hydration. Third-party services may include analytics, advertising, consent tools, chat systems, video, booking, payment, and embedded content.
A complete optimization project identifies which layer is causing the delay before changing the website.
Service overview
Core Web Vitals are Google’s current field-based metrics for loading performance, responsiveness, and visual stability. The current metrics and recommended “good” thresholds are Largest Contentful Paint at 2.5 seconds or less, Interaction to Next Paint at 200 milliseconds or less, and Cumulative Layout Shift at 0.1 or less. These thresholds are evaluated at the 75th percentile, meaning that the target should be met for at least 75% of measured page visits, with mobile and desktop considered separately.
LCP measures how quickly the main visible content is rendered. The LCP element may be a hero image, banner, large heading, content block, product image, background image, or video poster.
INP measures the responsiveness of a page across user interactions such as clicks, taps, keyboard input, menu opening, filter selection, form interaction, and product selection. INP includes input delay, event-processing time, and the delay before the browser displays the next visual response.
CLS measures unexpected movement of visible page content. Common causes include images without dimensions, advertisements without reserved space, embeds and iframes, dynamically inserted content, font changes, banners, and notices. Google recommends reserving predictable space for images and similar elements to reduce layout movement.
Related services
Natural next steps that connect this offering with related work on the same roadmap.
This service sits inside a wider SEO and digital growth system. Strong next steps usually include New Website Development and Server-Side Tracking Implementation Services.
Depending on priorities, also consider SEO Website Migration Services and Server Setup & DevOps Services.
Potential outcomes
A properly implemented performance project can improve the following—actual improvements depend on the website platform, hosting, third-party services, source code, content, and available development access:
Loading speed across important page templates
Core Web Vitals performance
Mobile usability on slower devices and networks
Responsiveness of menus, forms, filters, carts, and interactive components
Stability of images, embeds, banners, and dynamically loaded elements
Server response times and origin efficiency
Image, font, JavaScript, and CSS delivery
Cache efficiency and CDN usage
Performance monitoring and release quality control
The technical foundation supporting SEO, advertising, and conversion journeys
Service components
This helps prioritize pages that matter to users and the business.
Performance problems often differ significantly between templates.
Field data reflects real visitors, while laboratory tools provide controlled diagnostics that can help reproduce and investigate problems.
Mobile performance problems should not be assumed from desktop results.
The audit separates good, needs improvement, poor, insufficient data, lab-only findings, and field-confirmed findings.
Google’s LCP guidance recommends breaking the metric into its contributing delays rather than treating it as one unexplained number.
The analysis may identify main-thread blocking, long JavaScript tasks, expensive event handlers, rendering delays, layout recalculation, large DOM structures, third-party script work, and hydration delays.
The analysis distinguishes the element that moved from the element that caused the movement.
FCP supports diagnostics but is not one of the three current Core Web Vitals.
A frontend-only change cannot correct every TTFB problem.
Recommendations are adapted to the hosting platform.
Especially relevant for WordPress, WooCommerce, ecommerce, marketplaces, booking systems, and custom applications.
I review whether APIs should be cached, parallelized, deferred, combined, preloaded, moved server-side, removed from the critical path, or given fallback behaviour.
The strategy affects initial HTML, LCP, JavaScript requirements, hydration, caching, crawling, server cost, and data freshness.
Client components are used only where browser interactivity is required.
The objective is to reduce browser work without breaking application behaviour.
Plugins are not installed simply because they advertise performance optimization.
Cart, account, and checkout pages require different caching rules from public product content.
Scripts may be classified as essential, important but delayable, page-specific, conditional, unused, replaceable, or removable.
Bundle reduction should be validated through actual build output.
Possible improvements include breaking work into smaller tasks, yielding to the browser, deferring non-critical work, moving suitable processing to workers, and reducing the work performed.
Improvements may include reducing callback work, avoiding unnecessary state updates, debouncing, throttling, passive listeners, event delegation, separating visual updates, and removing duplicate listeners.
Each script is evaluated by business value, loading cost, execution cost, page coverage, consent dependency, and alternative implementation.
GTM does not automatically make third-party scripts lightweight.
Possible improvements include removing unused styles, splitting CSS by route, reducing framework output, loading page-specific styles, simplifying selectors, and minifying production files.
Critical CSS should not become an unmaintainable manually copied block.
Large images are prioritized according to their page position and business importance.
Modern formats do not replace correct image sizing. Fallback behaviour, image quality, transparency, browser support, and processing cost are considered.
Visitors should not download a desktop-size image for a small mobile container without a justified reason.
The implementation depends on how the image is rendered.
Above-the-fold LCP content should not normally be delayed through generic lazy-loading rules.
Unnecessary font families and weights are removed where possible.
This may use width and height attributes, aspect ratios, minimum heights, skeleton layouts, stable placeholders, and predictable component dimensions.
I test header behaviour throughout loading and interaction.
The implementation must still respect the customer’s approved consent requirements.
Possible improvements include facades, poster images, click-to-load behaviour, lazy loading, reserved space, conditional loading, simplified embeds, and linking externally where appropriate.
An unnecessarily large and complex DOM can increase style, layout, rendering, and interaction work.
Compression reduces transfer size but does not replace removing unnecessary resources.
Frequently changing HTML requires different rules from versioned static assets.
Possible environments include Cloudflare, hosting-platform CDN, ecommerce CDN, application platform, or a customer-approved provider.
This can be coordinated with the dedicated Cloudflare Setup and Optimization service.
Personalized and transaction-related content requires careful exclusion.
Every unnecessary request or redirect adds network work.
Preloading too many resources can compete with genuinely important requests.
Incorrect service-worker caching can preserve broken or outdated website versions.
Budgets may vary by page template.
Google’s web-vitals library can collect LCP, INP, and CLS directly for custom field measurement.
Automated testing should use a representative URL set rather than every URL without prioritization.
A website can become slow again after a successful optimization project when future changes are not monitored.
Lab improvements may appear immediately, while field datasets require new real-user visits before their reporting reflects the change.
A faster page that loses important functionality is not a successful optimization.
I verify that performance changes do not negatively affect crawlability and search controls.
I document findings, changes, results, limitations, and next steps.
Recommendations are written as implementation tasks rather than generic audit statements.
Unexpected changes are investigated against the production release.
Ongoing support keeps performance stable as the website evolves.
Search terminology
Page speed
The broader performance of loading, rendering, interacting with, and navigating a website.
Core Web Vitals
The field metrics LCP, INP, and CLS—representing loading, responsiveness, and visual stability.
PageSpeed Insights
A Google tool that evaluates a page on mobile and desktop and provides available real-user information and Lighthouse laboratory diagnostics.
Lighthouse
A controlled laboratory auditing tool that helps identify reproducible technical opportunities but does not represent every real visitor.
CrUX
An aggregated dataset based on eligible Chrome user experiences that supports real-world Core Web Vitals reporting.
Performance score
The Lighthouse performance score is a weighted laboratory score. It is not the same as passing the Core Web Vitals field assessment.
Ideal clients
Sites with poor or “needs improvement” LCP, INP, or CLS results.
Businesses whose pages load or respond poorly on mobile devices and slower networks.
Platforms with heavy product pages, filters, carts, checkout systems, scripts, and dynamic content.
Sites affected by themes, page builders, plugins, database load, scripts, and incorrect cache configuration.
Websites requiring bundle reduction, rendering review, hydration optimization, server-component planning, and interaction improvements.
Businesses relying on landing pages, forms, calls, messages, bookings, and advertising traffic.
Sites managing large images, advertisements, recommendations, videos, consent systems, and article templates.
Tour, hotel, transportation, and accommodation platforms with large visual assets and external booking tools.
Teams that want performance requirements included before development and launch.
Teams requiring prioritized tasks, evidence, acceptance criteria, and post-deployment validation.
What you receive
Concrete diagnostics, implementation work, technical requirements, validation, and monitoring—scoped to your website and platform.
Delivery process
I review the website, business-critical pages, devices, countries, traffic, conversions, hosting, and known performance problems.
I collect available real-user information and run controlled tests across representative page templates.
I identify whether the main problems come from the origin server, rendering, JavaScript, CSS, images, fonts, third parties, caching, or application architecture.
Issues are ranked according to user impact, page coverage, commercial importance, Core Web Vitals impact, technical risk, and implementation effort.
I apply approved frontend, backend, server, CDN, caching, and configuration improvements.
I verify that forms, bookings, ecommerce, navigation, analytics, consent, and integrations continue to work.
I repeat mobile and desktop tests and compare the results with the original baseline.
Approved changes are deployed through the customer’s development and release process.
I review production performance, errors, Core Web Vitals trends, and important customer journeys.
I document the changes, results, limitations, budgets, and ongoing monitoring requirements.
Specialist
I combine performance optimization with technical SEO, website development, Next.js, WordPress, analytics, tracking, Cloudflare, website migrations, AEO, GEO, and conversion measurement.
This allows me to evaluate both: Why the website is slow Which changes are safe for its business and search requirements
I review: Which page templates affect the most users Which resources delay important content Which JavaScript blocks interactions Which elements create layout shifts Whether the server or frontend is the main bottleneck Whether caching is safe for each page type Which third-party scripts are commercially necessary Whether a laboratory result reflects field performance Whether optimization changes affect crawling or tracking How future releases should be controlled
Rather than sending only a PageSpeed screenshot, I document: The technical cause The affected pages The user impact The recommended fix The implementation requirements The acceptance criteria The test method The measured result The remaining limitation The maintenance responsibility
I can support the process from performance diagnostics and prioritization through implementation, testing, deployment, monitoring, and long-term performance governance.
Common questions
Page speed optimization improves how quickly a website responds, loads its important content, becomes interactive, and maintains a stable visual layout.
The current Core Web Vitals are Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift. The recommended good thresholds are LCP at 2.5 seconds or less, INP at 200 milliseconds or less, and CLS at 0.1 or less.
No. A score of 100 is not required for a useful, successful, or search-visible website. The goal is to improve real-user experience and important website journeys without breaking business functionality.
The Lighthouse score is based on a controlled laboratory simulation. PageSpeed Insights may also display available field information from CrUX, which reflects eligible real-user experiences.
Results may vary because of server response, network conditions, test environment, third-party scripts, cache state, dynamic content, advertising, API responses, CPU simulation, and temporary platform behaviour. Several tests and field information are more useful than one isolated score.
Mobile testing commonly uses more limited processing and network conditions. The mobile version may also load different navigation, images, components, sticky buttons, advertising, widgets, and responsive layouts.
No. The achievable result depends on platform, hosting, website design, required functionality, third-party services, advertising, booking systems, ecommerce, consent tools, and content. I focus on meaningful, sustainable improvements rather than guaranteeing a test score.
Google recommends good Core Web Vitals as part of a strong page experience, but explains that there is no single page-experience signal and good scores do not guarantee top rankings.
A faster and more responsive experience may reduce friction, but conversion results also depend on offer, trust, content, pricing, design, audience, traffic quality, and sales process. Conversion improvements cannot be guaranteed from speed work alone.
Yes. WordPress optimization may include plugins, theme, page builder, PHP, database, images, fonts, scripts, caching, CDN, and hosting.
Yes. WooCommerce work may include product templates, category queries, cart, checkout, sessions, database, payment scripts, marketing tags, and cache exclusions.
Yes. Next.js optimization may include server and client components, rendering strategy, image delivery, fonts, data fetching, bundles, hydration, caching, middleware, and third-party scripts.
Yes. I can review bundle size, rendering, hydration, change detection, event handling, route loading, DOM complexity, and API dependencies.
Cloudflare may improve eligible asset delivery, caching, compression, geographic latency, and origin load. It cannot automatically repair slow application code, database queries, or excessive client-side JavaScript.
Yes. Possible improvements include removing unused code, splitting bundles, dynamic imports, replacing heavy dependencies, delaying third-party scripts, reducing event-handler work, and moving appropriate logic server-side.
Yes. Image work may include resizing, compression, WebP or AVIF, responsive variants, lazy loading, CDN delivery, LCP prioritization, and dimensions and aspect ratios.
Yes. Common fixes include reserving space for images, embeds, forms, advertisements, fonts, and dynamic content. Some layout movement may require changes to the design or third-party widget.
Yes. INP optimization may involve reducing long tasks, script evaluation, event-handler work, rendering complexity, layout work, and third-party JavaScript.
Yes, when the required hosting and application access is available. Possible work includes page caching, object caching, database queries, APIs, server resources, hosting location, rendering, middleware, and CDN.
No. A caching plugin may help selected WordPress pages, but it cannot automatically resolve heavy JavaScript, large images, layout shifts, third-party scripts, database problems, slow APIs, poor architecture, or checkout complexity.
Yes. Incorrect caching may affect user accounts, carts, checkout, prices, currency, bookings, personalized pages, forms, and dynamic data. Cache exclusions and validation are essential.
I identify their performance cost and business purpose. The customer decides whether an important advertising, analytics, chat, booking, or compliance service can be removed or replaced.
Laboratory tools may show changes shortly after deployment. Search Console Core Web Vitals relies on aggregated field information, so it requires sufficient new user data before the reporting reflects updated performance.
Some URLs and origins may not have enough eligible field data. Laboratory testing and custom real-user monitoring can still be used for diagnosis and validation.
I normally test representative templates and high-priority URLs. Testing every individual URL may be unnecessary when many pages share the same code and layout.
Yes. Monitoring may include Core Web Vitals, PageSpeed tests, Lighthouse, template sampling, bundle sizes, page weight, server response, errors, and performance budgets.
The service can include audit, developer-ready recommendations, direct implementation, quality assurance, and post-launch monitoring. The exact scope depends on platform access and the customer’s development workflow.
The initial process normally requires website URL, main page templates, business-priority pages, Search Console access, analytics access, CMS or repository access, hosting access where relevant, Cloudflare or CDN access where relevant, GTM access, staging environment, recent deployment information, known performance problems, required third-party services, and a development contact where relevant.