Website speed optimization and Core Web Vitals

Find the delay before choosing the fix

A slow page, a menu that hesitates and a button that moves under your finger need different investigations. We start with the pages and actions that matter: opening a service page, filtering products or submitting a form.

Core Web Vitals describe three parts of that experience. LCP, Largest Contentful Paint, measures loading performance. INP, Interaction to Next Paint, measures responsiveness to interactions. CLS, Cumulative Layout Shift, measures visual stability. We identify which problem is present before proposing changes.

The first deliverable is a prioritized diagnosis: affected pages, evidence, proposed fixes and anything that needs access or a third-party decision. That gives you a scope you can review before implementation.

Typical speed jobs

The work can be a focused repair or a wider pass across shared templates. We check the cause of each issue before applying a fix; installing another plugin is not a diagnosis. Typical requests include:

• Investigate a late-loading main image: inspect its size, format, request timing and loading instructions, then adjust the image delivery.
• Trace a slow menu, filter or cart interaction and reduce the JavaScript work that delays the response.
• Reserve space for images, embeds and banners so content does not unexpectedly move as they appear.
• Review font files, loading behavior and fallback styles when text appears late or shifts during loading.
• Audit duplicate tracking tags and unused app scripts, agreeing what can be removed with the people responsible for them.
• Investigate slow server responses and review caching rules, with explicit checks for carts, accounts and personalized content.

Work within your platform

For Shopify speed work, we inspect the theme, product media, app embeds and scripts around the storefront. For WordPress speed work, we review the theme, plugins, media and hosting configuration. WooCommerce work includes checking shopping flows after changes.

We also take on Webflow, Squarespace, Wix and custom websites. On hosted builders, we scope changes around the controls available in your account. If a suspected issue sits inside platform-managed code, we identify that dependency rather than quote for access we do not have.

For Next.js, React and TypeScript web apps, the investigation can include client-side code, rendering and API requests. We use PageSpeed Insights, Lighthouse and Chrome DevTools to inspect loading and interaction behavior. Existing Search Console data helps select affected page groups when it is available.

From first email to handoff

Send the affected pages, your platform, recent changes and any reports you already have. Tell us which features must stay, who manages the site and whether there is a staging environment. A description such as 'the mobile filter freezes after choosing a size' is useful. Send account invitations when needed; keep passwords out of the initial email.

We agree on the pages, interactions and test conditions in scope. Where the cause is unclear, paid discovery can come first. Before editing, we establish a backup or rollback route and use a staging site, theme copy or code branch where available.

After implementation, we repeat comparable measurements and check the affected forms, navigation and shopping flows. The handoff includes the changes, before-and-after evidence, remaining dependencies and rollback instructions. We agree on who deploys and who checks the live site.

Measure the experience, explain the limits

PageSpeed Insights combines lab testing with real visitor data where enough data is available. A lab score is a diagnostic signal; it does not establish that every visitor has a good experience. We keep the measurement type visible in the report and compare tests under similar conditions.

A standard Lighthouse page-load test does not measure INP. We investigate actual interactions separately and review field measurements when available. Google's field reporting uses a trailing 28-day window, so it will not immediately reflect only the updated site. Missing field data is an evidence gap, not a passing result.

We agree on measurable acceptance criteria after the initial review. Those might cover a specific loading bottleneck, a reproducible interaction delay or an unexpected layout shift.

A focused repair, with honest proof

This service suits owners with a slow site, marketers investigating a report and agencies needing help with a defined performance task. It is less suited to a brief that requires a predetermined score before anyone can inspect the site, or rules out changes to the components under investigation.

We design, build and maintain our own products end to end, including nine live App Store apps and this hand-built website. These are owned products, and we have no client performance case studies to present here.

We take on a small number of projects at a time, so you work directly with the people doing the work. Engagements can be fixed-price projects, hourly fixes or agency sub-projects, with ongoing maintenance available through a monthly retainer.