All notes

Website

Website Slow on Phones? Fix Core Web Vitals Starting with Your Most Important Pages

Translated from the Indonesian original.

A phone-sized wireframe card with a red stopwatch mark beside the main image, and a red pen.

The way to fix a website that is slow on phones is to measure the visitor experience, find the cause of the delay, then fix the pages closest to your business goals. Start with the service, product, or form pages that get used most often. A speed score is a diagnostic tool; visitors still need to be able to read and act comfortably.

The problem is not always the same. Some websites are slow to show content, some stutter when a button is tapped, and some make buttons shift when an image finishes loading. Each of the three needs a different fix.

Why does mobile performance still matter in 2026?

The Web Almanac 2025, published in 2026, reported that 48% of mobile websites in its analysis had good Core Web Vitals. The desktop figure reached 56%. This data shows there is still room for improvement, but it should not be taken to represent every business website in Indonesia. Source: HTTP Archive, Performance 2025.

For service businesses, look for friction along the customer journey: an offer page that opens slowly, a form that does not respond, or a consultation button that is hard to tap. Measure these on your own website before estimating their effect on sales.

What are Core Web Vitals?

Core Web Vitals are three measures of user experience: how quickly the main content appears, how responsive interactions are, and how stable the layout is. The thresholds for the good category are:

  • LCP, or Largest Contentful Paint: 2.5 seconds at most.
  • INP, or Interaction to Next Paint: 200 milliseconds at most.
  • CLS, or Cumulative Layout Shift: 0.1 at most.

The assessment uses the 75th percentile of user experience data, split by device. Do not draw conclusions from a single test on an office laptop. Source: web.dev, Web Vitals.

LCP: does the main content appear quickly?

Imagine a prospective customer opening a service page. The logo appears, but the large image and the offer headline only show up a moment later. This is the kind of experience to investigate through LCP.

Check the main image, server response time, and resources that hold back page rendering. Google recommends not lazy loading the image that is the LCP element. Techniques meant for images further down the page are not necessarily right for the main image. Source: web.dev, Optimize LCP.

INP: does the website respond when people use it?

A page can look fully loaded while the menu feels stuck or the form button is slow to react. Check the JavaScript work, including third-party scripts, that runs while the interaction is happening.

The INP guide covers reducing heavy work on the main thread so the browser can respond to interactions. In an audit, tie technical findings to actions such as opening the menu, choosing a package, or submitting a form. Source: web.dev, Optimize INP.

CLS: do elements shift unexpectedly?

Layout shifts can happen when images or extra elements appear without space reserved for them. Visitors may end up tapping the wrong thing.

Set image dimensions and reserve space for dynamic elements. Also check banners, fonts, and widgets that appear after the page opens. Source: web.dev, Optimize CLS.

How to audit a slow website without guessing the cause

Start with a list of priority URLs. For a service website, pick the homepage, one main service page, and the contact page. For an online store, add product pages and the most commonly used purchase journey.

Run the following checks:

  • Open PageSpeed Insights for each URL and check the mobile data.
  • Separate real-user data from lab test results.
  • Note the failing metrics along with the related elements or interactions.
  • Walk through the customer journey on a real phone.
  • Match the problems to the pages that generate enquiries or transactions.
  • Choose one group of fixes and measure again after they are applied.

If real-user data is not available yet, use lab testing for an initial diagnosis. Do not call it proof that every visitor is already getting a good experience.

Which fixes should come first?

Use three questions: how severe is the friction, how many customer journeys does it affect, and how much effort does the fix take?

An illustrative example: the homepage scores lower than a service page, but the form submit button on the service page does not respond. Fixing the form can take priority because it directly blocks an important action. After that, deal with the heavy resources on the homepage.

Another example: large images across all product pages are far bigger than the screen needs. Fixing the image template can be more useful than optimizing a single page that is rarely visited.

Avoid removing features just to raise the score. Contact buttons, service information, and accessibility must still work after optimization.

How do you measure the results?

Compare technical metrics and business behavior over comparable periods. Note any changes to design, campaigns, or traffic sources, because those factors can affect the results.

Monitor:

  • Changes in LCP, INP, and CLS on the same device type.
  • Successful form submissions, not just button clicks.
  • The share of service page visits that turn into enquiries.
  • User complaints about the menu, the display, or forms.
  • Performance of other templates after shared changes are applied.

Core Web Vitals are used by Google's ranking systems, but good results do not guarantee the top position. Content relevance and searcher needs still matter. Source: Google, Page Experience.

Frequently asked questions

Is a PageSpeed score of 100 required?

No. Make diagnosing problems and the user experience your goal. A perfect score is no substitute for testing whether visitors can actually get done what they came for.

Will moving to a new host definitely make the website fast?

Not always. If the main problem is images, JavaScript, or layout, moving hosts alone will not solve it. Identify the cause before changing infrastructure.

Does a fast website automatically increase conversions?

Not automatically. Speed can reduce friction, but results also depend on the offer, pricing, traffic relevance, and how easy it is to contact the business. Use your own website data to judge the impact.

Make performance part of building the website

Performance is easier to maintain when design, images, features, and integrations are considered together. Once the website is live, test again whenever there is a major change to the theme, plugins, or widgets.

Talk through your business website needs with Jelas Studio. Website development, SEO, and GEO can be planned with your prospective customers' experience in mind from the start.

Asking the same question about your website?Start a project