How Fast Should My Website Load, and What Is INP?

You did a speed test and got 98, and someone gave you a report stating your site is failing. Both are right. So how fast should my website load?

The answer is that a high lab score doesn’t always mean that it is what real visitors experience. The site can be optimized in a test environment but not on real devices and connections. To see how fast it is, look at lab data and real user data, particularly Google’s Core Web Vitals.

In this Guide

• The three thresholds
• What Interaction to Next Paint is
• Why test and report disagree
• What fixes each failure

 

How fast should my website load?

Three thresholds measured during real mobile visits: Largest Contentful Paint (LCP) < 2.5 seconds, Cumulative Layout Shift (CLS) < 0.1, and Interaction to Next Paint (INP) < 200 milliseconds.

The table lists the three thresholds as defined by Google on the web.dev Core Web Vitals documentation, the document that Google maintains and updates as the thresholds change.

The Core Web Vitals 2026 evaluation is based on the 75th percentile of visits, which means a site will be evaluated based on its slowest 25% of visits. Older phones on weak connections make the decision, and two of the three metrics are not load times.

What is Interaction to Next Paint?

INP is the time a page takes to visibly react to a tap or click, and it is considered good if it is less than 200 milliseconds.

What is a good INP score? Under 200 milliseconds. A tap is queued, the browser has one main thread, and if JavaScript execution keeps it busy with a long task, the response waits.

The visitor clicks on the menu, but nothing happens, and they click again.

INP is recorded across the whole visit, so a page that loads fast can still fail under chat widgets, popups, heat maps, review carousels and unaudited tag managers a third-party script problem that hits stores where every tap is a step to checkout hardest of all, because shoppers tap dozens of times in a single visit.

The most frequent one that we see is marketing scripts that have been installed over several years that no one ever removed.

Why does my speed test say 98 and my report say failing?

Because they measure different things. A lab test is one simulated load on a fast machine with no one interacting, and the report that decides is created based on thousands of real visitors.

The honest answer to website speed test accuracy is that the test was accurate on the one load that it simulated. A page can get close to 100% on a test that did not test anything and yet not pass for a person who did. Field data decides whether you pass.

The field window is a rolling 28 days, so a fix shows in weeks, not hours, and a low-traffic site may have too few visits for a field report at all. Use lab tools to diagnose and your own phone to judge.

Does speed affect my rankings?

While speed is a factor in how Google evaluates pages, it can make the difference between two similar results, but not if the page doesn’t answer the question.

The only citable basis for what speed does to rankings is the position Google states on experience signals in its page experience documentation, which is worth reading before you pay anyone to chase a score.

There’s no ranking strategy involved in passing Core Web Vitals, and a quarter of the time spent on trying to pass this test rather than publish anything useful is the cheapest feeling of progress. Look at where performance sits in search work before spending anything on it, so speed work supports useful pages instead of replacing them.

Does speed actually affect enquiries?

Yes, and more directly than it affects rankings, as a visitor who waits for a slow page or clicks on a button that doesn’t respond will leave, and the enquiry will go to the person who loaded.

On your phone, using mobile data, attempt to make an enquiry on your site. Time it.

Speed is one stage among many, and our guide on traffic without enquiries covers the rest of the conversion funnel from the first visit through to the moment someone calls you or leaves.

A landing page needs less to load than a full site, and pages built light on purpose leave little to slow down a visitor who is ready to call.

Why does the layout jump around while the page loads?

Cumulative Layout Shift occurs when content is being loaded without allocating space for it; therefore, when a visitor scrolls down to read the content, the image, banner, or font comes in late and pushes all the content down.

A visitor taps a button, the page shifts, and they hit something else. Predictable layout has nothing to do with a preference; it’s an accessibility need, and WCAG guidelines, published by the W3C, have established that for every public website.

What do I measure, and where?

Record the three numbers on the same day of the month; use the field report for the verdict and a lab tool for the diagnosis; check both on the mobile, and then record the three numbers on the same day each month to get a trend.

Write the values down with a date before any work starts. Without a baseline, you cannot prove a fix worked.

Why does my site fail Core Web Vitals? The field report names the metric, a lab tool finds the cause, and our audit offers a measured starting point for deciding what to fix first and in what order, before anyone is paid.

What fixes each one, and who does it?

Largest Contentful Paint is typically images and hosting; Cumulative Layout Shift is typically missing dimensions and injected content, and Interaction to Next Paint is typically too much JavaScript, the most difficult of the three.

Fix images first, because it is cheap and visible. Treat INP as a project, because removing scripts means deciding what the business stops using.

When you want a developer to take over the list, our technical team delivers performance work with the metrics attached, so every fix is tied to a number you recorded before the work began.

On WordPress, theme and plugin choices decide most of the weight, so a platform kept lean starts by removing every plugin that no longer earns its place on the site.

What if the theme itself is the problem?

Some sites cannot be made fast because the weight is in a page-builder theme and a stack of plugins doing overlapping jobs, and then performance work is a rebuild rather than an optimization.

Remove one heavy element and retest; if the number holds, the weight is structural.

If the weight in your theme is structural, our redesign team can tell you when performance needs a rebuild instead of another round of plugin tuning that never reaches the real weight.

For a new build, our web design team treats speed as a written requirement and delivers builds where speed is a requirement with the three metrics agreed in writing before any design work starts.

A rebuild can put your rankings at risk, so read our guide on rebuilding without losing what ranks before you commit to a new site and fix a launch date.

These thresholds are a floor, not a target, and a site that passes all three can still be unpleasant to use.

 

How fast should my website load? Setting a realistic target

Setting a realistic target takes eight steps:

1. Get the field report and write down the three numbers with today’s date.
2. Check which is failing.
3. Use a lab tool only to find the cause.
4. Fix images and hosting first.
5. Reserve space for late-loading content.
6. Remove every unused third-party script.
7. Retest in a month, not tomorrow.
8. If removing things changes nothing, the theme is the problem.

The metric most sites fail in 2026 measures whether the page answers when somebody touches it, a better definition of fast than load time.

We are a digital marketing agency at 13121 Louetta Rd, Cypress, Texas. Pink Socials can read your three numbers with you and name the first fix.