Why Is My Website Slow? Core Web Vitals, Rankings and How to Fix It

Most people who ask "why is my website slow" start by running a speed test, seeing a score like 41 out of 100, and installing the first caching plugin that promises to fix it. A week later the score is 46 and the site feels exactly the same. The score was never the problem. It is a summary, and summaries do not tell you what to fix.
The short answer: a website is usually slow for one of four reasons. The server takes too long to send the first byte of the page, the main image or headline is too big or discovered too late, JavaScript (much of it from third-party tools) keeps the browser too busy to respond, or the layout keeps moving as things load. Google measures the effects of these through three Core Web Vitals: Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS). Find out which of the three is failing on real visits, trace it back to its cause, and fix that cause first. Everything below shows you how.
This post is the technical diagnosis. For what a slow page costs in leads and sales, see why website speed impacts revenue.
What Core Web Vitals measure: LCP, INP and CLS explained
Core Web Vitals are three measurements of what a page feels like to a real person. Each covers a different frustration, so a page can pass one and fail another.
Largest Contentful Paint (LCP): when the main thing appears
LCP is the time from the moment someone starts loading the page until the largest piece of content in view, usually the hero image, a product photo or the main headline, has rendered. It is the closest metric to the feeling of "has this page loaded yet".
Interaction to Next Paint (INP): how quickly the page reacts
INP measures the delay between a click, tap or key press and the next visible screen update. Scrolling and hovering do not count. It reports roughly the slowest interaction of the visit (ignoring rare outliers), so one sluggish menu or "add to cart" button can drag it down. INP replaced the older First Input Delay metric as a Core Web Vital on 12 March 2024, which is why older guides still talk about FID.
Cumulative Layout Shift (CLS): how much the page jumps
CLS scores how much visible content moves unexpectedly while the page is being used. You know the feeling: you go to tap a link and an image or banner loads above it, pushing your target down. Movement that happens within half a second of the user's own click or tap is excluded, because the user expected it.
Good Core Web Vitals scores: the official thresholds
These are the thresholds Google publishes on web.dev. Our website speed test rates results against the same three bands.
| Metric | Good | Needs improvement | Poor |
|---|---|---|---|
| LCP (loading) | 2.5 seconds or less | Over 2.5 s, up to 4 s | Over 4 seconds |
| INP (responsiveness) | 200 milliseconds or less | Over 200 ms, up to 500 ms | Over 500 milliseconds |
| CLS (visual stability) | 0.1 or less | Over 0.1, up to 0.25 | Over 0.25 |
Three details change how you read these numbers:
- The 75th-percentile rule. Google does not judge your page on its best load or its average one. It takes the value that 75% of real page views meet or beat. If a quarter of your visitors are on slow phones with weak signal, they decide your score.
- All three must pass. A page passes the Core Web Vitals assessment only when LCP, INP and CLS are all in the good band at the 75th percentile.
- Mobile and desktop are judged separately. Plenty of sites pass on desktop and fail on mobile, and mobile is where most of them are weakest.
Field data versus lab data
Field data comes from real Chrome users visiting your page, collected in the Chrome UX Report and summarised over the previous 28 days. Lab data is one simulated load on a test device with fixed network conditions. The PageSpeed score out of 100 comes from the lab run. Core Web Vitals, as Google uses them, come from the field.
So a lab score of 55 with passing field data beats a 90 with failing field data. INP cannot be measured in the lab at all, because nobody clicks anything in a simulated load; lab tools show Total Blocking Time as the closest stand-in.
Are Core Web Vitals a ranking factor? What Google actually says
Yes, but it is easy to overstate. Google's Search Central documentation says Core Web Vitals are used by its ranking systems and highly recommends achieving good ones. It is just as clear on the limits:
- Good results in Search Console's Core Web Vitals report or in third-party tools do not guarantee your pages will rank at the top.
- Google always seeks to show the most relevant content, even if the page experience is sub-par.
- There is no single "page experience" signal, and Google evaluates pages mostly one at a time, with some site-wide assessments.
Our reading: if your content does not answer the search well, speed will not save it. Where several pages answer it equally well, a poor experience is a disadvantage you do not need. Be wary of anyone who promises that fixing Core Web Vitals alone will lift rankings. The stronger reason to fix them is that the same problems that fail the metrics are the ones that make visitors leave.
If rankings dropped suddenly, rule out simpler causes first. Our free website health check takes about 30 seconds and flags an accidental noindex tag, a missing canonical link or an unreachable sitemap, any of which hurts far more than a slow LCP.
Why is my website slow? The real causes, by metric
Each failing metric points to a short list of causes. Start with the metric, not a generic checklist.
When LCP fails
Google splits LCP into four parts, and knowing which one is long tells you exactly where to look:
- Time to First Byte. How long the server takes to start sending the page. Long here means slow hosting, no page caching, heavy database queries or a chain of redirects.
- Resource load delay. The gap before the browser starts fetching the main image, long when it is a CSS background, injected by a slider script or set to lazy-load. Google's guidance is blunt: never lazy-load your LCP image.
- Resource load duration. The download itself. A 3 MB hero photo served at full desktop size to a phone lives here. Google's own research says this part is usually not the main bottleneck, which is why compressing images alone often disappoints.
- Element render delay. The image has arrived but cannot be painted yet, because blocking stylesheets, fonts or long JavaScript tasks are in the way.
When INP fails
Poor INP is almost always too much JavaScript on the main thread. The usual culprits are chat widgets, tracking pixels, A/B testing scripts, consent banners and heavy page builders or themes that ship code the page never uses. Each interaction has three phases (waiting for the browser to be free, running the code, then drawing the result), and busy third-party scripts mostly hurt the first.
When CLS fails
Per web.dev, the common causes are images and videos without set dimensions, ads and embeds that resize themselves, content injected after the page starts rendering (cookie bars, promo strips, "you might also like" blocks), and web fonts that render at a different size from the fallback text.
Quick symptom finder
| What you notice | Likely metric | Where we would look first |
|---|---|---|
| Blank white screen for a few seconds | LCP | Server response time, hosting, caching, redirects |
| Text appears but the hero image comes in last | LCP | Lazy-loaded or CSS background hero, image size |
| Page looks ready but buttons ignore taps | INP | Third-party scripts, page builder JavaScript |
| Content jumps as you start reading | CLS | Image dimensions, banners, ads, font swaps |
How to fix a slow website, step by step
This is the order we work in on client sites. It stops you fixing things that were never the bottleneck.
- Test the pages that earn money, on mobile. Your home page, a key service or product page, a blog post and the checkout or contact page. Fix a template and every page using it improves. Our website speed test runs each page through Google PageSpeed Insights and shows LCP, INP and CLS with a good, average or poor rating, plus the biggest savings Google found. It takes about a minute, and you can have the full report emailed to you for your developer.
- Check if you are looking at field or lab data. The speed test tells you which one it is showing. If real-visitor data exists, trust it for the pass or fail decision and use the lab run for clues. If there is only lab data (common on smaller sites), treat Total Blocking Time as your INP warning light.
- Pick the worst metric and ignore the rest for now. Fix anything rated poor before anything rated "needs improvement", and do not touch the overall score yet.
- Trace it to its cause using the lists above. For LCP, find which of the four parts is long. For INP, list every third-party script on the page and ask who still uses it. For CLS, watch the page load on a phone and note what moves.
- Apply the specific fix. For LCP: put the main image in the HTML, add fetchpriority="high", remove lazy-loading from it, serve it at the size it is displayed in WebP or AVIF, and fix server response with page caching or better hosting. For INP: remove unused scripts, load chat and marketing tools only on the pages that need them, and delay the rest until after the page is usable. For CLS: give every image and embed a width and height, and reserve space for banners before they load.
- Retest the same pages in the lab. The lab run should improve straight away. If it does not, the fix missed the cause.
- Wait for the field data. Real-visitor data covers the previous 28 days, so a fix can take up to four weeks to show fully. In Search Console, open the Core Web Vitals report, which groups similar URLs together, and start validation on the issue you fixed. Google tracks it over a 28-day window before marking it fixed.
If you would rather have someone do this for you, our experienced team handles exactly this through our website speed optimization service, which starts from $497. We find the real bottleneck on your key templates and fix it without changing your design.
These are general timelines and prices for typical projects, not a quote. For an exact timeline, quote and pricing, book a free consultation.
Speed fixes we would skip (or do last)
Some popular advice wastes time, and some makes things worse.
Chasing a 100 score. Google ranks with field Core Web Vitals, not the lab score. On mobile, a score in the 70s or 80s with passing vitals is a respectable result. The last ten points often cost more than all the rest.
Lazy-loading everything. Lazy-loading is right for images below the fold and wrong for anything at the top of the page. Applied site-wide, it is one of the most common causes of a slow LCP we see.
Rebuilding the site first. A rebuild is sometimes right, usually when a heavy page builder or theme is the root cause on every template. But it should come out of the diagnosis, not replace it, and it carries its own risks, covered in website redesign mistakes that kill conversions.
Buying the cheapest hosting to save money. If Time to First Byte is your problem, no front-end work will fix it. Budget hosting is one of the costs that shows up later, a pattern we describe in the hidden costs of cheap ecommerce websites.
Frequently Asked Questions
Why is my website slow on mobile but fine on desktop? Phones have slower processors and weaker connections, so heavy images and JavaScript hurt them far more, and Google assesses each device type separately.
What are good Core Web Vitals scores? LCP of 2.5 seconds or less, INP of 200 milliseconds or less and CLS of 0.1 or less, measured at the 75th percentile of real page views. A page passes only when all three are good.
Are Core Web Vitals a ranking factor? Google says its ranking systems use them, but good scores do not guarantee top rankings and relevance comes first. They can hold a good page back; they will not lift a weak one.
Why is my PageSpeed score good but Core Web Vitals failing? The score comes from one simulated lab load, while Core Web Vitals come from real visitors over 28 days, on slower phones and networks, actually tapping things. Trust the field data.
How long does it take to fix Core Web Vitals? The fix itself can be quick, but the field data covers the previous 28 days, so allow up to four weeks for improvements to show fully in PageSpeed Insights and Search Console.
Will a caching plugin fix a slow WordPress site? It helps when server response is the problem, because it serves ready-made pages instead of building each one. It does little for a lazy-loaded hero image or too many scripts, and two caching layers at once often conflict, so use one.
Want someone to find what is slowing your site down?
Our website speed optimization work starts with the diagnosis above on your key templates, then fixes the causes in order of impact. If speed is only part of a ranking problem, our SEO and technical optimization service covers the rest.
We do not chase scores. We aim for passing field data on the pages that bring you leads and sales.
