Slow Website? What to Investigate Before Changing Your Hosting
Hosting affects one part of how fast a page loads: how quickly the server starts sending it, measured as Time to First Byte (TTFB). Everything after that – images, scripts, fonts, layout – happens in the visitor's browser and stays exactly the same on a new server. So measure before you move. If the server responds quickly and pages are still slow, new hosting will not help. If it responds slowly, find out why: a slow database query or a missing cache will be just as slow on more expensive hardware.
Step 1: Check what real visitors experience
"The site feels slow on the office computer" is not a diagnosis. Google's Core Web Vitals give three measurable targets. In the Search Console report, a page is "good" when Largest Contentful Paint (LCP, when the main content appears) is 2.5 seconds or less, Interaction to Next Paint (INP, how quickly the page reacts to taps and clicks) is 200 milliseconds or less, and Cumulative Layout Shift (CLS, how much the layout jumps around) is 0.1 or less.
Two free tools show these numbers:
- PageSpeed Insights shows field data from real Chrome users over the previous 28 days, reported at the 75th percentile so you see the more frustrating experiences, alongside a lab test. The performance score comes from the lab test in a simulated environment, and Google notes that good lab data does not necessarily mean real users have a good experience. Use field data to decide whether there is a problem, and lab data to find out why.
- The Core Web Vitals report in Search Console groups pages with a similar experience and gives each group a status, separately for mobile and desktop. A group takes the status of its weakest metric.
Smaller sites may not have enough visits for field data. PageSpeed Insights then falls back to data for the whole site or shows none, and Search Console may show "No data available". In that case, run lab tests several times on your most important pages – and open them on a real mid-range phone.
Step 2: Split the load time into parts
Google's guide to optimising LCP breaks the time until the main content appears into four parts and suggests roughly how big each should be:
| Part | What it is | Suggested share of LCP | Can hosting fix it? |
|---|---|---|---|
| Time to First Byte | From the start of the load until the first byte of HTML arrives | about 40% | Partly |
| Resource load delay | Time before the browser starts loading the main image or other LCP resource | under 10% | No – it depends on how the page is built |
| Resource load duration | Time to download that resource | about 40% | Mostly no – file size matters most; a CDN can help |
| Element render delay | Time between the resource arriving and it appearing on screen | under 10% | No – CSS and JavaScript |
The guide calls these percentages guidelines, not strict rules. What matters is the shape: if TTFB is a small part of a slow LCP, the bottleneck is in the page, not the server.
Step 3: If TTFB is high, find out what is inside it
TTFB is not just the server "thinking". It includes redirects, the DNS lookup, connection and TLS setup, and the wait until the server sends the first byte. web.dev suggests most sites aim for 0.8 seconds or less and treats more than 1.8 seconds as poor, but calls these thresholds a rough guide – TTFB is not a Core Web Vital itself. Check these causes, roughly in this order:
- Redirect chains. http → https → www → a language folder can add several round trips before the page even starts. Link to the final address directly.
- Distance. Google's LCP guide lists distant visitors among the causes of high TTFB. If your customers are in Bulgaria and the server is on another continent, location matters.
- No page caching. If every visit rebuilds the page from the database, even a fast server spends time on work that could be done once.
- Slow database queries. One badly indexed query on a large table can cost more time than any hosting upgrade can win back. On our product search engine gega.bg, with around 2.7 million products, search became faster after we replaced LIKE queries with a FULLTEXT index on InnoDB – a change in the code, not the server. We describe it in how a product search engine with millions of products works.
- Heavy plugins or extensions that run on every request, and scheduled tasks that run at busy times.
A simple test: compare TTFB for a static file, such as an image, with TTFB for a dynamic page. If the static file is fast and the page is slow, the time is spent in the application, and moving hosts will change little.
Step 4: Look at what the page makes the browser download
According to the HTTP Archive's Web Almanac 2025, the median home page in July 2025 weighed 2.86 MB on desktop and 2.56 MB on mobile. On mobile home pages, the median was 911 KB of images and 632 KB of JavaScript. None of that changes with a new host. The usual suspects:
- Images uploaded straight from a camera or phone, without resizing or compression.
- A main image that loads late. web.dev is blunt: never lazy-load your LCP image. Images inserted by JavaScript or referenced only in CSS are also discovered late, so keep the main image in the HTML.
- Render-blocking CSS and JavaScript. Large stylesheets and synchronous scripts in the page head delay the moment anything appears.
- Fonts – several families and weights, each a separate download.
- Third-party scripts – chat widgets, trackers, embedded videos and social feeds: code you don't control, running on your page.
Step 5: Check responsiveness, not just loading
A page can load quickly and still feel slow when buttons and menus react late. That is what INP measures. Google's INP guide explains that long tasks on the browser's main thread – script evaluation during loading is one source – delay the response to input, and that large pages with many elements take more work to update. The fix is less and better-organised JavaScript, not a bigger server: the delay after a tap usually happens on the visitor's device.
When hosting really is the problem
Changing hosting is the right decision when the evidence points to the server:
- TTFB is high even for static files and cached pages;
- response times get much worse at busy times, or the site slows down for reasons you cannot find in your own code;
- your hosting plan regularly hits its CPU, memory or process limits;
- the server is far from your visitors and a CDN is not an option;
- the host cannot provide supported versions of the software your site needs.
Even then, upgrading the plan with your current host is sometimes cheaper than moving, because a migration takes work and carries its own risks. And sometimes the honest answer is that the site needs development work, not new infrastructure.
Symptoms and likely causes
| Symptom | Likely cause | Where to look |
|---|---|---|
| Every page is slow, even simple ones | Server, distance or missing cache | TTFB in lab tests; resource usage in the hosting panel |
| Home page is fast, search or category pages are slow | Database queries | Slow query log, indexes |
| Text appears quickly, the main image late | Image size, lazy-loading, late discovery | The LCP element in a PageSpeed Insights lab test |
| The page loads, but taps react late | JavaScript, third-party scripts | INP field data; the browser's performance tools |
| Fine in the office, slow for customers | Mobile devices and networks | Mobile field data |
If you want a clear answer before paying for a migration, we can measure where the time goes on your key pages and tell you whether the fix is in the code, the content or the server. Speed is part of every website we build, and slow database work is the kind of problem we solve in web-based software. If your site runs on a CMS, WordPress vs custom development covers the trade-offs. Send us a few example addresses.