Lorixen.
EN

Slow Server Response on Cheap Hosting: Diagnosing Time to First Byte

Short answer: Time to first byte (TTFB) is how long the browser waits before the first piece of the page arrives. Measure it from several locations, and compare a cached page with an uncached one such as the basket or a search result. If cached pages are fast but uncached ones are slow, the server is struggling to build pages: check the PHP version, the number of PHP workers on your plan, database size and heavy plugins. If even cached pages are slow, the host itself or the distance to visitors is the issue, and a CDN with edge caching or a better plan will help. Gather numbers before you contact the host.

What TTFB includes

TTFB covers everything before the first byte of HTML reaches the browser: DNS lookup, connecting, the TLS handshake, any redirects, and the time the server spends building the page. MDN's glossary entry defines it, and web.dev's TTFB guide suggests aiming for roughly 0.8 seconds or less for most visitors.

A slow TTFB delays everything else. Images, fonts and scripts cannot start downloading until the HTML arrives, so a 2.5-second wait at the start makes even a lightweight page feel sluggish.

Step 1: Measure from several places

One measurement from your office proves little. Use:

  • Browser developer tools: open the Network tab, reload, click the main document and read "Waiting for server response".
  • PageSpeed Insights: the field data section shows real-visitor TTFB if your site has enough traffic.
  • A multi-location speed test: many free tools test from London, Frankfurt, New York and other cities at once.

Run each test three times and note the median. Record the time of day too; shared servers are often slower in the evening when neighbouring sites are busy.

Step 2: Compare cached and uncached pages

This is the most useful single test. Page caching stores a ready-made copy of a page so the server does not rebuild it for every visitor.

Page Usually cached? What a slow result means
Homepage, About, blog post Yes The server or network is slow even for simple work
Basket, checkout, account No Page building is slow
Search results, filtered shop pages Often no Database or plugin queries are heavy
Admin dashboard No Same as above, felt by your staff

To test a page uncached, add a random query string such as ?nocache=123 to the URL. Many caches skip such requests, though not all do. Logged-in sessions also usually bypass caching.

Fast cached, slow uncached: focus on PHP, the database and plugins. Slow everywhere: focus on the host, location or network layer.

Step 3: Check PHP version and workers

On shared or entry-level managed hosting, two settings matter most:

  • PHP version. Newer supported versions are noticeably faster than old ones and still receive security fixes. In your hosting panel, find the PHP selector and check which version the site uses. Test the site on a staging copy before upgrading.
  • PHP workers or processes. This is how many uncached requests the server can handle at the same moment. If your plan allows only a couple and three visitors check out at once, the third waits in a queue. Hosting dashboards sometimes show this limit, or "entry processes", with a graph of when it was reached.

Also check memory limits. A site hitting its memory ceiling slows down or fails, and the error log will usually say so.

Step 4: Look at the database and plugins

For WordPress and WooCommerce shops, typical slow spots:

  • Bloated options table full of leftover data from removed plugins, loaded on every request
  • Huge tables of logs, sessions, revisions or expired transients
  • Search and filter plugins running expensive queries on every product listing
  • Plugins calling external services while the page is built, such as stock or currency feeds

A query monitoring plugin, used on a staging copy, shows which plugins and queries take the longest. Remove what you do not need before buying more server power.

Step 5: Use CDN edge caching

A content delivery network keeps copies of your files on servers close to visitors. Standard CDN setups cache images, CSS and scripts, but not the HTML page, so TTFB barely changes. Edge HTML caching, offered by several CDNs and some hosts, caches the page itself at the edge, which can cut TTFB sharply for visitors far from your server.

Set it up carefully: basket, checkout and account pages must never be served from the cache, and the cache must be cleared when you publish changes.

Step 6: Bring evidence to the host

Hosting support takes a ticket seriously when it includes facts. Send:

  1. URLs tested, cached and uncached, with TTFB results from at least two locations
  2. Times and dates of the slow measurements
  3. Your PHP version and any resource-limit warnings from the dashboard
  4. Confirmation that you tested with plugins disabled on staging, if you did
  5. A direct question: "Is the server overloaded, and are we hitting our PHP worker or CPU limit?"

If the reply is that you need a higher plan, compare that cost with a host that suits the site better. A small business site with a shop often outgrows the cheapest shared tier.

When to move host

Consider moving when cached pages are slow from nearby locations, resource limits are hit daily after you have removed waste, the PHP version cannot be updated, or support cannot explain the slowness. Test a new host with a copy of the site before switching DNS.

FAQ

What is a good TTFB?

Google's guidance suggests about 0.8 seconds or less for most visitors. Lower is better, especially for shops.

Will a caching plugin fix slow TTFB?

It helps a lot for cached pages. It does not help basket, checkout or logged-in pages, which rely on server speed.

Does a CDN reduce TTFB?

Only if it caches HTML at the edge. Caching images and scripts alone mainly improves other parts of loading.

Is cheap hosting always slow?

Not always, but limits on workers, CPU and memory are tight. Busy or dynamic sites feel those limits first.

For a clear answer on whether hosting is to blame, send us your URL and we will measure it properly.

Sources

WhatsAppBook a call