Why is your site slow: common causes and how to find them with free tools
Practical guide for non-sysadmins: from TTFB to plugins, how to read PageSpeed Insights, GTmetrix and the Network tab without jargon.
UptimeMag editorial team · 20 November 2026 · 5 min read

If you manage a website without being a sysadmin, you usually only notice it's "slow" when someone complains or traffic drops. The problem is understanding where to look, because "slow" can mean five different things: the server responding late, too many plugins, heavy images, no cache, or a struggling database. This guide walks through the checks in order, using just two free tools (PageSpeed Insights and GTmetrix) and the browser's Network tab—enough to isolate the culprit without opening a support ticket.
The first number to check: TTFB
TTFB (Time to First Byte) is the number of milliseconds it takes a browser to receive the first byte of the response from your website's server. If this number is high, there's a problem before anything else matters: there's no point optimising images if the server takes a full second just to start responding.
The correct threshold, confirmed on Google's official TTFB page (web.dev), is: good up to 0.8 seconds, poor beyond 1.8 seconds. Several Italian sector websites confirm the same: Google recommends keeping it below 0.8 seconds, and beyond that threshold Google suggests staying under 0.8 seconds of TTFB... if we exceed 1800 milliseconds we're in negative territory.
The 200-millisecond myth
Many hosting blogs repeat a different figure: 200 milliseconds as a "Google-recommended" threshold. That's not what the official source says. Google's web.dev page, updated 18 November 2025, explicitly refers to 0.8 seconds as the general benchmark and 1.8 seconds as the negative threshold; the 200 ms figure is a stricter goal that some providers set for backend processing time alone, excluding network latency—not an official Google limit. If a provider tells you that you're "out of bounds" at 250 ms, they're not quoting Google: they're applying their own internal standard, which is legitimate but different.
Plugins and extensions: every added piece is one more request
Every plugin on WordPress (or extension on other CMSs) adds code the server must execute and often JavaScript and CSS files the browser must download. Performance metrics are usually affected by oversized images, complex and often unnecessary JavaScript, and poorly optimised CSS. In practice: disable one plugin at a time and re-run the test. If TTFB or load time drops noticeably, you've found the culprit.
Unoptimised images: the most common cause of LCP problems
The content users see first (LCP, Largest Contentful Paint) is almost always an image. The "good" threshold is 2.5 seconds: a website is slow when the main content takes more than 2.5 seconds to appear, or when the page responds sluggishly to clicks and touches, beyond 200 milliseconds. These are the thresholds Google uses in Core Web Vitals. The second figure mentioned here refers to INP (responsiveness to clicks), not TTFB: they're two different metrics, don't confuse them.
The practical consequences of not taking action are measurable: a Google study found that 53% of mobile visits are abandoned if the page takes more than 3 seconds to load.
No cache: every visit redoes all the work from scratch
Without a cache system (cache plugin, server-side cache, OPcache for PHP), every request regenerates the page from zero: database queries, template processing, rendering. With cache enabled, the server serves a ready-made copy. It's often the single change that lowers TTFB the most, because it eliminates repeated work.
Overloaded shared hosting
On shared hosting, resources (CPU, disk I/O, database connections) are split between dozens or hundreds of sites on the same server. If a neighbouring site consumes resources, your site's TTFB rises even if you've changed nothing. A typical symptom: the site is slow at different times than usual, with no obvious cause in your own code.
Slow database queries
If TTFB remains high even after disabling plugins and enabling cache, the problem may lie in queries: poorly indexed searches, oversized tables, plugins that query the database on every load instead of using their own cache. This is the most technical check on the list: if you've reached here without results, it's the point where involving whoever manages your server makes sense.
How to read PageSpeed Insights without getting lost
PageSpeed Insights shows two types of data, and confusing them leads to wrong conclusions. On one hand, field data: this shows your pages' real-world performance based on actual usage data, collected from real site visitors over the past 4 weeks. On the other, lab data, generated at test time: it's measured in a controlled environment (not by real users). If your site is new or gets little traffic, you'll only see lab data: that's normal, it's not a tool error.
The tool also displays First Contentful Paint (FCP) and Time to First Byte (TTFB): they track how long the first content takes to appear and how long the server takes to respond. These are the first two rows to check: if TTFB is high, start there, not with images.
Thresholds to remember
| Metric | What it measures | "Good" threshold |
|---|---|---|
| TTFB | server response time to first byte | ≤ 0.8 seconds |
| LCP | time for main content to appear | 2.5 seconds or less |
| CLS | visual stability of page during load | 0.1 or less |
| INP | response time to clicks and taps | 200 milliseconds or less |
GTmetrix and the Network tab: manual checking
GTmetrix is PageSpeed Insights' free complement: with a free account you can choose the test server location and browser, and it shows a waterfall chart of each individual page request with timings. It's the right tool when PageSpeed Insights says "slow server" but you need to understand why.
For a quick check, without installing anything, just use your browser's Network tab (F12, then Network, then reload the page). The first row in the list is the main HTML document: the time marked as "Waiting" or "TTFB" on that row is the same number PageSpeed Insights reports as TTFB. If it's above 0.8 seconds even on a fast connection and without other sites open, the problem is the server, not the user's browser.
No Prime Software product is mentioned in this guide: the tools described (PageSpeed Insights, GTmetrix, browser Network tab) are all free and third-party.
Written with the help of artificial intelligence and checked by the editors (EU AI Act, art. 50).