What TTFB measures
TTFB measures how long the visitor waits before the page can even start to load. Nothing can render until the first bytes of HTML arrive, so every millisecond of TTFB is added to FCP and LCP.
It is the sum of several phases of the request:
- Redirects, for example from http to https or from a tracking link to the page.
- Service worker startup, if the site has one.
- DNS lookup.
- Connection and TLS negotiation.
- The request itself, until the first byte of the response arrives: mostly your server's time to build the page.
TTFB is not a Core Web Vital. Google treats it as a diagnostic: a site doesn't have to meet the TTFB threshold as long as it still gets good FCP and LCP.
Good, needs improvement and poor
| Rating | TTFB |
|---|---|
| Good | 0.8 s or less |
| Needs improvement | Over 0.8 s, up to 1.8 s |
| Poor | Over 1.8 s |
As with the other metrics, read it at the 75th percentile of real page loads, split by mobile and desktop.
Example
An ad campaign links to a page through two redirects: a link shortener and an http-to-https hop. Each costs a round trip, and visitors on mobile networks in another region see a p75 TTFB of 1.6 s, against 0.5 s for direct visits. Linking straight to the final https URL and caching the page at a CDN brings the campaign's TTFB to 0.6 s, and its LCP improves by about a second.
How to measure and improve TTFB
Lab tools show TTFB for one request from one place. Real user monitoring shows it for every visitor, which matters because TTFB depends on distance and network. The usual fixes are a CDN or edge cache for HTML, fewer redirects, faster database queries and server rendering, and HTTP/2 or HTTP/3 to speed up connection setup.
Common TTFB mistakes
- Testing only from your own city. Visitors far from your server see much higher TTFB.
- Forgetting redirects. Tracking links, shorteners and trailing-slash redirects all count.
- Optimizing TTFB for its own sake. A fast server doesn't help if render-blocking scripts delay FCP; check what TTFB is holding back.
How MetricRaven measures TTFB
With Page speed turned on in Settings (it is off by default, on every plan), MetricRaven's add-on sends one snapshot per real page load with TTFB alongside LCP, INP, CLS and FCP, using Google's web-vitals library. It rates TTFB with Google's thresholds: good at 0.8 s or less, poor over 1.8 s.
TTFB isn't a column. In Pages → view all, hover a page's LCP or INP value to see its TTFB at the 75th percentile with the number of loads behind it, and the API returns p50, p75 and p90 per page with include=speed. The comparison of slow and fast visitors' purchase rates uses LCP and INP only. Filter the dashboard by country or campaign to see whether far-away visitors or one ad's landing page get a slow first byte and a lower conversion rate. See Page speed vs conversion.
Questions
- What is a good TTFB?
- Google suggests 0.8 seconds or less at the 75th percentile as good. Between 0.8 and 1.8 seconds needs improvement, and over 1.8 seconds is poor.
- Is TTFB a Core Web Vital?
- No. The Core Web Vitals are LCP, INP and CLS. TTFB is a diagnostic metric: a slow first byte delays FCP and LCP, which are what Google rates.
- What does TTFB include?
- Redirect time, service worker startup if there is one, DNS lookup, connection and TLS setup, and the request until the first byte of the response arrives.
- How do I reduce TTFB?
- Cache HTML at a CDN close to visitors, remove redirects, speed up server rendering and database queries, and use HTTP/2 or HTTP/3. Measure real visitors, because TTFB varies by location and network.