Time To First Byte (TTFB) is the time it takes for a web server to respond to a request. It’s a metric reported by several page speed testers. It is also often quoted as being a primary means for measuring how fast a site is. The idea being that the faster a web server responds, the quicker a site will load.
However, numerous groups have found that TTFB isn’t that important. When looked at in isolation, the figure provides an appealing way to grade your site or hosting provider. However, when looked at in conjunction with other metrics, there seems to be a disconnect. This is especially true with regards to SEO rankings and improved user experience.
Here, we’re going to look at why TTFB can be easily manipulated, what metrics actually matter, and how knowing these things can help you to improve your site’s SEO, user experience, and more.
Time To First Byte Test
In some cases of testing site speed, you’ll find TTFB test durations far longer than what you would expect. This is despite actual page load times seeming much faster. This is the first indication that something is wrong with how TTFB measures speed.
A deeper look shows TTFB actually measures the time it takes for the first HTTP response to be received. It does not measure the time it takes for the whole page to be sent.
In the Time To First Byte test above, TTFB is measured at 0.417 seconds, which seems very quick. However, looking at the waterfall, we can see that this figure only correlates with the HTML loading time. Afterward, page load speed takes much longer for other assets on the page and we’re seeing DOM content loaded at around 1.6 seconds.
This is because the TTFB value is incredibly easy to manipulate. HTML HTTP response headers can be generated and sent incredibly quickly but they have absolutely no bearing on how fast a user will be able to see or interact with a page. For all practical purposes, they are invisible.
By loading HTTP response headers to speed up TTFB, it’s easy to create a ‘false’ view of a site’s speed. It also doesn’t necessarily mean that the rest of the waterfall will load quickly as well.
Speed Tests With NGINX
A good example of how Time To First Byte testing can be manipulated with HTTP headers is when looking at the page load times of NGINX in conjunction with compression.
Compressed pages are smaller and so they download from a server faster when compared with uncompressed pages. This ultimately means that page loads times to interactivity are much faster. However, from the perspective of TTFB, this is not true.
This is because HTTP headers can be generated and sent relatively quickly before the main page content.
This is an especially significant figure for those that make use of the Nexcess Cloud Accelerator, as this makes use of NGINX in order to speed up caching speeds on optimized Nexcess platforms.
Continue reading to find out what metrics you should be using to check page load times.
What About Page Load Speed Does Matter?
In a 2013 study by Moz, it was found that Time To First Byte does have a significant correlation with SEO rankings. The faster TTFB was, the higher ranked pages would be.
This being said (and as Moz themselves make clear) correlation and causation are not the same thing. The actual methods Google (and other search engines) use to crawl web pages and build out SERPs are not known to the public.
It’s been deemed by many that page load times to interactivity are actually a lot more important. When looking at page speed tests, it’s important to look at all the figures available as a whole and not just TTFB.
So, with regards to TTFB tests, SEO, and user experience:
Slow and Steady Wins the Race
Ok, all this doesn’t mean that you should let your site crawl to a halt. This isn’t a childhood fable or a call to reduce quick internet. Fast internet is one of the wonders of the modern age and you still want your site to load as quickly as possible.
What we’re saying is that if you’re trying to find how to improve Time To First Byte, stop.
It’s far more important for you to start looking at page load time in their entirety and not just the time it takes for a server to respond. At Nexcess, we’re proud of how fast our data center serves content, and work our hardest to make sure that our servers are optimized for providing a great user experience and helping to boost your SEO as much as a hosting company can.
We highly recommend checking out the Nexcess Cloud and seeing how Nexcess can help.