I’ve been trying to improve the load times of my website, and I’m using WebPageTest as a resource. Something that I’m finding weird is the grading system for TTFB.
My website has a <1s TTFB. Ideally I’d like it to be lower, but I’m having trouble bringing it down. My TTFB gets an F on WebPageTest.
I was wondering if my CMS is what was causing the long TTFB, so I ran a test on another website that I know uses the same CMS. It has a >1.6s TTFB, but for some reason it gets a B grade.
Can anyone explain this?
I am sure Patrick will confirm this, but I think the grade on the TTFB is graded based on the percentage of time it takes to load the entire site. For instance, if your site takes 1.2 seconds to load but your TTFB is the first second, your probably going to get a bad TTFB grade. If your site takes 15 seconds to load, then that same 1 second TTFB doesn’t look as bad and you’ll get a better grade.
I also think that it is weighted on page size as well - if your returning a tiny page, then a higher TTFB is bad, but if your sending 500kb of data, then the TTFB may be graded accordingly.
But those are just my guesses really.
Did you check the grading info? You can click the grade image to view more details on that grade. For example, it might say something like:
First Byte Time (back-end processing): 47/100
630 ms First Byte Time
102 ms Target First Byte Time
You might also want to check related posts on this topic:
The target is set based on a multiple of the estimated round trip time to the server (uses the socket connect time). If one of the sites is using a CDN like cloudflare for the main page it will throw off the calculation which can cause the inconsistency you are seeing.