Making NodeBB faster and getting higher Google PageSpeed score
-
You're testing against a page that has a 300kb image on it...
RE: PageSpeed, at some point we were in the 80-90s for both. Probably have to revisit that for our next version and see what we did wrong since.
-
@psychobunny Here are tests without an (big) image:
http://www.webpagetest.org/result/150810_WR_137N/
it is not better.
On the other hand, the software should resize the image or show smaller version.
-
My perceived loading speed for Nodebb is satisifying. In fact, you are the first person I heard complaining about speed. Use your same method and tell us what you find for Discourse
-
@Ziggy said:
My perceived loading speed for Nodebb is satisifying. In fact, you are the first person I heard complaining about speed. Use your same method and tell us what you find for Discourse
Discourse seems to have similar Google PageSpeed results and is slower on Pingdom Tools and WebPageTest.org
But should we compare to slower websites? Wordpress websites are much faster, Stackoverflow is also much faster.
Speed is an important ranking factor, content is important ranking factor.
The difference is, once you have a fast cms you don't need to bother. You can create your content.
Other way around. You can spend thousands of hours for your content and community, but the speed destroys your visibility on search engines.
-
Well, this is what I got when I tested this 552.2 KB page:
-
@Ziggy
that is fast. my results are somehow different:
If the speed can be simply improved, then all we need is a better google insight score, especially on mobile devices.
-
I haven't tested the mobile experience yet.
On desktop, my browsing experience is pleasant.
Let's see what others have to say.
-
Nick picking at its finest
Just for comparison heres a discourse page without an image on it....
http://www.webpagetest.org/result/150811_NW_N3Q/
First View
(5.770s)
Repeat View
(4.699s)
Compared to the nodebb one you listed....
First View
(4.117s)
Repeat View
(2.946s)Have a look at the output of PageSpeed and say what you think they should do for each issue?...bet you dont come up with much
this site isnt really doing that much wrong.
The only thing that was obvious is that all the images for the emoji plugin should be optimized. And minimize the css and html and save a whole 1.5k, whoope. -
Why do you compare it to discourse? Compare it to stackoverflow or wordpress. In the real SEO world we compete against the fastest, not the slowest.
said it before: "Discourse seems to have similar Google PageSpeed results and is slower on Pingdom Tools and WebPageTest.."
Quote: this site isnt really doing that much wrong.
I know and I really like nodebb, but why not doing everything right? It is not really much to change to make google happy.
Every single minute of your time you invest in your community and content is a lost time, if your site is not as fast and SEO friendly as your hardest competitors.
-
I come from a cryptocurrency background... and some of the recent posts around here smell of FUD. This is a tactic oft used by the competition of a product on their home turf to sow discord and distract from progress.
-
@JonnyBlack I think you should open a new topic and discuss it there. Off topic.
Here is a discussion of the possible improvements, to make nodebb even better. More SEO-Optimized (= Google Optimized) nodeBB would result in higher Google rankings of nodebb communities and this would automatically lead to higher installation rates. Win-Win for all.
-
@prog We'll always place an emphasis on getting NodeBB loading quickly, as user experience is one of the features we'd like to focus on.
That said, the Google pagespeed test is at best still machine driven and does not accurately represent real-world usage.
I don't mean to disregard the test results -- not at all! As @psychobunny mentioned, we used to score higher.
Re: serving smaller images - this is unfeasible in core as we merely store post content, and do not rehost images. If someone links to an imgur image, we wouldn't intercept the image and host multiple versions of it locally (one for each dpi/device), because that is squarely in plugin territory.
-
@prog I would also dispute blogs having superior content. While articles (especially long form ones) are great from an SEO perspective, you can often get much better quality content from a healthy discussion, this here being a perfect example.
