How to tell if shared hosting is REALLY hurting page speed???

.
I have read in many places that upgrading a site’s webhosting plan will improve it’s page speed.

But in reality, will it really matter? My tiny experiment below says “no”.

Please share your knowledge and experience on this topic…

I have not seen any solid studies or examples of pagespeed differences that result directly from boosting webhosting plan levels.

  • If you know of one, please share a link.
  • If you know of a normal website on a Premium hosting plan, please share the site’s name and host.
  • I would kinda like to be proved wrong about this :blush:

(links to the tests these screenshots came from are at the bottom of this post)

Here is a partial screenshot of my site’s test results using Mobile settings.
[ Dulles, VA - Moto G4 - Chrome - Emulated Motorola G (gen 4) - 3GFast – Mobile ]
This site uses the lowest level shared hosting plan that GreenGeeks.com offers (the Lite plan)

And here is a screen shot of the test results of a GreenGeeks.com page using the same Mobile settings.
This site is on a DEDICATED SERVER
I thought it would be super fast compared to my cheap basic shared plan. But take a look…

The “light” areas in each bar of the waterfall are just as long from this Dedicated Server as they are from my basic shared hosting. It just doesn’t make any sense to me. :huh:

GreenGeeks webhosting service is very well rated for speed and uptime. Here are their published differences between the bottom (Lite) and top (Premium) shared hosting plans:
Disk I/O : Lite = 10 MB/sec, Premium = 20 MB/sec
Inodes : Lite = 150000, Premium = 500000
CPU Cores : Lite = 2, Premium = 4
I am assuming that their Dedicated Server’s stats are same or better than their Premium shared hosting

Am I looking at the wrong things?

Are the “light” areas in the waterfall not affected by hosting levels or servers speeds?

Would upgrading my site’s hosting plan really make a big difference to my page speed? HOW?

Below are the webpagetest.org test results I used:

My site with Mobile settings:
https://www.webpagetest.org/result/210305_Ai77_e8834aec1c9c58bdf1d226db9881279b/

GreenGeeks interior page with Mobile settings:
https://www.webpagetest.org/result/210305_AiSG_480b02a1cf641ef41046adec659f8f53/

My site with Chrome Desktop settings:
https://www.webpagetest.org/result/210305_AiCS_83f6f0a77c43d3d1e2a048c07243f19d/

GreenGeeks interior page with Chrome Desktop settings:
https://www.webpagetest.org/result/210305_AiB8_9382276799d43d9f17d6c59628ccb130/

Thank you for your help :cool:

Greg

1 Like

Not quite sure what are you about. Using Google’s metrics for the Core Web Vitals, I get these reports:

  1. WPT https://www.webpagetest.org/result/210305_AiCE_a966d921421b20a611ce07a9525cce04/

  2. Lighthouse https://www.webpagetest.org/lighthouse.php?test=210305_AiCE_a966d921421b20a611ce07a9525cce04&run=1

And do not see any server issues or so ever.

Hope above help.

Wish you good luck Greg.

1 Like

Hi Greg,

I agree with you. From my experience, if you have a site where almost all the traffic is anonymous (not logged-in), there is no performance benefit in more than a shared host.

Maybe there comes a point with web traffic that you need a more powerful server but if you have a static site that rarely changes, you can just use a CDN (I believe that even on a free plan) and cache everything. Normally your HTML is not cached on a CDN, just the web assets like images, JS, CSS, but you can choose to cache everything - so there’s no hit on your server.

It’s when user log into your site that you need a VPS or more. My experience is with a shared host with about 30K pageviews per month. Maybe it’s a different story if you have a million visitors but a CDN will definitely help in scaling.

btw Well done Greg on creating a performant site and for example thinking about how you embed Youtube videos without impacting performance.

David

1 Like

Pretty much echo what everyone else has said. The best you can expect for the “light” part of the bars is the same as the socket connect time to that server (orange bar) which is the native round trip time on the connection. If you’re anywhere near that they you aren’t constrained by the server resources and upgrading won’t make it faster.

Using a CDN is the only way to improve it if you want to make the round-trips shorter, particularly for the static content as that’s their whole purpose for existing.

Usually when you see a hosting problem, the base HTML has a REALLY long TTFB (light part) - like in the seconds.

1 Like