[Closed] superb gtmetrix vs average gtmetrix

Hello,

GTmetrix works based on the input it's given and if you give one site in one language and another site URL in another language then you'll likely see different results.

The first url you sent was this one: hidden link

which points to a specific Product which in EN is this one: hidden link

In order to compare the results, you should be using that link i just sent. i ran a test with that link and here are the results: hidden link

Very similar to the original language (remember that you will never have 2 exact results because the server resource usage will always be different and there are many factors to take into consideration).

----------------------

Interestingly the second URL is different. it's heavier because i seems it contains more images and the product page is heavier in general (bigger DOM, etc).

I wonder if you have a caching mechanism in place and which one because if i access the site normally, the HE version actually loads slower, but it's not that different from the EN version (which loads a second and a half earlier), but it's not the same that i see on GTMETRIX, where sometimes i get a 404 error, sometimes i get a 7s load (mediocre, as you mentioned), sometimes i get a B score and sometimes an A.

In Hebrew, i get more A and B than anything else but it seems it always loads the CACHED version so the loading time from the server is lower.

We do not have anything that handles cache in one language or the other, the cache plugins take care of that and it's only possible configurations that could alter this (We do not really touch any of that)

No, we're not. On the front-end they are nothing but a "new page" and the fact that with another product the stats are the same between languages tell me that maybe there's something different between product 1 (showing the same between languages) and product 2 ( showing differently between languages) and that might be what's creating this issue.

If there's a plugin adding something only to this product and not the other, then maybe that's the clue here, maybe that forces a cache flush or something creating this issue.

I'd try to test on a staging site, with all plugins disabled except WooCommerce and our WPML plugins. in this case, it's probable that the grade is almost to the top, but that's not what you'd be checking there but instead checking for DIFFERENCES between the languages. checking if in one language the test takes considerably longer to load than in the other, for example.

This should be first tested WITHOUT any type of cache, because again, you're not going to be checking for an A grade but for big differences between the languages.

You need to use a few products as samples (product 1 vs product 2, both in EN and in HE).

If you don't notice any change, take a few mins and enable some of the WC-related plugins (not all of them at once but in small batches). wait a few minutes to let the server reset the resource usage and then try again and so on.

At the same time, you can ask the hosting for the server logs of your installation so you can see if there are particular thing happening at the time of your testing when you notice the increase.

I would kindly ask you to read my last answer again. i wasn't testing anything on a staging site (i don't even know the URL of a testing site).

I also don't have access to neither of your sites and i'm just studying and testing on the front-end using the 2 products you used as references, which are both on the front-end.

At the same time, my research focused on TTFB, DOMContentLoaded and other JS events and caching on some of the pages.

--------------------

What i can tell you is this. when you are testing for performance, you need to find trends and what you found on the second product is currently not a trend. the reason why is because if you check the first product, you'll see it works normally in both languages (which means that the same cache plugins or systems work normally with WPML).

That's why i suggested you tested a few things because there is very likely something happening on that product that has a bad grade that doesn't exist on other products and it's there where the clue might be. so far, if a problem is not happening on all the products but just some (i tested more than just the 2 and i only found a problem in the one you shared) then the problem is most likely something else.

Regards.

The DOM contains everything, it's the entire site. what i meant there is that the second product is heavier as a website in relation with the first product.

There were more images loaded and they were sometimes cached, in both languages, sometimes they weren't.

Have you tried asking the litespeed cache authors about this? they are the ones that created the WPML integration and they are the ones that maintain it too so they can probably have better insights about how it works.

Then you can let us know here in case you need more information.