Question: Landing Original Loading Speed Vs. CDN (10)
Hey amazing people,
My apologies if this was answered somewhere else.
A quick one, please.
If I have for example a LP that have high loading speed like this one:
But because I use CDN (Cloudflare), if I tested the speed again after 1 sec, it shows me a lot better speed, here you go:
This LP runs in one country only, therefore I think the original speed does not matter as long as the CDN will serve it from the same server location. Is my thought correct?
I would appreciate some thoughts, please
cheers,
Are you asking why running a speed test twice gets different results? It could be because of browser caching since it uses real browsers to test.
For native traffic its on the slow side, but its not awful since users are generally on fast connections anyway.
Why so many requests though? Tons of images?
I mean:
When I run the test twice, the second time gives better speed because the testing server was served by CDN not my server.
So in this case I should not worry about the actual loading speed of my page as long as CDN loading speed is fast, is my understanding correct?
So many images cuz of review profile images.
It's the same page? You're not testing one without a CDN and one with?
Then in that case the 2nd result loading faster is almost certainly due to the assets being loaded from the browser cache.
So ignore the 2nd result. The vast majority of your traffic is going be 1st time visitors
Try using kraken.io to reduce the file size of all your images.
I am gonna have to take this further
Let’s say my LP and the whole campaign is targeting London, UK only.
I use CDN for the LP, and such CDN has got only 1 Server in London.
Here is my understanding: the first user ever user located in London will see the first result I have showed above, however everyone else in London will be served faster ( the second result) because it is already cached in the CDN server.
I might be wrong but please let me know if I have explained myself clearer.
Cheers
A server means your content/assets are located in one location.
A CDN means your assets are distributed across many locations.
If a visitor has not received an asset before, they still have to download it, regardless of whether its on a CDN or not. Then if you have settings enabled to tell user to save that asset, they will save it which will make for faster load times on repeat visits, but it only saves it for them on their device.
Agree with thedudabides regarding image compression. They definitely tend to add weight on page loading speed.
Besides images, pay some attention to the jQuery files attached to your lander. If you’re using standard libraries like those hosted on Google. Try linking the Google CDN hosted versions instead of serving them from your server. Chances are the visitor already has a copy of that library in his browser cache.
What about waterfall results of your page loading analysis? There you should be able to see which component adds more weight to it.
To clarify here, the Pingdom test should not be serving from a browser cache -- this would be bad for any testing system to do as they would make all the results become pointless.
If you test once, then again, and the second result is faster + you have your assets on a CDN, the second result is undoubtedly faster because the CDN edge servers have now cached your content on their servers, i.e. doing what a CDN should do.
If you test from other regions you should see the same effect where the first test is slow and primes the CDN, the second test is faster. If you look at the waterfall you should see the connectivity timings on your assets get much smaller, particularly the connect times.
So in a live campaign in a single country, yes the speed of loading will move toward the second result.
To test if it's browser caching, just open Pingdom in incognito like an hour later and run the same test with the same location. If the results are still fast, its becuase the CDN is working properly. If the result is the first test is slow and the second, third, etc. are much faster, something weird is going on as a CDN should be caching things for a long time, not less than an hour.
A little off the topic:
CloudFlare does NOT cache HTML pages. Although you can define a page rule instructing CloudFlare to cache html, it is still useless if your tracker use query string to pass tracking information between LP & tracker (like
If the size of your html page is big (for ex. you use minifier to inline css/js into html), the page download speed will not be good. I personally have not found a perfect solution for this problem, just cascade 2 CDNs as a workaround (S3 --> CloudFront --> CloudFlare).
Originally Posted by khzidan
Let’s say my LP and the whole campaign is targeting London, UK only.
I use CDN for the LP, and such CDN has got only 1 Server in London.
Here is my understanding: the first user ever user located in London will see the first result I have showed above, however everyone else in London will be served faster ( the second result) because it is already cached in the CDN server.
I might be wrong but please let me know if I have explained myself clearer.
Cheers
If you buy UK traffic and you don't use a CDN and your server is in the US , files will be served from the US. If you buy UK traffic and don't use a CDN and your server is in London, files will be served from London.
The latency is the proportional to the distance of the location of the user to the location of the server serving content. The bigger the distance , the bigger the latency.
So in your example, if you buy UK traffic and your server is in the UK you don't really need a CDN.
However if you buy UK traffic and your server is in the US, having a CDN will improve things.
Here's a case study I ran a few years ago and has some interesting results inside :
https://stmforum.com/forum/showthrea...results-inside
Home > Design - Imagery, Banners & Landers > Landing Pages