im running my landing pages from a server in singapore through cloudflare, i have noticed the following for south korea
1. the one with the cloudflare script loads faster on IE and safari
2. the one without the cloudflare script loads faster on chrome
and the following for brazil
1. the one with cloudflare script loads faster on IE
2. the one without cloudflare script loads faster on chrome
maybe chrome already has this function? while the other browsers dont have this "faster loading" function?
suppose here that maxcdn is better, are u going to use different cdn for different region? like maxcdn for say brazil and cloudflare for south korea? thats a tonne of work to remember where all your files are?
the challenge is that most ad networks do not allow browser targeting, so u need to take all the traffic, in this case which browsers should we optimize for?
Woah there, I think you have some crazy things going on in terms of data and thoughts.
Firstly, CloudFlare is not a script by any description.
CloudFlare is a service that works at the DNS level. You set your domain name (at your registrar) to use CloudFlare's nameservers. Now, when you try to visit your domain, CloudFlare basically tunnels traffic between your server and the user but also caches stuff like images, i.e. acts as a CDN.
When users request your pages, CloudFlare will deliver image content to them from their edge location (i.e. nearby) but will still get the page itself from your server (if .php, not sure if it caches .html files).
All else being equal, two things happen here:
1. Your initial time take to contact your server may increase or decrease by a little bit due to the different networking involved.
2. You should see reduced page load speed due to static content being cached and delivered by CloudFlare's CDN.
Both of these things have nothing to do with your browser, providing you don't have anything abnormal set up like no caching or customised DNS configs.
How exactly have you determined your conclusions that the load times are better/worse with and without CloudFlare? How did you test this?
Have you similarly looked at the actual landing page CTRs and EPCs for pages with and without CloudFlare on a real campaign?
Again, CloudFlare is not a script so there is no point pondering whether browser functions are having an effect here - the browsers have little to do with what is happening on the backend.
Anyway, most people will stick to one CDN. If page load speeds are too slow in a specific region then you could use a different CDN for that - e.g. one with a PoP in that region, but this would be an infrequent thing to do.
As for browser targeting, you should a) track browser-based stats with your tracking system, e.g.
If IE for example does terrible you can take that as a hint that your page design might not display well in IE (have you considered that any real-world browser-variation results may have nothing to do with CloudFlare?)
You could also just funnel all IE traffic to whatever page it seems to work best on. IE users are often a different demographic to the average user so this isn't a silly idea.