- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
Should I put the same snippet javascript file on my own server?
Should I put the same snippet javascript file on my own server?
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report
For example, a snippet file is http://cdn.optimizely.com/js/xyz.js
Should I download the xyz.js file, and then put the file on my own web server for web page to load, instead of the remote file in cdn?
Re: Should I put the same snippet javascript file on my own server?
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report
You will find that you constantly have to keep updating the file on your CDN every time you make a change.
There is no real performance issue with loading the snippet from the Optimizely CDN.
Re: Should I put the same snippet javascript file on my own server?
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report
David has nailed it.
In general there is no real benefit to self-hosting our snippet and additionally you would need to download the snippet from Optimizely and update the self-hosted snippet file every time you change the settings for your project or make changes to an experiment.
If you've any concerns about snippet performance, feel free to let us know so we can put your mind at ease!
-Kathryn
Re: Should I put the same snippet javascript file on my own server?
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report
Well seems casper.com hosted the file themselves and it shaved off 1.7 seconds on page load time so seems there is value.
Re: Should I put the same snippet javascript file on my own server?
[ Edited ]- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report
The 1.7s mentioned in the article includes changes other than just "self hosting".
They also did not mention if they used preconnect and prefetch tags to help move these handshakes earlier.
Isolating the "self hosted" pieces from the data they provided, it looks like ~400ms were saved by self hosting.
The other improvements resulted from other changes.
If 0.4s on first-page-hit is important to you (e.g., we cache ours for 15 minutes, so the user only fetches it once every 15 minutes - typically their very first page hit of their session), then self-hosting may be worthwhile to pursue.
Analytics and Testing Guru



