Lord Kelvin, the famous British scientist and mathematician, once said: “You cannot improve what you cannot measure.” Just as web developers track the quality of their code using unit testing, regression suites, and bug trackers to improve their work, performance metrics are needed as well.
So we created the Zoompf Performance Rank. This rank easily quantifies the performance status of your website. You can compare Performance Rank across different sites, or track how the performance of the same site varies over time.
Zoompf’s Performance Rank for a website is calculated by scoring each response individually on a scale from 0 to 100. The score of all the responses are averaged together to yield the Zoompf Performance Rank . If a response has no performance problems it scores 100 points. For every performance problem a resource has points are deducted from that resource’s score. The number of points deducted is based on the severity of the performance issue. Critical defects, like not HTTP compression for a compressable response subtract 25 points.High severity issues like Unoptimized images are 15 points. The lowest score any resource can have is a 0. This only happens if the page has a large number of performance issues. No page will have a score less than zero. This means if your website only consists of 2 pages, one that is completely awful and the other is perfect, the lowest Performance Rank you can have is 50, regardless of how awful that awful page is.
Zoompf Performance Rank is a very flexible scoring system that more accurately reflects the performance of a site than other scoring systems. Let’s say you run a photography site that consists of 1 HTML page and 19 images. Let’s say you don’t have HTTP compression turned on so your HTML page receives a score of 75 while the images have no problems and each receive a score of 100. The Zoompf Performance Rank would be 98.75. Other tools might take off a large number of points because such a critical issue as HTTP compression is not turned on. However for this photography site, HTTP compression is not a big concern. The vast majority of the site, both in bandwidth and page count, would not benefit from HTTP compression. Now consider the same photography site. HTTP compression is now turned on for the HTML page (100 points), but all of the images are unoptimized (85 points each). The Zoompf Performance Rank is 85.75. Even though the performance problem is less severe, it affects the vast majority of the photography website and so has a lower Zoompf Performance Rank.
You can find our your Zoompf Performance Rank and test your website for over 300 performance issues right now using our free web performance scanning service. We also setup a Twitter account for the performance scanner (@zoompfauto) which tweets out the Performance Rank and a link to the full Zoompf performance report for every werbsite it scans. This is a great way to see how your site compares with others. As always feedback and questions are always welcome. Enjoy!
Apparently, my Zoompf rank is 87.50. I can live with that!
Will there be a hall of fame / high-score list kind of thing?
Yep. Still working on that. You can game the system by getting a perfect score by simply having a single page with no issues. Thinking about ordering the hall of fame by score, and then by number of pages scanned. Any ideas?
Blacklisting cheaters so they won’t appear in the rankings anymore after fraud has been discovered is probably a good idea too.
[...] This post was mentioned on Twitter by Bruno Sabot, Zoompf Inc.. Zoompf Inc. said: What's your Web Performance Rank? Find out now http://bit.ly/ci9wu2 Follow @zoompfauto and see how fast other websites are. [...]
Got me a 90.11, yeah!
http://zoompf.com/scans/472cd5641138ad8a03ac95d15f4c5fa7/report.html
Here’s my feedback Billy:
> Duplicate Resource (Image)
It flags on http://www.aaronpeters.nl/images/logo_AaronPeters.png.
But there is no dupe.
> Invalid HTML Tag in
On this URL: http://www.aaronpeters.nl/en/
Which tag is that?
I do know about the http://validator.w3.org/check?verbose=1&uri=http%3A%2F%2Fwww.aaronpeters.nl%2Fen%2F
but that is in the body.
Ah, wait, I removed the tag.
That explains it.
> JavaScript Not Minified ( Block)
Huh?
> Potentially Cacheable Text Response (iPhone)
The index.html is 15418 bytes and caching is set at +10 minutes, so this is a valid flag.
Actually makes me wanna trim that 0.1 KB off!
> Robots.txt Without Crawl-Delay
With a site like mine (only 2 pages, very little objects) this flag makes no sense, agree?
Keep up the good work Billy!
My Zoompf rank in original (bulgarian) version of site is 96.43. But I think, that some Zoompf issues are not exact:
1. Unoptimized Image (GIF) -> if site is compatible with IE6, transparent PNG is not directly supported (only with additional filters).
2. Without Defer -> if I use Defer with jQuery in IE, browser freezes.
3. Unknown HTTP Header in -> Zoompf doesn’t supports Content-Style-Type and Content-Script-Type (they are exact – see http://www.w3.org/TR/REC-html40/present/styles.html)
4. Unoptimized Resource Location (Script) -> if I place jQuery in the end of page, accordion menu does not work.
5. iPhone cache responses issues -> Sprites are the best way to speed-up transfer. But sometimes is impossible to make sprites less then 15K and 20K to work great with iPhone. I think, that is better to make bigger then 15K and 20K sprites, then many small images for iPhone.
Excuse me for my bad English, but I am Bulgarian
Aaron,
I’m glad you are so happy with our free service. Here’s some additional info to help with the results:
Duplicate Resource (Image)
–
There is a dupe but it is not obvious. The file “logo_AaronPeters.png” embedded in your STYLE tag using a data URI and it is also referenced as a hyperlink for the background image. It looks like you are trying to use a browser hack so browsers that don’t understand data URIs still download the image.
JavaScript Not Minified (Block)
–
The two SCRIPT tags for you Google Analytics code have a number of spaces that can be removed.
Robots.txt
–
Totally agree.
I think some of the confusion comes from limits of our free service. In Zoompf WPO, our full performance scanner offering, we show you the actualy server response and then highlight the text that is causing the specific problem. This is how I determined where your unminified script blocks were. Zoompf WPO also has a “Duplicate Items” report which shows which URLs/embedded resources have the same content. You can learn more at http://zoompf.com/wpo.
Thanks again for the valuable feedback Aaron! I’m glad you like Zoompf so much.
Billy