How did Awwwards winning websites deal with performance optimization in February 2016?
We have tested 29 winners of the Awwwards Site of the Day. (Check the end of the article for How we test.)
Top 10
1. Exoskills – 83/100
- PageSpeed Insights Mobile: 78
- PageSpeed Insights Desktop: 89
- WebPagetest: 83
2. Codeology – 79/100
- PageSpeed Insights Mobile: 89
- PageSpeed Insights Desktop: 95
- WebPagetest: 67
3. 2016 Make Me Pulse – 75/100
- PageSpeed Insights Mobile: 83
- PageSpeed Insights Desktop: 91
- WebPagetest: 63
4th – 10th place
- Quiver – 73/100
- Red Collar – 73/100
- CLEVER°FRANKE – 72/100
- Publicis90 – 69/100
- Petra in StreetView – 68/100
- Concrete LCDA – 65/100
- Centurion-Magazine – 65/100
Last 3
Without calling the names, the lowest 3 results in our test were:
- 37/100
- 23/100
- 19/100
Google PageSpeed Insights
The following table shows how many sites passed Google PageSpeed Insights rules:
| Mobile | Desktop | |||||
|---|---|---|---|---|---|---|
| Passed | Consider Fixing | Should Fix | Passed | Consider Fixing | Should Fix | |
| Avoid landing page redirects | 90% | 10% | 0% | 97% | 3% | 0% |
| Reduce server response time | 76% | 24% | 0% | 83% | 17% | 0% |
| Leverage browser caching | 0% | 45% | 55% | 0% | 62% | 38% |
| Enable compression | 52% | 14% | 34% | 55% | 10% | 34% |
| Minify HTML | 48% | 52% | 0% | 45% | 55% | 0% |
| Minify CSS | 79% | 21% | 0% | 79% | 21% | 0% |
| Minify JavaScript | 38% | 48% | 14% | 38% | 52% | 10% |
| Optimize images | 17% | 59% | 24% | 17% | 62% | 21% |
| Prioritize visible content | 38% | 59% | 3% | 48% | 52% | 0% |
| Eliminate render-blocking JS and CSS in above-the-fold content | 3% | 14% | 83% | 3% | 38% | 59% |
Legend:
- Passed – No significant issues found. Good job!
- Consider Fixing – Consider fixing this if it is not a lot of work
- Should Fx – Fixing this would have a measurable impact on page performance
WebPagetest
What optimization grades did websites get from WebPagetest?
| A-B | C-D | F (X) | |
|---|---|---|---|
| First Byte Time | 86% | 4% | 10% |
| Keep-alive Enabled | 100% | 0% | 0% |
| Compress Transfer | 58% | 21% | 21% |
| Compress Images | 38% | 24% | 38% |
| Cache static content | 7% | 10% | 83% |
| Effective use of CDN | 28% | – | 72% |
Commentary
- Surprisingly, a large number of websites don’t have gzip compression enabled. This is just a matter of simple server configuration and definitely shouldn’t be omitted.
- Many sites don’t set caching properly. Again, this is a configuration thing, and even though it’s a bit more complicated, it’s worth having a look at.
- Most of the sites could have also done a better job at image optimization. Often they could save hundreds of kB or sometimes even a few MB if they properly resize and optimize their images better.
- Usage of CDN is one of the most recommended performance optimization techniques, however only a small portion of websites use it effectively.
- None of the websites defer JavaScript execution or inline critical CSS. Presumably, the complexity of this technique scares off most websites creators.
Fail of the month
A photo saved as PNG with size ~ 4MB. If it was properly resized and compressed, it could save 3.5MB (94% reduction). There were a few similar sized images making the site home page over 16MB.
Next month
Some websites did a good job on performance optimization but none of them really shined – we are waiting for a score of over 90, and also for a little bit more color in the first 3 places 🙂
How we test
Google PageSpeed Insights
We test each website home page with Google PageSpeed Insights and count the average score from the Mobile and Desktop results.
WebPagetest
We test each website home page with WebPagetest and calculate the score from the result grades as follows:
- A – 100
- B – 80
- C – 60
- D – 40
- F – 0
Effective use of CDN criteria is either Yes (100) or X (No – 0).
Total Score
The total score is calculated as an average from the Google PageSpeed Insights score and the WebPagetest score. The maximum score any site can get is 100.
About the author
Related blog posts
Web Performance Awards – March 2016
How did Awwwards winning websites deal with performance optimization in March 2016?
Web Performance Awards – April 2016
How did Awwwards winning websites deal with performance optimization in April 2016?
Comments (2)
Anon
Quick note on this page's performance, I found that the Pingdom beacon took > 1.5s to load, meaning the whole page loaded in > 3s.
Apr 12, 2016
Lubos Kmetko
@anon that's possible but that image is loaded from Pingdom JS which is loaded asynchronously so it shouldn't affect perceived performance, should it? Or was it blocking loading and rendering the whole page? Thanks for the report.
Apr 12, 2016