Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
Data Driven Performance Optimisation
Search
Stefan Baumgartner
June 10, 2015
Technology
560
2
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Data Driven Performance Optimisation
Which key metrics should we choose to get an idea on our site's performance?
Stefan Baumgartner
June 10, 2015
More Decks by Stefan Baumgartner
See All by Stefan Baumgartner
Idiomatic Rust
ddprrt
0
190
Automating UI development
ddprrt
0
200
The hero's journey in JavaScript frameworks
ddprrt
0
880
Get a grid for flex' sake
ddprrt
1
620
Technical Writing
ddprrt
0
620
Web performance optimisation for single page applications
ddprrt
1
810
The Super Mario Bros. Principle - Extended Edition
ddprrt
0
210
HTTP/2 is coming! Unbundle all the things?!?
ddprrt
4
670
The Super Mario Bros. Principle
ddprrt
2
650
Other Decks in Technology
See All in Technology
視野はAIに、視座は人に / ai-for-breadth-humans-for-perspective
nkzn
6
3.2k
bpf_arenaとstruct_opsを使うのが難しかった話
masu_mi
0
210
ゼロから作るWebアプリの新定番!?Amplify Gen 2 × AWS Blocksで実現する「開発速度とRDBの柔軟性のいいとこ取り」
amixedcolor
1
170
Building Audio Plugin Ecosystem for Android in 2026
atsushieno
2
160
Content Mapper APIの応用例~CSS Modules、言語ファイル~
uhyo
2
620
Kiro 入門 (1 つのエージェントを、どの画面からでも) | 目指せ Kiro マスター 01
inariku
PRO
5
1.1k
FinOpsを導入してマルチプロダクトのコスト最適化を実現する
kworkdev
PRO
0
140
続・ソフトウェアアーキテクトのための意思決定術: Decision Readiness, Revisited — When AI Changes the Economics of Software Architecture
snoozer05
PRO
9
2.7k
サボる子はいねぇがー!AgentCoreハーネスの脆弱性に学ぶ、エージェントの入力バリデーション
iwamot
PRO
5
300
バックオフィスに染み出すTypeScript ── clasp で安全に仕組みを作る
nealle
1
180
複雑な業務システムを高速でつくるため、チーム全員をつくれる人へ
ayakurabuchi
0
140
SLOを阻む不安
sraku2159
0
230
Featured
See All Featured
10 Git Anti Patterns You Should be Aware of
lemiorhan
PRO
659
62k
The Pragmatic Product Professional
lauravandoore
37
7.5k
How to Build an AI Search Optimization Roadmap - Criteria and Steps to Take #SEOIRL
aleyda
1
2.2k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
35
2.6k
The Power of CSS Pseudo Elements
geoffreycrofte
82
6.6k
How Software Deployment tools have changed in the past 20 years
geshan
3
35k
The AI Search Optimization Roadmap by Aleyda Solis
aleyda
2
6.4k
Chasing Engaging Ingredients in Design
codingconduct
0
390
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
33
5.6k
[Rails World 2026] Durable orchestration on Rails: from continuation to workflow
palkan
3
1.2k
What the history of the web can teach us about the future of AI
inesmontani
PRO
3
790
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
141
35k
Transcript
Data Driven Performance Optimisation Stefan Baumgartner | @ddprrt
ruxit.com
None
2011 2015 4 times a week in the fitness studio
0, nada, niente, null, void, undefined able to run 10 kilometers able to get to the supermarket across the corner 20% body fat this is getting too intimate (and embarrassing) 74kg
Faster than a speeding bullet?
Our websites are too heavy!!
Images 1.310 Other 4 Video 227 HTML 56 Fonts 97
Scripts 329 Stylesheets 63 http://httparchive.org/interesting.php Total: 2099 kB
Users expect fast sites. After 3 seconds of load time,
40% will abandon your site https://speakerdeck.com/lara/designing-for-performance
Etsy found out that: + 160kb on a page +
12% bounce rate https://speakerdeck.com/lara/designing-for-performance
Also: - 1 redirect on a page + 12% click-through
rate https://speakerdeck.com/lara/designing-for-performance
Performance matters!
But what should we be aiming for?
We need good, reliable and replicable key metrics
1 Pageweight
“Weight” is an easy one: The bigger your site is,
the more it has to transfer over the wire.
So the goal is: Reduce weight!
Images 1.310 Other 4 Video 227 HTML 56 Fonts 97
Scripts 329 Stylesheets 63
http://imageoptim.com
<img src=“path/to/image/800.jpg“ srcset=“path/to/image/200.jpg 200w, path/to/image/400.jpg 400w, path/to/image/600.jpg 600w, path/to/image/800.jpg 800w,
path/to/image/1000.jpg 1000w," sizes="(min-width: 992px) 60vw, 80vw" >
How to keep track?
None
None
https://github.com/macbre/phantomas
http://www.filamentgroup.com/lab/weight-wait.html
This menu at the top holds 94 product images The
requests alone stall the browser for roughly 14 sec
But it just has to be available when JavaScript is
on … so we load it asynchronously in our JavaScript at the bottom. This reduces time to glass to 1.3s
2 Page Speed Score
“Page Speed Score” tells you if you’ve followed a series
of best practices for every component of your website
None
https://developers.google.com/speed/pagespeed/insights
Eliminate render-blocking JS and CSS Optimize images Minify HTML, CSS,
JavaScript
Reduce server response time Leverage browser caching Avoid landing page
redirects Enable compression
None
None
None
How safe is Page Speed Score?
blog.catchpoint.com/2011/12/27/biggest_misconception_about_google_page_speed/
blog.ruxit.com/pagespeed-score-matters-less-think/
http://mobiforge.com/research-analysis/googles-mobile-friendly-test-can-spruce-goose-really-fly
None
None
None
What the …?
Solution: Page Speed Score is a good indicator on your
dev style … but we need more!
3 Speed Index
The Speed Index is the average time at which visible
parts of the page are displayed. It is expressed in milliseconds and dependent on size of the view port.
∫ 1 - visuallycomplete/100 0 end
What does this mean?
Visually complete (%) 0 25 50 75 100 Time in
Seconds 0s 1s 2s 3s 4s 5s 6s 7s 8s
Visually complete (%) 0 25 50 75 100 Time in
Seconds 0s 1s 2s 3s 4s 5s 6s 7s 8s
Visually complete (%) 0 25 50 75 100 Time in
Seconds 0s 1s 2s 3s 4s 5s 6s 7s 8s
Every page is fully loaded in 10 seconds 93% linear
after 1 second 100% linear after 10 seconds 20% linear after 9 seconds 7% linear after 9 seconds 80% linear after 1 second 881 5000 8610
The lower Speed Index: The lower the better
http://webpagetest.org
None
Critical CSS
None
This is where the content is loaded So let’s show
the user everything we’ve got so far
None
CSS is blocking. That means: No content shown until the
whole CSS is loaded
Open up your website at the desired viewport. Look at
the above the fold content.
Open a headless browser to find the CSS for that
content.
npm install --save-dev critical
critical.generate({ src: 'index.html', dest: 'styles/main.css', width: 1300, height: 900 });
critical.generate({ src: 'index.html', dest: 'styles/main.css', width: 1300, height: 900 });
critical.generate({ src: 'index.html', dest: 'styles/main.css', width: 1300, height: 900 });
http://jonassebastianohlsson.com/criticalpathcssgenerator/
<!doctype html> <html> <head> <style id=“critical”> </style> </head>
<script id=“load-main-css"> var ms=document.createElement(“link"); ms.rel=“stylesheet"; ms.href=“http://path/to/main.min.css”; document.getElementsByTagName(“head”)[0].appendChild(ms); </script>
None
None
One more optimisation
None
What happened?
On loading a new page… why should we (even if
it’s cached) check everything again and again?
Just load the things we need and tell the browser
where we are without loading all the burden
As an enhancement!
if(window.history && history.pushState) { … }
$('body').on('click', '#content a:not([href*=#]):not([href^=http])', function(e) { e.preventDefault(); retrieveContent(this.href); });
var retrieveContent = function(loc) { $.get(loc) .done(function(data) { updateContent(data, loc);
}) .fail(function(err) { window.location.href = loc; }); };
var updateContent = function(data, url) { var mkup = $(data);
url = url.toString(); history.pushState({ url: url, title: title }, mkup.filter('title'), url); $('#content').html(mkup.filter('#content').html()) .remove(); $('title').text(title); };
Speed Index equals regular Speed Index of the site without
JS and CSS
Take it even further…
Do contents change during the period of a session?
var updateContent = function(data, url) { var mkup = $(data);
url = url.toString(); history.pushState({ url: url, title: title }, mkup.filter('title'), url); $('#content').html(mkup.filter('#content').html()) .remove(); $(‘title').text(title); sessionStorage.setItem(url, data); };
var retrieveContent = function(loc) { if(sessionStorage.key(loc)) { updateContent(sessionStorage.getItem(loc), loc); }
else { $.get(loc) .done(function(data) { updateContent(data, loc); }) } };
Speed Index? Probably 0
+ What’s more?
Apdex Network latency Connection speed Bandwidth
THX! @ddprrt - fettblog.eu - workingdraft.de