Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Establishing Performance Contexts
Search
Ianfeather
September 28, 2014
Technology
140
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Establishing Performance Contexts
Segmenting your site to target specific user performance contexts
Ianfeather
September 28, 2014
More Decks by Ianfeather
See All by Ianfeather
Building Resilient Front End Systems (Smashingconf)
ianfeather
0
140
Building Resilient Frontend Systems (All Day Hey)
ianfeather
1
580
Building Resilient Frontend Systems (@frontendne)
ianfeather
1
210
Testing Without Assertions
ianfeather
0
150
Building Resilient Frontend Systems - NationJS
ianfeather
0
220
Reducing complexity with a Component API
ianfeather
0
210
Web Fonts and Performance
ianfeather
0
240
Other Decks in Technology
See All in Technology
ここは地獄!つらい朝会を体験することで、チームとしてのより良い振る舞いに気づくワークショップ / The stand-up meeting from hell in the game industry
scrummasudar
0
220
AI x 開発生産性を取り巻く予算戦略と投資対効果
i35_267
7
2.8k
SRENEXT_2026_Chairs__Talks_in_Tamachi.sre.pdf
srenext
1
140
Multicaで30個のミニプロジェクトをAIエージェント運用して見えてきたこと
eiei114
1
620
【公開用】AI_Dev_Ex2026_AI_登壇資料
matsuritechnologies
PRO
1
460
AICoEでAIネイティブ組織への進化
yukiogawa
0
210
「守りたい体験」を渡すだけで E2E を生成させられるようになった話
hinac0
2
1k
2026年のソフトウェア開発を考える(2026/07版) / Agentic Software Engineering 2026-07 Findy Edition
twada
PRO
24
10k
AIが当たり前の組織で エンジニアはどう育つか
nishihira
1
950
事業成長とAI活用を止めないデータ基盤アーキテクチャの設計思想
hiracky16
0
460
AI時代のPlaywright活用(システムテストを自動化する ー 実行エンジンにPla ywrightを選んだ理由)
ynisqa1988
2
930
伝票作成AIエージェントを支える、LLMOpsとインフラの選択肢 / AICon2026_takeda
rakus_dev
0
260
Featured
See All Featured
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
162
16k
How to Build an AI Search Optimization Roadmap - Criteria and Steps to Take #SEOIRL
aleyda
1
2.1k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
141
35k
The Cult of Friendly URLs
andyhume
79
7k
Claude Code どこまでも/ Claude Code Everywhere
nwiizo
65
57k
The Illustrated Children's Guide to Kubernetes
chrisshort
51
53k
Taking LLMs out of the black box: A practical guide to human-in-the-loop distillation
inesmontani
PRO
3
2.3k
End of SEO as We Know It (SMX Advanced Version)
ipullrank
3
4.3k
Statistics for Hackers
jakevdp
799
230k
SEO Brein meetup: CTRL+C is not how to scale international SEO
lindahogenes
1
2.8k
Making Projects Easy
brettharned
120
6.7k
Gemini Prompt Engineering: Practical Techniques for Tangible AI Outcomes
mfonobong
2
470
Transcript
Establishing performance contexts Ian Feather - @ianfeather
Performance at ! !
We don’t know how important speed is to our users
at a given time
None
None
None
None
It’s not that hard to get fast
It’s not that hard to get fast
It’s not that hard to get fast
It’s not that hard to get fast
None
You can use these techniques to optimise (to a certain
extent)
None
Total: weight: 674 KB ! Until you scroll down
the page…
None
!
Performance vs Design is all about compromise
Performance vs Design is all about compromise
You can only get as fast as the design allows
Performance Contexts
We have to accept there are things we don’t know
We are constantly required to make assumptions…
Assumption 1: Users on mobile devices want a different experience
to those on desktop
Assumption 1: Users on mobile devices want a different experience
to those on desktop Viewport !== context Viewport !== network
Assumption 2: Users want every site to load within X
seconds
None
None
None
None
Assumption 2: Users want every site to load within X
seconds Users want a site to load in-line with their expectations
Serving the same experience across multiple unknowns is a compromise
Currently we don’t have the tools to do anything about
this
But… ! The Web is for everyone
How can we determine a user’s performance context?
Why not ask them?
Hostel Mode
http://blog.chriszacharias.com/page-weight-matters
None
“We even burned through our monthly data plans in 40
minutes.”
Light Mode Default
Segmenting in to two ‘modes’ can be very powerful but…
Are we still making assumptions?
None
Did we ever ask users with retina screens if they
want a slower experience?
And are we still compromising?
We can’t leverage the best of the platform… whilst limiting
page weight.
None
Light Mode Default
Light Mode Default Enhanced
Segmenting allows us to design focused User Experiences
Asking the user
None
None
None
How could this look?
None
None
None
None
Implementation Advice
Don’t build three different websites!
Abstract out the development complexity
if('querySelector' in document && 'localStorage' in window && 'addEventListener' in
window && 'classList' in document.documentElement && lpHostelMode() !== "light") { } Global feature switches
<div class=“my-module”> <h1>{{ moduleTitle }}</h1> {{ img_tag(moduleImage) }} <p>{{ moduleContent
}}</p> </div> Authoring must stay the same
def img_tag(image_url, opts={}) return if @hostelMode ! image_tag(image_url, opts) end
Use helpers to abstract switches
Stateless vs Stateful
Consider the preloader
Future: Could the browser expose these user options?
Recap
You can only get as fast as the design allows
Your application may be entirely inaccessible to parts of the
world
Segmenting the user experience can be very powerful
Don’t guess. Ask.
Don’t do it at the expense of complexity
Thank you Ian Feather - @ianfeather
Creative Commons ! ! ! Arrano : https://flic.kr/p/4WjZDf Patrick Quinn-Graham
(thepatrick): https://flic.kr/p/Tk4Ha Leonardo Rizzi (L1010913): https://flic.kr/p/cVRiZm