<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Owen Campbell-Moore on Medium]]></title>
        <description><![CDATA[Stories by Owen Campbell-Moore on Medium]]></description>
        <link>https://medium.com/@owencm?source=rss-94a116f73676------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*xAZlL1PB1ygB78Ni0DRA9w.jpeg</url>
            <title>Stories by Owen Campbell-Moore on Medium</title>
            <link>https://medium.com/@owencm?source=rss-94a116f73676------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 20:40:10 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@owencm/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Reactive Web Design: The secret to building web apps that feel amazing]]></title>
            <link>https://medium.com/@owencm/reactive-web-design-the-secret-to-building-web-apps-that-feel-amazing-b5cbfe9b7c50?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5cbfe9b7c50</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[progressive-web-app]]></category>
            <category><![CDATA[react]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Mon, 03 Apr 2017 16:48:04 GMT</pubDate>
            <atom:updated>2017-04-03T16:48:04.314Z</atom:updated>
            <content:encoded><![CDATA[<p>In the last year I’ve observed <strong>two subtle techniques</strong> being used by some developers that take a web app from feeling slow and janky to highly reactive and polished.</p><p>I believe these techniques are important enough that they need a name: <strong>Reactive Web Design</strong>.</p><p>In summary, reactive web design is a set of techniques that can be used to build sites that always feel fast and responsive to user input regardless of the network speed or latency.</p><p>As web developers and framework authors, I believe finding ways to make these patterns default in everything we build is a top priority for improving UX and perceived performance on the web.</p><h3>Technique 1: Instant loads with skeleton screens</h3><p>When used well, this technique is almost never noticed, but has a huge impact on perceived performance of a site.</p><p>Interestingly, the technique is used by almost all native apps and makes them feel very reactive even on terrible networks, but it is almost never used on the web!</p><p><em>Opportunity this way lies!</em>💡</p><p>In short, skeleton screens ensure that whenever the user taps any button or link, the page reacts immediately by transitioning the user to that new page and then <em>loading in content to that page </em>as the content becomes available.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*tAz8By32NRGGypJIF6_sGg.png" /><figcaption>Facebook using a skeleton screen to improve perceived performance when you first open it</figcaption></figure><p>Skeleton screens are a key <em>perceived performance </em>technique as they make applications <em>feel </em>much faster, dramatically reducing the number of moments where the user is left wondering:</p><blockquote>What is going on? Is it even loading? Did I tap it right? <em>🤔</em></blockquote><p><a href="http://www.flipkart.com/">Flipkart.com</a> is a rare example of a website that makes use of this approach. Browsing through categories or tapping on products therefore feels lightning fast, even when the actual results take a few seconds to load:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/360/1*54hYHKhgcLpA5D9yOi0Waw.gif" /><figcaption>A screencapture of <a href="http://flipkart.com/">flipkart.com</a> launched from the home screen in standalone mode on Android</figcaption></figure><p>When this technique is used best, content that is already available such as thumbnails or article titles can be re-used to improve the perceived performance even further, making loads feel truly instant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*RivCwhp1kMqjD2AsM5-83A.gif" /><figcaption><a href="http://app.jalantikus.com/">app.jalantikus.com</a> uses the Skeleton Screens pattern and reuses titles and thumbnails across transitions</figcaption></figure><h4>Testing sites with skeleton screens</h4><p>Testing how well sites use this technique is easy: simply use Chrome network emulation to make the network as slow as possible and then click around a site. If it is doing this well, the site will still feel snappy and responsive to your input.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/310/1*NC3XoTYVOEbWmNDhttn4TA.png" /><figcaption>The slowest speed supported in Chrome network emulation</figcaption></figure><h3>Technique 2: “Stable loads” via predefined sizes on elements</h3><p>You know that feeling where a website is jumping around while you’re trying to use it? You’re just trying to read an article and the text keeps moving around? That’s what we call an “unstable load”, and we need to <strong><em>burn it with fire </em></strong>🔥🔥🔥.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*L_QFBt0rlRNkWvZoxMMmxA.gif" /><figcaption>slate.com content jumps around very aggressively as the page loads. The slower the network you’re on, the longer it jumps for.</figcaption></figure><p>Unstable loads make websites hard to interact with, and makes them feel… well… <em>Unstable!</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*W0UPZSTKYMWEHaNbBCyY4A.gif" /><figcaption>Browsing an unstable site always reminds me of how I imagine it would feel to walk around during an earthquake</figcaption></figure><p>Unstable loads are caused by images and ads embedded into a page but not including any sizing information. By default the browser only knows the size of these once they have loaded, so as soon as an image loads in, <strong>THUNK!</strong>,<strong> </strong>the whole page slides down 😡.</p><p>To prevent this, all &lt;img&gt; tags on a page must proactively include the dimensions of the image they will contain.</p><p>In many cases images used on certain pages are always the same size and so their size can simply be included in the HTML template, but in some cases the size of images is dynamic and thus their size should be calculated when the image is uploaded then templated into the HTML when it is created.</p><pre>&lt;!-- Always include sizes on images to prevent unstable loads --&gt;<br>&lt;img src=&#39;/thumbnail.png&#39; style=&#39;width: 100px; height: 84px&#39;&gt;</pre><p>The same is true for ads, often a culprit when it comes to unstable loads. Wherever possible, create a div that will contain an ad, and in your templating set it to be sized with your best guess at how big this ad will be.</p><p>Note that unstable loads are at their worst on slow networks as you have just settled into reading content when suddenly it jumps, and you can <em>never quite be sure that you’re safe</em>.</p><h3>Putting it all together</h3><p>I’ve build a small demo site at <a href="http://reactive.surge.sh/">reactive.surge.sh</a> to demonstrate the difference between conventional and reactive web design.</p><h4>Conventional article loading</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/320/1*vSbJQ42cy-3Q0IdJE8KDbg.gif" /><figcaption>Note how sluggish it feels and how frustrating content jumping is. Interestingly I find this orders of magnitude more annoying on mobile devices when tapping the screen and not seeing it react.</figcaption></figure><h4>Loading an article with reactive web design</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/320/1*xxGYpV_Qn82m9jJIYBqq-g.gif" /><figcaption>With reactive design the load feels instant and the site remains reactive when tapping the back icon and the article title multiple times</figcaption></figure><h3>Wrapping up</h3><p>The slower the network is, the worse the user experience becomes when page transitions block on the network and pages jump around for extended periods.</p><p>With <strong>Reactive Web Design</strong> we can make our experience feel snappy and responsive (“Responsive Design” as a name was already taken, d’oh!) even on slow and painful networks.</p><p>I’d love to hear about data from the community on the effect of perceived performance on KPIs such as engagement and revenue!</p><p>Additionally, I’d encourage framework and library authors to consider how to make skeleton screens and stable loads the default, also known as the <a href="https://blog.codinghorror.com/falling-into-the-pit-of-success/"><em>pit of success</em></a><em>.</em></p><p>If you have thoughts about this, please tweet me <a href="https://twitter.com/owencm">@owencm</a>, and if you enjoyed this please give it a ♥!</p><p><em>P.S. be sure to check out the demo site </em><a href="http://reactive.surge.sh/"><em>reactive.surge.sh</em></a><em> on a mobile device for it’s full glory!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5cbfe9b7c50" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One weird trick to performant touch-response animations with React]]></title>
            <link>https://medium.com/@owencm/one-weird-trick-to-performant-touch-response-animations-with-react-9fe4a0838116?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/9fe4a0838116</guid>
            <category><![CDATA[material-design]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[ui]]></category>
            <category><![CDATA[animation]]></category>
            <category><![CDATA[web-development]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Tue, 24 Jan 2017 02:04:26 GMT</pubDate>
            <atom:updated>2017-01-24T19:41:11.457Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve seen a single fatal performance issue on many sites, including Twitter.com and in high quality UI toolkits including Material UI.</p><p>This post includes the TL;DR on the issue and how to solve it, resulting in super performant animations or CSS changes that apply in response to touches.</p><h4>Overview of the problem</h4><p>Many sites implement pretty buttons or tabs that respond with some CSS change or animation when tapped. These look great during development when the onclick handler does nothing, and on fast Macbook Pros that can steam quickly through main thread work.</p><p>The issue is that these kind of touch-response animations often follow the following pattern:</p><ol><li>Detect a touch of some kind</li><li>Trigger a CSS toggle, start an animation etc</li><li>Call something like an onclick callback that triggers whatever logic the button exists to perform in the app (e.g. change some React state that causes some part of the app to be re-rendererd)</li></ol><p>The issue lies in between 2 &amp; 3, as changes to CSS and animations do not actually apply/start synchronously, but rather only start by the beginning of the next frame. This means that if step 3 blocks the main thread for any time, say 300ms, then the CSS change or animation will be blocked for 300ms resulting in your app feeling kludgy and slow.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/498/1*6oxL6urT99GUnH1rdCq99A.png" /><figcaption>WAIT!! I can feel some of you think you know this issue and are about to leave, but read on and I promise the solution is probably not what you think</figcaption></figure><h4>The commonly stated, but incorrect solution</h4><p>At this point experienced developers will put their <em>magic incantation hat</em> on and say: “<em>I recognize this, we solve it by calling the callback in </em><strong><em>requestAnimationFrame()</em></strong><em> which ensures the animation has started (or CSS change is committed) before the callback is even called.</em>”</p><p>Unfortunately, due to <a href="https://bugs.chromium.org/p/chromium/issues/detail?id=675795">a bug in Safari and Chrome</a>, this will not solve the problem (!)</p><p><em>If your inner voice is screaming right now, don’t worry, mine is too.</em></p><p>You can prove to yourself that this solution fails by creating a small repro where in response to a touch event you toggle a CSS class and then pass a function to <strong>requestAnimationFrame</strong> that blocks the thread for 1s.</p><h4>The actual solution</h4><p>The actual solution is to make the CSS changes or start the animation and then use a <em>double </em><strong>requestAnimationFrame</strong> to call any work that may block the main thread. This guarantees that your animation will have started, and will dramatically improve the performance of your application.</p><p>The solution will look something like:</p><pre>requestAnimationFrame(() =&gt; {<br>  requestAnimationFrame(onButtonClicked)<br>})</pre><blockquote><em>Now go forth, and make your interactions feel great!</em></blockquote><p><em>Update (2017/01/24): I’ve now written up </em><a href="https://medium.com/@owencm/well-there-are-definitely-issues-here-ordered-least-significant-to-most-1f0931b03024#.f9qigf9n6"><em>some thoughts on whether this actually is a bug or is how things should be</em></a><em> following some great discussion on Twitter.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9fe4a0838116" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Issues with Open Whisper Systems approach to Private Contact Discovery]]></title>
            <link>https://medium.com/@owencm/issues-with-open-whisper-systems-approach-to-private-contact-discovery-190eb9c5168?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/190eb9c5168</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[privacy]]></category>
            <category><![CDATA[browsers]]></category>
            <category><![CDATA[whatsapp]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Sat, 21 Jan 2017 01:42:10 GMT</pubDate>
            <atom:updated>2017-01-23T20:54:39.804Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>Context: </em></strong><em>I am currently researching Private Contact Discovery in attempt to determine if there’s a feasible way to bring an API to enable this to the Web Platform. I believe it’s important for us to enable social applications such as future messaging clients to be built on the Open Web, and for these to be able to bootstrap their social graph from the local contact book.</em></p><p>This short post discusses the Open Whisper Systems post ‘<a href="https://whispersystems.org/blog/contact-discovery/">The Difficulty Of Private Contact Discovery</a>’, a classic in this space that I have referred to often and am very grateful to them for publishing.</p><h4>TL;DR on the problem we’re trying to solve</h4><p>A user, Alice, is signing up to a social service run by Bob, and would like to discover which contacts in her contact book have signed up to Bob’s service, so Alice can add/follow/message them etc.</p><h4>An attack on the <a href="https://whispersystems.org/blog/contact-discovery/">Open Whisper Systems (OWS) proposals</a></h4><p>So far as I’m aware, most privacy-preserving intersection schemes assume that each participant has a set of private data, not known to the other participant. If this was true in the Private Contact Discovery case, then I think the suggested OWS approaches would work well, but <strong>this is not the case.</strong></p><p>Consider for illustrative purposes an attack where a malicious actor, Mallory, provides a social service that Alice wishes to sign up to and perform Private Contact Discovery with.</p><p>Mallory, noticing that the phone numbers are enumerable (they’re just numbers with special formatting, after all), enumerates every valid phone number and “pretends” (for the purposes of these schemes) they every phone number is associated with a user signed up to the service.</p><p>Regardless of the scheme proposed in the OWS post, I believe that Alice’s entire contact set would be exposed to Mallory.</p><h4>Does this attack work?</h4><p>What do you think? Does this attack work? Did I miss something?</p><p>I note that the OWS approach <em>does</em> protect against the threat model of malicious eavesdroppers and National Security Letters, so perhaps my threat model (malicious service) was simply implicitly not the same threat model considered by them, but naturally I don’t want to have to <em>trust</em> services run by strangers online.</p><h4>My personal conclusions</h4><p>I have spent a long time exploring potential solutions to this problem and am fairly convinced that phone numbers have a few properties, such as their enumerability, that mean no system will ever be able to perform anything like perfect Private Contact Discovery.</p><p>That said, I believe there are a number of <em>OK-</em>to-<em>pretty reasonable </em>approaches that do better than simply sharing contact books directly. I believe that we as a community are going to have to come to a conclusion about whether the value of being able to discover people we know on social services is worth the risk.</p><p>I personally believe that if a platform can guarantee users give explicit permission then the value is worth the risk imposed by <em>OK </em>schemes to solve this problem.</p><p>I hope to share some of these alternate approaches that I’m aware of in a follow up post when I get time, but for now wanted to open the discussion on the above and hear what you all think.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=190eb9c5168" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Rethinking URL bars as primary browser UI]]></title>
            <link>https://medium.com/@owencm/rethinking-url-bars-as-primary-browser-ui-e2118339d2c0?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/e2118339d2c0</guid>
            <category><![CDATA[domains]]></category>
            <category><![CDATA[browser-security]]></category>
            <category><![CDATA[design-in-the-browser]]></category>
            <category><![CDATA[computer-security]]></category>
            <category><![CDATA[browsers]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Fri, 30 Sep 2016 15:46:03 GMT</pubDate>
            <atom:updated>2016-09-30T20:51:25.589Z</atom:updated>
            <content:encoded><![CDATA[<p>This post consists of my notes on why web browser vendors show our users the URL at the top of every screen. It exists to capture the security and usability constraints that affect URL bars and examine what alternatives may also satisfy those constraints.</p><h4>Why do we show URLs at the top of every page?</h4><p>A URL is made up of an origin and a path. These strings are <em>the very foundation of the Web, </em>allowing us to seamlessly link content together.</p><p>Let’s now put on our inquisitive hat and ask ourselves: why <em>do</em> we show users the URL at the very top of every page? Is that the UI we would design if web browsers were invented today?</p><p>I’d say we do it for 3 main reasons (beyond the historical reason that “we’ve always done it that way”, which shouldn’t be underestimated).</p><p>We do it:</p><ol><li>So users understand who they’re talking to</li><li>So users can know <em>where</em> on a site they are, and navigate by modifying the URL</li><li>So users can share a link to what they’re currently looking at</li></ol><p>In this post I will unpack these one at a time, attempting to capture the underlying constraints implied by these use cases, and thus what other potential approaches could satisfy them</p><h4>Reason we show URLs #1: so users understand who they’re talking to</h4><p>Let’s start with a thought exercise. Off hand, can you confidently state the origin used by Bank of America?</p><p>bofa.com? bankofamerica.com? bankofamerica.bank?</p><p>Considering this question led to me to the following conclusion:</p><blockquote>When we want to find an entity online for the first time we don’t just magically know the correct origin — we use a search engine.</blockquote><p>We start by searching for an entity, and subconsciously trust that the top result is actually the entity we were looking for.</p><p>If we can’t reasonably assume that the top result is who we’re looking for then for all intents and purposes the battle for security is already lost.</p><p>Thus we reach our first, important conclusion of search-engine-dependence:</p><blockquote>Our ability to understand who we’re talking to on the web is underpinned by search engines.</blockquote><h4>When you revisit the origin later</h4><p>The next time we want to access the same site we just type its name into the URL bar and trust either autocomplete or returning to the original search engine will do the rest.</p><p>If we somehow end up on a site via a link and wonder who it is, we now have a look at the URL bar and attempt to recall whether the origin is the same as it was when we landed on the site originally via a search engine.</p><p>This approach is theoretically secure, but assumes and requires that our users are technical, paying full attention and have a good memory. <em>But most of our users</em> —<em> the billions of people that use web browsers</em> — <em>are not, they are just trying to get something done online.</em></p><h4>So who are our users?</h4><p>Billions of users are coming online for the first time via mobile devices in emerging markets. Their first experience is one of a grid of logos labeled with the words “Facebook” and “WhatsApp”.</p><p>If we’re lucky these new users venture to the browser. At this point they’re suddenly presented with a distinctly confusing box containing runs of characters like “https://”, “www.”, “.co.in”. The simple world of logos and app names is gone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*Q55uWfDeF6IwWQGNU6fdrg.png" /><figcaption>To new users of the web, the URL bar is about as meaningful as this**</figcaption></figure><p>Many users manage to fumble through using search, and features like Safe Browsing save them from some phishing attacks, but the UI is not very user friendly.</p><h4>Can we better serve our users needs to understand who they’re talking to?</h4><p>Suspending implementation questions for one moment, let&#39;s consider a hypothetical question about how we could improve the UX that conveys the origin’s “content”.</p><blockquote>When a user is on Facebook, why doesn’t the URL bar just say “Facebook”, with a small Facebook logo? Would that not more clearly help the user understand the literal “origin” of the content?</blockquote><p>You may now reasonably think: “sure, build that. Then I’ll immediately make a fake site named Facebook with the same logo and how foolish you will look!”</p><p>Clearly we would need a way for browsers to determine the locally known brand name and logo of all sites*. Let’s call such a hypothetical system an <em>origin-to-local-brand-mapping</em>.</p><p>Now recall how earlier in this post we established search-engine-dependence: the fact that our ability to understand who we’re talking to on the web is already underpinned by search engines.</p><p>Therefore, I think it’s worth exploring whether we can use search engines to build such an <em>origin-to-local-brand-mapping </em>system.</p><p>Now it’s time to move on from helping our users understand <em>who </em>they’re talking to. Let’s talk about navigation.</p><h4>Reason we show URLs #2: So users can know where on a site they are, and navigate by modifying the URL</h4><p>Billions of users are currently in the process of coming online on mobile devices for the first time.</p><p>If you personally manually modify URLs to browse around a site and parse the parameters to determine where you are, then you are likely not in the majority of our users.</p><p>You do deserve the right to this ability, but I’d argue we should optimize for the 99%, rather than the 1% of URL-hackers.</p><h4>Reason we show URLs #3: So users can share what they’re looking at</h4><p>The ability to share what we’re looking at is extremely important. Sharing and navigating is at the core of the web and is fundamentally enabled by URLs and the web’s principles of Security, Linkability, Indexability, Composability and Ephemerality, collectively known as <a href="https://paul.kinlan.me/slice-the-web/">SLICE</a>.</p><p>But let’s think for a moment of reasons #1 and #2 that we show URLs. If #1 is — at least theoretically — better served by rendering the locally known brand and we can agree that #2 shouldn’t be a top priority for the precious pixels that we have on mobile, then it makes me wonder:</p><blockquote>Is the current UX of an editable text box persistently containing a URL the best approach to help people share what they’re looking at?</blockquote><p>Could we instead build a persistent “share” or “get link” button? Would that work too?</p><h4>My personal conclusions</h4><ul><li>Users need a way to share a link to what they’re currently looking at, but that doesn’t necessarily need to be in the form of a text box already containing the link</li><li>Origins are not very user-friendly, especially for users coming online for the first time on mobile</li><li>Search engines are already implicitly trusted when it comes to web security as they tell us what the origin should be</li><li>If we could determine the locally known brand name and logo for a subset of sites we may be able to build a more user-friendly, and hence safer, browsing experience</li></ul><p>Given the above, I think there’s a real and exciting possibility to revolutionize browser UI.</p><p>That said, this idea is all underpinned by being able to build an origin-to-local-brand-mapping system for at least some subset of the web. Is that even remotely possible? Assuming you had full access to the knowledge of search engines, or otherwise, how would you design such a system?</p><p>Responses or Tweets (I’m <a href="https://twitter.com/owencm">@owencm</a>) much appreciated. Let’s explore these constraints and opportunities together!</p><p><em>Thanks to Emily Schechter, Alex Ainslie, Alex Komoroske, Elizabeth Harwood, and the many awesome folk on Chrome’s Enamel and Fizz teams for the insightful discussions and feedback on this post.</em></p><h4>Footnotes*</h4><p>*To some extent EV certificates aimed to solve this problem, but company legal names rarely correspond to known brands. Last I checked the mint.com EV cert read “Intiuit USA Corp”. Not exactly user-friendly.</p><p>**This line of questioning also begs the question of why we ever show “www.” or “m.” in origins. Are there any cases where that meaningfully aids users in identifying that they are talking with who they mean to be talking with? They are additional confusing syntax and have a real cost. Are they delivering sufficient value to remain?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e2118339d2c0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The surprising tradeoff at the center of the question whether to build a Native or Web App]]></title>
            <link>https://medium.com/@owencm/the-surprising-tradeoff-at-the-center-of-question-whether-to-build-an-native-or-web-app-d2ad00c40fb2?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2ad00c40fb2</guid>
            <category><![CDATA[mobile]]></category>
            <category><![CDATA[progressive-web-app]]></category>
            <category><![CDATA[web-development]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Thu, 08 Sep 2016 16:58:19 GMT</pubDate>
            <atom:updated>2016-11-14T18:05:52.223Z</atom:updated>
            <content:encoded><![CDATA[<p>The choice to build a native app over a website is often made because apps “feel” so much nicer than web sites and that “<em>they’re the future”, </em>but here’s a secret I want to share with you:</p><blockquote>Great UX isn’t worth anything if you can’t acquire users</blockquote><p>And it is <em>way</em> harder to get users to try your native app than your web site.</p><h4>Acquiring users is much easier on the web</h4><p>On average, just <a href="https://splitmetrics.com/blog/whats-a-good-app-store-page-conversion-rate/">26.4% of users that visit your app store page will install your app</a>. The other 73.6% of users are now lost to you and didn’t even get to try your app.</p><p>According to a study by ComScore, <a href="http://www.slideshare.net/VictorKongCisneros/the-2015-us-mobile-app-report-by-comscore">the top 1000 mobile websites receive 270% more monthly visitors than the top 1000 native apps</a>. This data makes it clear that native apps are leaving users on the table.</p><p>At this point things aren’t looking too good for native apps, but that’s nowhere near the full story.</p><h4>You can’t acquire as many app users, but they engage heavily</h4><p>According to the same ComScore data, the average user spends ~18.5X more time than in the average top 1000 native app than in the equivalent mobile web app (<a href="http://venturebeat.com/2016/04/06/why-pinterest-forces-you-off-its-mobile-site-and-into-its-app/">Pinterest found this number to be ~3X and Yelp found it to be ~10X</a>). There are are couple reasons for this difference:</p><ol><li>Apps are easily launched from the user’s home screen</li><li>Apps can send engaging push notifications</li><li>Apps generally have better UX</li><li>Companies generally invest more in their apps than their web property</li><li><strong>These metrics only consider the time spent by the 26.4% of potential users that chose to install the app. <em>76.3% of potential users spent exactly zero time in the app. Hence these engagement numbers are heavily inflated by a selection bias</em></strong></li></ol><p>Regardless of the selection bias, there seems to be a very real tradeoff between <strong>reaching </strong>users are <strong>engaging </strong>them when thinking about platform strategy.</p><blockquote>Conventional web sites let you <strong>reach </strong>users, while apps let you deeply <strong>engage </strong>users. This is the heart of the tradeoff between mobile web vs apps.</blockquote><p>The most important question is therefore: for <strong>your </strong>product specifically, what is the optimal way to trade off <strong>reach </strong>and <strong>engagement</strong>?</p><h3>Making the tradeoff for your product</h3><p>The key to answering this trade off is to ask yourself:</p><blockquote>What interactions within your product ultimately create value<em> </em>for your company?</blockquote><p>Do users “create value” on their first visit (e.g. by viewing ads)? And is the session value of a returning user very similar to that of a new user (e.g. e-commerce, publishing)? If so, then <strong>reaching</strong> as many users as possible is the most important factor.</p><p>Alternatively, do users only become valuable to you after they have used your product for a while (e.g. messaging, social networks)? If so then you’ll want to focus on <strong>engaging </strong>a smaller group of potential customers.</p><p>(I have another post in mind that digs into this question from a more mathematical perspective, leave a highlight if you’d like to read that and I’ll work on it!)</p><h3>What if we could get the best of both?</h3><p>In an ideal world you’d get the reach of a web app <em>and </em>the full engagement of a native app, but unfortunately that maximal point is never obtainable.</p><p>This is because app engagement numbers <em>have selection bias built in</em> and because your target customers <em>aren’t all equally interested in your product.</em></p><p>Some of your target customers will want to use your product infrequently, or even only once while some users will want to use your product every day.</p><blockquote>Additionally, almost all of your users won’t know if they will want to use your product every day until they have tried it.</blockquote><p>To make progress from here it’s helpful to think about interaction with a product as going through <em>two distinct phases</em>:</p><ol><li><strong>New or low engagement users</strong> that don’t know much about your product, but are hopefully willing to try it and engage a little</li><li><strong>Engaged and convinced users </strong>that have decided they like your product and are willing to commit to using it more often</li></ol><h4>Conventional web-only or app-only strategies</h4><p>A conventional web-only or app-only strategy would each only address one of the above phases, which is almost always suboptimal.</p><p>The only case in which a web-only strategy could make sense are in certain verticals (e.g. publishing and e-commerce) where there are so many potential new or low engagement users that it is optimal to focus all resources towards those users, rather than focusing on the ‘whales’, a much smaller group of users that each consume very heavily. This is changing though, as we’ll see in the section below ‘<strong>Innovative strategies to address users in both phases</strong>’.</p><h4>An app-only strategy approximately never makes sense</h4><p>By observing the strategy of almost every <em>successful</em> mobile product is should be clear that approximately none of them use an app-only strategy, and those that do are increasingly <a href="https://www.techinasia.com/flipkart-myntra-app-only-disaster">realizing that it isn’t working</a> which should surprise nobody once you make the realization that the users in phase two start out as <strong>new or low engagement users.</strong></p><p>Some companies <em>have</em> made a <strong>desktop web + native app strategy</strong> work by first engaging users on desktop as a way to transition them from phase one to phase two and then convert them into installing an app. In that case it is possible that on mobile a hard sell for a native app <em>can</em> be optimal, as <a href="http://venturebeat.com/2016/04/06/why-pinterest-forces-you-off-its-mobile-site-and-into-its-app/">was found by Pinterest</a>, although this strategy almost certainly depends on having a pre-established brand and a well known value proposition.</p><p>While app-only may be a reasonable way to first demonstrate product value, very quickly it will be necessary to expand to the web in order to increase growth. (At this point I’d predict the engineering team will realize how much easier it is to ship new code and experiment on the web and regrets not doing this earlier)</p><h4>The conventional web and app strategy</h4><p>The most conventional strategy is to build both web and app versions. This allows you to <strong>reach</strong> large groups of users with your product and then once some users are convinced upsell them to your app, where they will be <strong>engaged</strong> and retained. This is the strategy used by companies such as GrubHub, Twitter, LinkedIn and others.</p><h3>Innovative strategies to address users in both phases</h3><blockquote>The main problem with the conventional web + app strategy is that it requires you to build your product three times: iOS, Android and web</blockquote><p>In the past year a new option has emerged. The major web browsers (with the notable exception of Safari) have centralized around a new set of web technologies which collectively enable developers to build <a href="https://developers.google.com/web/progressive-web-apps/">Progressive Web Apps</a>.</p><blockquote>Progressive Web Apps are fast-loading web apps that can be accessed with low friction just like any other web app, but on Android can also be added to the user’s home screen, work offline, and use push messaging.</blockquote><p>With a Progressive Web App strategy, the product can <strong>reach</strong> the large number of new or low engagement users with their Progressive Web App, and then offer engaged or convinced users to <em>install it </em>or<em> enable push notifications, </em>driving the <strong>engagement</strong> benefit of a native app but with the same single platform.</p><p>Unfortunately Safari doesn’t currently support the installation, push notifications or offline features of Progressive Web Apps, so upselling to an iOS native app is still important to address the segment of engaged or convinced iOS users.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*efmFpzS7q3E11ZNYNx3JRQ.png" /></figure><p>You may want to try the following examples to get a sense of how Progressive Web Apps can work: <a href="http://flipkart.com/">Flipkart</a>, <a href="https://www.washingtonpost.com/pwa/">Washington Post</a>, <a href="https://www.aliexpress.com/">AliExpress</a>.</p><h4>Benefits of the Progressive Web App strategy</h4><p>This strategy has a number of benefits:</p><ul><li>You can address users in both phases across all three platforms (Desktop, iOS and Android) with one fewer implementation of your product rather (as the Progressive Web App can take the place of the Android native app)</li><li>Investments in the Progressive Web App now pay off across all three platforms. For example, when AliExpress built a Progressive Web App they saw an <a href="https://developers.google.com/web/showcase/2016/aliexpress"><strong>82% increase in conversion rates on iOS</strong></a> compared to their old mobile web site</li><li><a href="https://developers.google.com/web/showcase/2016/konga">Progressive Web Apps consume <strong>92% less data</strong> compared to a native app</a>, allowing them to better address users with slower phones or limited storage</li><li>Progressive Web Apps allow engineering teams to maintain a high development velocity thanks to fast deploys and easy experimentation on the web compared to app stores</li></ul><h4>Drawbacks of the Progressive Web App strategy</h4><ul><li>Today, browsers tightly control your ability to prompt users to install the Progressive Web App. While there have been hints that this may be getting easier, Chrome users must have already spent 5 minutes in the Progressive Web App before you can prompt them to install it</li><li>While you can buy ads on Facebook to direct users to try your Progressive Web App, the Facebook WebView doesn’t support the installation or push notification opt in flow, and you can’t take advantage of the app-install ad format</li></ul><h3>Measuring your platform strategy</h3><p>With this understanding of the <strong>reach </strong>and <strong>engagement</strong> benefits of different platforms in mind, the final question is how to measure how your platform strategy is working.</p><p>Rather than fallacious approaches like simply comparing average engagement between platforms, the key is actually fairly simple: run A/B tests!</p><p>Use A/B tests to measure the expected lifetime value (in engagement, dollars, etc) of your users in different experimental groups such as:</p><ul><li>Letting users experience the web version uninterrupted</li><li>Showing a small banner encouraging users to install the native app or Progressive Web App</li><li>Letting users interact with the Progressive Web App but require they install it (or the native app) at key conversion points in the flow</li></ul><h3>Wrapping up</h3><p>When you go out into the world to build your product, remember that while your native app experience may be very appealing that…</p><blockquote>Great UX isn’t worth anything if you can’t acquire users</blockquote><p>and companies like <a href="https://www.flipkart.com/">Flipkart</a> have proved that you can build <a href="https://medium.com/@owencm/designing-great-uis-for-progressive-web-apps-dd38c1d20f7">great app-like UX</a> with Progressive Web Apps.</p><p>Remember that when it comes to <strong>reach </strong>and <strong>engagement</strong>:</p><ul><li>In most cases a hybrid web/app strategy works best to <strong>reach </strong>a large group of phase one users and then upsell those interested to phase two, where the average <strong>engagement</strong> per user will go up dramatically</li><li>The exact balance of web and app that works best for you will depend on your business model and whether you can derive value from users that visit infrequently (such as advertising-supported products that don’t have high sign-up/set-up friction)</li><li><a href="https://github.com/TalAter/awesome-progressive-web-apps">Progressive Web Apps</a> are an innovative new way to achieve this strategy with lower costs</li></ul><p>That’s it!</p><p>Don’t forget to give a ♥, and for more, <a href="https://twitter.com/owencm">follow me on Twitter (@owencm)</a> and Medium.</p><p><em>This post is shared under </em><a href="https://creativecommons.org/licenses/by-sa/4.0/"><em>Creative Commons Attribution-ShareAlike 4.0</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2ad00c40fb2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Designing Great UIs for Progressive Web Apps]]></title>
            <link>https://medium.com/@owencm/designing-great-uis-for-progressive-web-apps-dd38c1d20f7?source=rss-94a116f73676------2</link>
            <guid isPermaLink="false">https://medium.com/p/dd38c1d20f7</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[progressive-web-app]]></category>
            <dc:creator><![CDATA[Owen Campbell-Moore]]></dc:creator>
            <pubDate>Sat, 21 May 2016 23:42:25 GMT</pubDate>
            <atom:updated>2016-09-06T17:20:37.643Z</atom:updated>
            <content:encoded><![CDATA[<p>Progressive Web Apps give us an opportunity to reset our expectations and to loudly declare that We Can Do Better when designing user experiences on the web.</p><h3>The keys to great Progressive Web App UX</h3><ol><li>Always be testing on real world hardware</li><li>Take UX inspiration from native apps</li><li>Use the following checklist to ensure you’ve avoided common errors</li></ol><h3>Always be testing on real world hardware</h3><p>When you start each new project, find the crappiest mobile device you can and set up <a href="https://developer.chrome.com/devtools/docs/remote-debugging">remote debugging with Chrome</a>. Test every change to ensure you start fast and stay fast.</p><h3>Take UX inspiration from native apps</h3><p>It’s totally possible to take a poor mobile website and slap on service worker caching to improve performance. This is worth doing, but falls well short of providing the full benefit of a Progressive Web App.</p><p>When it comes to visual design and interaction patterns, start by forgetting everything you know about conventional web design, and instead imagine you’re actually designing a native app. It’s important to pay attention to detail here since native apps have given users expectations around touch interactions and information hierarchy which are important to match to avoid creating a jarring experience.</p><p>Try related native apps on iOS and Android and browse sites like <a href="https://dribbble.com/">Dribbble</a> for design inspiration. Spend some time browsing the <a href="https://www.google.com/design/spec/material-design/introduction.html">Material Design Specification</a> to level up your familiarity with common UI components and their interaction.</p><h3>Use a checklist</h3><p>There are a number of UX checklists for native apps that are worth Googling and verifying. I’ve personally noticed a number of common mistakes Progressive Web App developers make, which have been condensed down into the following recommendations:</p><h4><strong>Recommendation 1: Screen transitions shouldn’t feel slow due to blocking on the network</strong></h4><p>When a user taps a button or link in a web app, especially a server rendered web app, they are often made to wait staring at the current screen before suddenly jumping to a whole new screen of content.</p><p>This feels unresponsive and is unacceptable in a PWA (Progressive Web App) since betrays the idea that the app itself is running locally on the device.</p><p><strong>Aim to</strong><em> </em>provide the perception that the whole “app” is stored locally on the phone and that only the content is being dynamically loaded.</p><p><strong>Example of this done well:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/370/1*JkSwFcMnrBONQCPsgWGXCg.png" /></figure><p>When a user taps on an article in <a href="http://app.jalantikus.com/">Jalantikus</a>, the PWA <em>immediately</em> navigates the user to the article page without blocking on the network. The user is shown the thumbnail and title which were already available locally, and then the article content and a higher resolution image fade in when they become available.</p><p><strong>Tip</strong>: use a cache-first network strategy wherever possible as network-first strategies cause severe performance issues on slow or flaky networks.</p><p><strong>Tip</strong>: Aditya Punjani gave a <a href="https://www.youtube.com/watch?v=m2tvYGCdOzs">talk</a> about how they avoided this problem in the <a href="http://www.flipkart.com/">Flipkart Lite</a> PWA by using the App Shell architecture with React.</p><h4>Recommendation 2: Tappable areas should give touch feedback</h4><p>When a button or tappable area is tapped, the user shouldn’t be left wondering if the tap was recognized.</p><p><strong>Aim to </strong>make it immediately clear that the taps were recognized. For example this can be done by highlighting the tapped region in grey, showing a material ripple or having a page transition start immediately.</p><p><strong>Tip: </strong>by default some browsers show<em> their own </em>highlight when elements are tapped. This default styling is unattractive and may conflict with your custom feedback (for example, <a href="https://elements.polymer-project.org/elements/paper-button?view=demo:demo/index.html&amp;active=paper-button">Polymer Paper Buttons</a> flash blue when tapped in Chrome for Android 😞), so either customize it or disable it by setting:</p><pre>-webkit-tap-highlight-color: rgba(0,0,0,0);</pre><p><strong>Example of touch feedback done well:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/370/1*SciS9goCxhROIYcUz4jBSQ.png" /></figure><p>The Google I/O 2016 web app provides an immediate grey highlight when a talk is tapped and the default blue tap color is disabled.</p><h4>Recommendation 3: Touching an element while scrolling shouldn’t trigger touch feedback</h4><p>When implementing the touch feedback it is important that it doesn’t trigger mistakenly when the user is only touching an element to scroll the page.</p><p><strong>Example of touch feedback given at the wrong time</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/320/1*A6CG3A1AYTyGFASTtjxqfw.gif" /></figure><p>An early version of the popular <a href="http://www.material-ui.com/">Material-UI</a> library triggered ripples on touchable elements when they were touched during scrolls. The issue was fixed by aborting the ripple animation if the user’s touch location moved more than 6 pixels in any direction within the first few hundred milliseconds.</p><h4>Recommendation 4: Content shouldn’t jump as the page loads</h4><p>All img tags on a page<em> </em>should proactively include the dimensions of the image. This allows the browser to layout the screen correctly even before the image is loaded.</p><p>If dimensions aren’t specified on the tag then content will jump when the image is downloaded, which makes for a poor user experience.</p><p><strong>Tip:</strong> show a placeholder where the image will go, such as a grey square, or a blurred thumbnail of the image if it is available on the client due to being shown on a previous screen. Fading in the image when it is downloaded is also a nice touch of polish.</p><h4>Recommendation 5: Pressing back from a detail page should retain scroll position on the previous list page</h4><p>When a user selects an item from a list and is viewing it’s details, tapping back should bring back to the list page <em>with the same scroll position they were at before they pressed on the item.</em></p><p>It is very difficult to browse a list of articles, products or other items when pressing back always returns the user to the top of the list.</p><p><strong>Tip: </strong>routers built into some frameworks may provide this functionality by default for you, but verify that they do.</p><h4>Recommendation 6: Buttons and ‘non-content’ shouldn’t be selectable</h4><p>In Chrome, single taps on text are sometimes interpreted as selections. This creates the bug that taps on buttons and other ‘non-content’ are sometimes identified as selections, resulting in a strange experience.</p><p>To resolve this issue, mark non-content elements with the following CSS property:</p><pre>user-select: none;</pre><p><strong>Example of non-content accidentally being selected</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tBVRKx7JpjGTd3yZEXcjIQ.png" /></figure><p>Here you can see an icon on a button has been accidentally selected by the user, resulting in a weird experience where the expected action doesn’t occur and the ‘contextual search’ feature in Chrome appears at the bottom of the screen.</p><h4>Recommendation 7: Ensure inputs aren’t obscured by keyboard</h4><p>Sometimes when a user taps on an input field the keyboard will be brought up in such a way that it obscures the input. Be sure to try scrolling to different positions on the page and then tapping the input element to be sure it is never obscured.</p><p><strong>Tip</strong>: If the element is being obscured, explore using features like Element.scrollIntoView() to solve it.</p><h4>Recommendation 8: Provide an easy way to share content</h4><p>PWAs are often shown in a context where the current URL isn’t easily accessible, so it is important to ensure the user can easily share what they’re currently looking at. Implement a share button that allows users to copy the URL to the clipboard, or share it with popular social networks.</p><p><strong>Tip</strong>: In many browsers you can <a href="https://developers.google.com/web/updates/2015/04/cut-and-copy-commands?hl=en">automatically put content into users clipboards</a>. Providing a share button that copies the URL of the current page to the clipboard is a nice model.</p><p><strong>Tip: </strong>If you’re implementing social sharing buttons, be sure to delay the loading of the 3rd party JavaScript until after the primary content in the page has loaded.</p><h4>Recommendation 9: Use system fonts</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*1koL68q47csvJkH0jO2UnA.png" /><figcaption>Each platform has a recommended font. Using it in your PWA can make the app feel more at home.</figcaption></figure><p>Users will feel more at home in your Progressive Web App if you tweak it’s style to match their operating system. You can go so far as to implement different buttons and menus per platform, or simply get a head start by using the system’s preferred font:</p><pre>/* iOS */<br>font-family: &#39;Helvetica Neue&#39;, Helvetica, Arial, sans-serif;<br><br>/* Android */<br>font-family: &#39;RobotoRegular&#39;, &#39;Droid Sans&#39;, sans-serif;<br><br>/* Windows Phone */<br>font-family: &#39;Segoe UI&#39;, Segoe, Tahoma, Geneva, sans-serif;</pre><h4>Recommendation 10: Avoid overly “web-like” design</h4><p>“Classic” web design depends heavily on links in text, and static elements available on every page such as a header and a footer. These design elements are rarely found in native apps, so in order to match user’s mental models it is recommended to avoid these and use more app-like design instead.</p><p><strong>Example: </strong>Links should be used sparingly. Use careful placement of buttons, or tappable regions of cards instead.</p><p><strong>Example: </strong>Static page footers that contain navigation links and copyright information should be avoided. Take inspiration from native app design to see how similar information can moved into a menu or otherwise not included on every page.</p><h4>Recommendation 11: Touch interactions should be implemented very well, or not at all</h4><p>It is notoriously difficult, although not impossible, to implement advanced touch interactions such as pull to refresh and swipe to dismiss navigation menus.</p><p>If your app design requires advanced touch interactions, ensure the implementation is battle hardened and works well on real devices. Generally this means finding an implementation of this feature that somebody else has already crafted with great effort, such as the excellent <a href="http://oliviertassinari.github.io/react-swipeable-views/">react-swipable-views</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*xhWU4KyzM6mp9ncNVuGMIA.png" /><figcaption>A bottom navigation bar doesn’t suggest the user is able to swipe left and right to move between sections</figcaption></figure><p>Often the problem can be avoided altogether modifying the design so it doesn’t suggest advanced touch interactions are possible, for example by using a <a href="https://www.google.com/design/spec/components/bottom-navigation.html">bottom navigation bar</a> in place of <a href="https://www.google.com/design/spec/components/tabs.html#">tabs</a>.</p><p><strong>Tip: </strong>Much has been written about how <a href="https://lmjabreu.com/post/why-and-how-to-avoid-hamburger-menus/">‘Hamburger Menus’ should be avoided</a> for UX reasons. They have even been <a href="http://www.youmobile.org/blogs/entry/Google-Updates-the-YouTube-App-and-removing-the-Hamburger-Menu">stripped out of a number of Google apps</a>. Given the additional challenge of implementing their swipe interactions on the web, I would recommend against using them in Progressive Web Apps entirely.</p><h3>We can build amazing user experiences in Progressive Web Apps</h3><p>I’m excited that Progressive Web Apps allow us to build a new neighborhood on the web with <a href="https://en.wikipedia.org/wiki/Broken_windows_theory">no broken windows</a>.</p><p>By testing on real devices, taking inspiration from native app designs and applying checklists we can build truly amazing experiences on the web.</p><p>Follow me on Twitter <a href="https://twitter.com/owencm">@owencm</a> for more about Progressive Web Apps.</p><h4>Credits</h4><p>Thanks to the Chrome XFN team and the <a href="https://github.com/GoogleChrome/lighthouse">Lighthouse team</a> for all your inspiration.</p><h4>Related links for further reading</h4><ul><li><a href="http://pwa.rocks/">PWA Rocks</a> — a collection of Progressive Web Apps</li><li><a href="https://developer.apple.com/library/ios/documentation/UserExperience/Conceptual/MobileHIG/">iOS human interface guidelines</a></li><li><a href="http://pttrns.com/">PTTRNS</a> — example UIs for inspiration</li><li><a href="https://www.google.com/design/spec/patterns/empty-states.html">Android design patterns</a></li><li><a href="https://developer.android.com/design/patterns/notifications.html#guidelines">Android notification UX guidelines</a></li><li>[Video] <a href="https://www.youtube.com/watch?v=srdKq0DckXQ">Progressive Web Apps across all frameworks — Google I/O 2016</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dd38c1d20f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>