What is it like to work at Netflix as a developer? How do they think about culture, customers and engineering productivity?
In this incredible episode of Dev Interrupted, I bring in Kathryn Koehler, the Director of Productivity Engineering at Netflix, to chat about what makes Netflix so unique and why they are standardizing data-driven engineering today.
Advertising knows you better than your friends, better than your family, perhaps even better than your partner.
Look up pizza recipes, and advertising will show you promotions for pizza ovens. Download a marathon training plan, and advertising will show you the latest running shoes. Buy a car, and advertising will show you adverts for other cars because no system’s perfect.
Advertising does this with a simple trick: it watches you constantly. It’s watching you right now. The web is one giant machine for making money, and you’re the fuel.
On the one hand, advertising’s insidious invasion of our privacy is enough to make you paranoid; on the other hand, I really love my pizza oven.
The largest facilitator of advertising on the web is Google Ads — reportedly worth $134.8 billion per annum; it’s Alphabet Inc’s primary source of revenue.
Last year, Google Ads announced that it would be ending its reliance on third-party cookies for delivering targeted advertising as part of a wider industry trend towards greater privacy protection for individuals. This week, we received more details confirming that Google Ads will not replace third-party cookies with comparable tracking technology.
Google Ads intends to maintain relevant advertising, without user tracking, by anonymizing your identity within a crowd. The technical term is a Federated Learning of Cohorts (FLoC), essentially Asimov’s Psychohistory, in capitalist form, some 45,000 years before Hari Seldon is due to be born.
In simplistic terms, someone who buys a pair of running shoes can reasonably be expected to be interested in GPS watches. The complexity arises when grouping becomes more complex: people who watch Netflix on a Tuesday evening purchase a particular soup brand and read the Washington Post, for example. The system requires billions of groupings that are too complex to express in English. And yet Google claims to already be making some progress.
As with any fledgling technology, the implications of its widespread adoption are unclear. FLoC is Chrome-based, so there’s the looming specter of a monopoly. Then there’s the issue of how groups are built; does Google need individual tracking to generate crowds of individuals? It’s unclear, but what is clear is that if Google succeeds — and it’s likely that it will — other networks will have no choice but to follow suit. It seems inevitable that there will be a wide-ranging impact across not just advertising but analytics and marketing as a whole.
The back door that’s being held open is one-to-one relationships. If you visit a site, that site can attempt to entice you back with targeted advertising. This means the next few years will see a growth in the number of companies developing ongoing relationships in the form of newsletters and memberships.
How ever it plays out, a fundamental change to the system that funds most of the web is certain to have a long-term impact on day-to-day user experience.
Featured image via Pexels.
The post Google’s FLoC Promises a Radically Different Web first appeared on Webdesigner Depot.
L’intelligence artificielle (IA) s’impose comme le nouvel intermédiaire entre les entreprises et les clients dans la décision d’achat.
La complexité de la vie moderne soumet notre cerveau à une vraie surcharge cognitive. Pour y remédier, les clients se tournent vers les IA et se fient à leurs recommandations, délaissant les interactions traditionnelles avec les marques.
Ainsi acceptons-nous volontiers les recommandations de Netflix pour les séries, de Spotify pour la musique ou les résultats affichés par Google. Nous faisons tellement confiance à ces suggestions que 35% des achats sur Amazon et 75% des visionnages sur Netflix sont issus des recommandations du Machine Learning, selon McKinsey (1).
Pourquoi ? Parce que, dans l’expérience, les clients distinguent de plus en plus les aspects routiniers des aspects riches de sens (voir « Donner du sens à l’expérience »). Et à mesure que la précision et l’utilité des IA augmentent, les clients délèguent les décisions qu’ils jugent « sans importance ». Certaines entreprises se voient alors coupées de leurs clients et forcées de composer avec des IA agissant comme des gardiens.
Les IA étant capables de regrouper, d’indexer et d’analyser des montagnes de données, les clients ont de plus en plus confiance dans leur capacité à interpréter leurs préférences personnelles et à faire des choix à leur place. Aux recherches en arrière-plan réalisées par les IA s’ajouteront bientôt les achats en arrière-plan.
Le marketing émotionnel peut susciter la fidélité chez les humains mais pas chez les IA. À mesure que les clients confieront les décisions d’achat de routine aux IA, les entreprises auront besoin de nouvelles stratégies pour conquérir ces nouvelles audiences automatisées.
On regroupe souvent les tâches ménagères et administratives en catégories, comme le nettoyage de la maison ou le paiement des factures. Mais avec des IA qui prennent en charge différentes tâches à la fois, ces frontières disparaîtront. Les entreprises incapables de s’adapter à ce changement dans la façon d’appréhender les tâches perdront des clients.
Créez une intelligence artificielle spécialisée qui apporte des connaissances uniques à vos clients. Appropriez-vous les fonctionnalités des IA pour mettre ces connaissances et compétences spécialisées à disposition des clients, directement et par le biais d’IA généralistes.
Au fur et à mesure que les points de contact avec les clients seront relégués au second plan, les produits joueront un rôle plus central en tant qu’ambassadeurs de la marque. Le Big Data, l’IoT et le Machine Learning permettent d’améliorer l’expérience client en augmentant la satisfaction, en réduisant la charge cognitive et en éliminant les actions inutiles. Trouvez des moyens de créer un lien avec vos clients par le biais des produits.
Les systèmes basés sur l’IA constitueront le prochain « espace linéaire » numérique où optimiser le placement des produits. Cette discipline d’optimisation est encore balbutiante mais vous pouvez déjà vous appuyer sur votre équipe de référenceurs SEO afin de prendre le leadership dès le début. Des descriptions factuelles de vos produits et services contribueront à votre succès auprès des IA.
Utilisez des outils basés sur l’IA pour animer les interactions avec vos clients en fournissant aux collaborateurs de 1ère ligne des informations client et des suggestions de messages. Par exemple, les outils d’analyse et le Machine Learning peuvent servir à offrir aux collaborateurs une visibilité approfondie sur chaque client, améliorant ainsi la personnalisation des interactions.
Publié en anglais sur insights.sap.com
Références
(1) Ian MacKenzie, Chris Meyer et Steve Noble, « How Retailers Can Keep Up with Consumers » (Comment les détaillants peuvent répondre aux besoins des clients), McKinsey & Co., octobre 2013, https://www.mckinsey.com/industries/retail/our-insights
/how-retailers-can-keep-up-with-consumers.
The post L’expérience client future : l’IA comme intermédiaire appeared first on SAP France News.
We are gathered here today….
Today I write in memory of Adobe Flash (née Macromedia), something that a bunch of people are actually too young to remember. I write this with love, longing, and a palpable sense of relief that it’s all over. I have come to praise Flash, to curse it, and finally to bury it.
We’ve been hearing about the death of Flash for a long time. We know it’s coming. December 2020 has been announced as the official timeframe for removal, but let’s be real about this: it’s dead. It’s super-dead. It’s people-are-selling-Flash-game-archives-on-Steam dead.
That last bit actually makes me happy, because Flash games were a huge part of my childhood, and the archives must be preserved. Before I’d ever heard of video cards, frames per second, and “git gud”, I was whiling away many an hour on disney.com, cartoonnetwork.com, MiniClip, Kongregate, and other sites, looking for games.
I think we’ve established in my previous work that even as a missionary kid, I did not have a social life.
The Internet itself gave me a way to reach out and see beyond my house, my city, and my world, and it was wonderful. Flash was a part of that era when the Internet felt new, fresh, and loaded with potential. Flash never sent anyone abuse, or death threats. Flash was for silly animations, and games that my parent’s computer could just barely handle, after half an hour of downloading.
I even built my first animated navigation menus in Flash, because I didn’t know any better. At all. But those menus looked exactly like the ones I’d designed in Photoshop, so that’s what mattered to me, young as I was.
That was a part of Flash’s charm, really.
Funny story, JavaScript was only about a year old when Flash was released. While HTML5 and JS are the de-facto technologies for getting things done now, Flash was, for many, the better option at launch. JS had inconsistent support across browsers, and didn’t come with a handy application that would let you draw and animate whatever you wanted.
It was (in part) Flash that opened up a world of online business possibilities, that made people realize the Internet had potential rivalling that of television. It brought a wave of financial and social investment that wouldn’t be seen again until the advent of mainstream social networks like MySpace.
The Internet was already big business, but Flash design became an industry unto itself.
Yeah, Flash websites could be reliably responsive (and still fancy!) before purely HTML-based sites pulled it off. Of course, it was called by other names back then, names like “Liquid Design”, or “Flex Design”. But you could reliably build a website in Flash, and you knew it would look good on everything from 800×600 monitors, to the devastatingly huge 1024×768 screens.
You know, before those darned kids with their “wide screens” took over. Even then, Flash still looked good, even if a bunch of people suddenly had to stop making their sites with a square-ish aspect ratio.
On top of being pseudo-responsive, the plugin-based Flash player was almost guaranteed to work the same in every major browser. Back in a time when Netscape and Internet Explorer didn’t have anything that remotely resembled feature parity, the ability to guarantee a consistent website experience was to be treasured. When FireFox and Chrome came out, with IE lagging further behind, that didn’t change.
While the CSS Working Group and others fought long and hard for the web to become something usable, Flash skated by on its sheer convenience. If your site was built in Flash, you didn’t have to care which browsers supported the <marquee> tag, or whatever other ill-conceived gimmick was new and trendy.
Remember when YouTube had a Flash-based video player? Long before YouTube, pretty much every site with video was using Flash to play videos online. It started with some sites I probably shouldn’t mention around the kids, and then everyone was doing it.
Some of my fondest memories are of watching cartoon clips as a teenager. I’d never gotten to watch Gargoyles or Batman: The Animated Series as a young kid, those experience came via the Internet, and yes… Flash. Flash video players brought me Avatar: The Last Airbender, which never ever had a live action adaptation.
Anyway, my point: Flash made online video streaming happen. If you’ve ever loved a Netflix or Prime original show (bring back The Tick!), you can thank Macromedia.
Obviously, not everything was rosy and golden. If it was, we’d have never moved on to bigger, better things. Flash had problems that ultimately killed it, giving me the chance, nay, the responsibility of eulogizing one of the Internet’s most important formative technologies.
Firstly, it was buggy and insecure: This is not necessarily a deal-breaker in the tech world, and Microsoft is doing just fine, thank you. Still, as Flash matured and the code-base expanded, the bugs became more pronounced. The fact that it was prone to myriad security issues made it a hard sell to any company that wanted to make money.
Which is, you know, all of them.
Secondly, it was SEO-unfriendly: Here was a more serious problem, sales-wise. While we’re mostly past the era when everyone and their dog was running a shady SEO company, search engines are still the lifeblood of most online businesses. Having a site that Google can’t index is just a no-go. By the time Google had managed to index SWF files, it was already too late.
Thirdly, its performance steadily got worse: With an expanding set of features and code, the Flash plugin just took more and more resources to run. Pair it with Chrome during that browser’s worst RAM-devouring days, and you have a problem.
Then, while desktops were getting more and more powerful just (I assume) to keep up with Flash, Apple went and introduced the iPhone. Flash. Sucked. On. Mobile. Even the vendors that went out of their way to include a Flash implementation on their smartphones almost never did it well.
It was so much of a hassle that when Apple officially dropped Flash support, the entire world said, “Okay, yeah, that’s fair.”
Side note: Flash always sucked on Linux. I’m just saying.
Flash was, for its time, a good thing for the Internet as a whole. We’ve outgrown it now, but it would be reckless of us to ignore the good things it brought to the world. Like the creativity of a million amateur animators, and especially that one cartoon called “End of Ze World”.
Goodbye Flash, you sucked. And you were great. Rest in peace. Rest in pieces. Good riddance. I’ll miss you.
Featured image via Fabio Ballasina and Daniel Korpai.
L’ancien président des Etats-Unis et son épouse ont signé un contrat avec Netflix qui prévoit la création de contenu audiovisuel, comme des reportages et des docu-séries.
Source de l’article sur ZDNet
Le service de SVOD va consacrer l’essentiel de sa puissance financière dans la création de films et séries originaux.
Source de l’article sur ZDNet
Nouveau sur ZDNet.fr : avec Décideurs TV, voici le 4e numéro de notre debrief de l’actualité IT des dernières semaines. Avec la complicité de Stéphane Larcher du magazine L’Informaticien.
Source de l’article sur ZDNet
