mobile internet

image_print

Better Than The Apple MVNO Fantasy

by Jean-Louis Gassée

An Apple cellular network is a nice but unrealistic fantasy. Today we explore better ways of achieving the carrier-independence we dream of. More

What The Ad Blocker Debate Reveals

378_hdr

by Jean-Louis Gassée

iOS 9 and OS X 10.11 (“El Capitan”) carry ad-blocking technology that delivers an experience that stands in stark contrast to current advertising and tracking practices. Users are beginning to notice…and advertisers aren’t happy about it.

More

News Sites Are Fatter and Slower Than Ever

375_cover

by Frédéric Filloux

An analysis of download times highlights how poorly designed news sites are. That’s more evidence of poor implementation of ads… and a strong case for ad blockers. More

Samsung’s Mobile OS Dilemma

372_samsung

by Jean-Louis Gassée

Scroogled if you do, HPed if you don’t. To differentiate itself from aggressive Android competitors, Samsung would need to build its own mobile OS…but can it overcome the odds? More

The redistribution game for news

371_redist_graph
by Frédéric Filloux
Forget the 70-30 split for subscription between publishers and distributors. Today, for publishers, the new norm is a 100%-70% split of ad revenues, depending on who sells the ad. For news distribution, re-intermediation will be intensely competitive.

The chart above illustrates the upcoming shift in news distribution. No doubt: We’re heading towards a new phase of massive re-intermediation, of reshuffling the layers between the news producers (traditional media houses or pure players) and readers. This raises important questions: What will publishers gain or lose in the process? Will they end up handcuffed to a cluster of gatekeepers or will they reap decisive gains for their business model.

Who becomes the dominant player in this new structure? More

More facts on mobile 

wireframe-mobile-blueprint

by Frederic Filloux

New mobile internet trends have caught my attention this week. Today, we look at their impact on the news business. (1)

In developing countries, mobile growth keeps accelerating.

  • In Nigeria (pop: 173 million, median age: 18), based on page views count, 76% of internet traffic now comes from mobile.
  • In India (pop: 1.25 billion), the rate is 65%. As a comparison, the proportion is 23% in the UK, 22% in the United States and 14% in France.

At least, four elements drive mobile growth in emerging countries: More

Circa: What went wrong

by Frederic Filloux

365_ciracBW

Circa, the clever smartphone news app, failed to live up to its promises. The fiasco stems from the smartphone advertising market’s inherent weakness, from Circa’s inability to catch up with evolving reading habits, and from an insufficient editorial proposition.  More

Featured On Tim Cook’s Keynote – What It Takes

by Frederic Filloux

At last October’s introduction of the new iPad Air, the creators of a clever iOS app named Replay were invited on stage. To get there, they went through a selection process that illustrates Apple’s perfectionism — and hidden application sophistication.

In September 2014, while at the Stupeflix Paris office, Nicolas Steegmann got a call from Apple in Cupertino. Once the caller identified herself, Nicolas knew something up. The contact came after Stupeflix presentations to Apple’s team in Paris. In rather elliptic terms, Steegman’s interlocutor said it would be great if two members of the company, a developer and a designer, could be in Cupertino the next day. ‘They will have to stay at least two weeks’, she said. 48 hours later, the team was on Apple’s campus. They quickly found themselves in a windowless room and given a straightforward brief: Devise the coolest possible demo for your app. No more details, no promises whatsoever.

364_main364

[Stupeflix founder Nicolas Steegmans (right) and motion engineer Jean Patry (left)] 

Replay is a clever iOS application that focuses on a “simple” issue: Automating the process of making of videos, without going through the convoluted steps of a dedicated movie editing app. With Replay, you shoot with your iPhone (or your iPad), select the clips you want, pick one of the proposed theme and you’re done. The app will assemble the clips in the smartest possible way, making visual corrections, adjusting the soundtrack selected from your iTunes library (or drawn from a proposed catalogue) to the pace of the movie. If you have the time and inclination, additional settings let you fine-tune your production. But even if you just stay with basic preset themes, the result is stunning. In literally a few seconds, you end up with a clip perfectly suited to quick sharing on YouTube, Instagram or Facebook.

Behind Replay’s simplicity are years of work and a great deal of sophisticated programming. The company’s roots are in an automated video generation system originally designed for completely different goals.

As often, a company’s final product has little to do with the original intent.

Stupeflix is a pure engineers’ startup. It was created in 2008 by Nicolas Steegmann, an engineering and mathematics graduate from Ecole Centrale de Paris, and Francois Lagunas who holds a PhD in computer sciences and linguistics from Polytechnique and Ecole des Mines. Their first product was an automated video generator that scrapped images and text from Wikipedia and other sources, inserted text-to-speech voice-over, to create 45 sec. glances at various cities and places around the world. The result was more a demonstrator than a commercial product (you can still access hundreds of automatically generated videos here on YouTube.)

The concept paved the way for a much more bankable product: a system to create videos entirely online, with presets themes — the ancestor to the Replay app. Its business model was (and still is) based on the proven freemium mechanism: Casual use is free, scaling to a professional/intensive use requires a subscription.

The same went for the next iteration: a home-brewed API allowing third parties to use all the tools Stupeflix developed to create videos. As a result, digital advertising agencies such as Publicis, Saatchi and TBWA jumped on it. Hundreds of thousands of videos were created for Coca-Cola, Red Bull or Sprint, to be used in countless promotional operations. Stupeflix still derives significant revenue from its API business.

364_bandeau_apis

Technically speaking, editing and rendering a video is CPU intensive — GPU intensive to be more precise; Stupeflix’s APIs suck a lot of graphic processing power. At the time, explains Nicolas Steegmann, graphics rendering was outsourced to a specialized server farm in Texas (where the oil industry consumes loads of computational power for geophysics modeling). Now, Stupeflix relies on Amazon Web Services, which has since cornered the CPU/GPU for-hire market.

It took 18 months to port the video rendering engine to iOS. Many invisible features had to be pared down to fit the power of the iPad/iPhone processor. Unbeknownst to the user, Replay performs many complex graphics tasks. For instance, it analyses each piece of raw media material picked by the user. Color palette and saturation, exposure, motion, pace are decomposed and translated into mathematically useable components. These chunks of data are then fed into a “cinematographic grammar” hard-coded by Replay’s programmers (all movie enthusiasts). Each theme or skin selected by the user reflects a Quentin Tarantino or Alfred Hitchcock inspiration that will direct transitions, colorimetry, beat, as well as soundtrack sync. And an embedded machine learning engine also devises new rules by itself.

The fluidity of Replay’s performance caught Apple’ attention during the summer of 2014, a couple of months before Tim Cook’s unveiling of the new iPad Air.

Now secluded in their room of the Apple campus, always escorted when they had to walk in and out, Stupeflix’s team is hard at work devising the most mind-blowing demonstration of their app. Early on, they had a hunch that the whole process was in fact a competition among applications (12 contenders, as they would later discover.) For two weeks, a quiet selection process took place, with a stream of people visiting the team, now allowed to test its work on the last version of a new iPad camouflaged in a thick neoprene enclosure to conceal its size and shape. Each successive visit was made by someone ranking higher and higher in the chain of command — as the team realized after Googling the reviewers. They knew they were on the short list when their demo was shown to Phil Schiller, Apple SVP for Worldwide Marketing. The next day, the pair was taken to a conference room where their work was reviewed by Tim Cook in person. They knew it was a go. It was time for a series of full rehearsals.

On D-day, the two-minute presentation was to be made by Jeff Boudier, the Stupeflix man in San Fransisco (and co-founder of the company), assisted by François Lagunas controlling the iPad. It went well, except when a slip of a finger (due to an excess of makeup applied to the demonstrator’s hand) caused the auto-correct to transform the title “Utah Road Trip” into a weirder “It’s a road trip”… After the show, Apple staff asked to re-record the demo for a spotless posterity (the re-edited version visible here on Apple’s site, while the original is here; time code about 00:55:10 on the two keynotes). This says much about Apple’s attention to details.

An epilogue: Replay became a hit, generating substantial revenue thanks to the in-app purchase system. Stupeflix now employs 23 people, all of the same caliber as the founders. To stimulate the team’s creativity, management keeps holding internal hackathons, and they continue to build on the uniqueness of their video algorithms and rendering engines. The company recently came up with Steady, a spectacular app that gives the impression your iPhone is mounted on a Steadicam (a complex system crops each frame in real-time to compensate for unwanted motion) and Legend, that animates texts on the fly. They are now working on another movie capture app that will further transfer the burden of filmmaking from the user to the software. Call it talent by proxy.

frederic.filloux@mondaynote.com

Beware of Airbnb entering the hyperlocal travel guide business

 

by Frédéric Filloux 

Airbnb has every reason to enter the news services sector – and to threaten a broad range of media/services such as Trip Advisor or Yelp. 

Seen through the eyes of travel information publishers, Airbnb holds a dream position: a huge base of 25 million potential readers/users, spread over 34,000 cities in 190 countries, well in tune with the brand’s core product and attributes.

359_heatmap

For a start, should Airbnb develop a publishing arm, this unparalleled notoriety would spare it tens of millions dollars otherwise required to promote its content services: In the travel industry, advertising and marketing demand high spending. To put things in perspective, this year, according to the trade site Skift.com, HomeAway, one of Airbnb’s key competitors, plans to spend $100m (after shelling out $60m in 2014) “To show it’s not Airbnb“. HomeAway was created in 2005 and received $504m in funding before going public in 2011 — the stock lost 18% since. Airbnb has yet to go public after raising more than $800m. Its latest confirmed round closed in October was for $50M with a $13bn valuation. Airbnb is now rumored to be raising a $1bn “war chest” at an even higher price: $20bn.

Today, Airbnb’s expansion is a matter of concern not only for its direct competitors but also for large players in the travel information segment.

Last November, Airbnb fired the first shot: Pineapple, a 128 page, ad-free, glossy quarterly magazine, mostly written by hosts and guests gently discussing their experiences. With a tiny print run of 20,000 copies, it was distributed for free in some Airbnb hotspots and sold at selected bookstores such as WH Smith in Paris, where I got my copy for €12. According to Pineapple publisher Christopher Lunekzic (interview on the FIPP site), “The idea came from a desire to capture the unique nature of traveling through Airbnb. We wanted to bring that sense of creativity, culture and connection to life”. All the talking points of an elegant PR campaign are checked, part of Airbnb recent rebranding. Nothing to freak out travel press behemoths.

Except… Now, a sizable part of Airbnb community is aware of the company’s recent push into the magazine business. Which could make a serious difference.

That said, print is certainly not a key part of Airbnb’s media future. Mobile apps are. Last year, only 20% of Airbnb traffic came from mobile. That was before last December’s app redesign. Today, the Airbnb ranks in the top 5 apps in 7 countries, and in the top 100 in… 152 countries. That’s where the real potential is. And the San Francisco-based icon of the sharing economy is betting heavily on mobile: it recently announced a partnership with Deutsche Telekom’s T-Mobile to pre-install its app on Android smartphone across 13 markets in Europe (TechCrunch story here). A decisive move for Airbnb’s mobile expansion.

Now let’s indulge in a little bit of fiction — from the perspective of an Airbnb guest.

I’m booking a flat in, say in London’s Marylebone district. I’m not familiar with the best places to go out. In the dedicated section of the new Airbnb app, I enter the host’s postal code, which is precise down to the building (likewise, the American “Zip+4″ code provides a city block location; in other cities, the Lat-Long associated to a street address does the job). My host has listed her preferred spots: organic groceries, wine bars, galleries, shopping places, movie theaters… Every place shows up on a Google map. The practical details I might need appear over my host’s comments: business hours, booking information, etc. I can also check the reviews on third party sites, but given my host’s profile, I assume our tastes will match; plus, I don’t want to get drowned into a tedious (and too often dubious) series of five-stars searches.

On my phone, I now enjoy a mini-guide of the neighborhood where I’m going to spend the next few days, filled with trusted, non-commercially-induced recommendations. Right from the app, I can make reservation via Open Table, and even call a Uber car. That’s what API’s are for: connecting applications together, in a mutually beneficial way. The third-party service provider expands its reach and the app publisher offers a wider range of services while keeping the customer “inside”. Similarly, a gallery or museum can make its program available within the app.
APIs will be a major development engine for the apps ecosystem as the number of features and services that can be added is boundless. For example, Uber has recently made its API system much more accessible and now partners with 11 companies, including Hyatt, OpenTable, Starbucks, Time Out, TripAdvisor, TripCase — curiously, not Airbnb nor Yelp.

Why would such integration threaten large travel business publishers?

Beyond developing its gigantic global footprint, Airbnb wants to build a community of users, itself structured in homogenous layers (e.g. young families looking for budget rentals, yuppies aiming at trendy places…) There’s even the growing crowd of professionals who prefer an Airbnb apartment free of the check-in/out hassles of hotels, and who will gladly trade unexciting room service for a super-fast DSL connection. (I’m told a growing number of Googlers do so for their business trips, with their employer’s blessing.) Each of these sub-communities will be far more likely to trust their peers than the usual travel guides where it’s always difficult to sort actual user opinions from bogus reviews and paid-for insertions, disguised advertorials, etc.

The beauty of this powerful combination lies in its scalability. Airbnb listings contain a broad range of properties, including high-end, luxury items. Those who might be willing to cough up €2,000 a night for a two-bedroom unit with a stunning view of the Hong Kong harbor might also want a special cicerone, a much more sophisticated one than the peer-to-peer guide described above.

Sometime ago, Louis Vuitton — the main brand of LVMH luxury conglomerate — had the idea of creating a dedicated app aimed at its rich Chinese clientèle, at those able to spend €20,000 or more in a single afternoon shopping stroll through Avenue Montaigne in Paris. This special application was to offer the services of a personal shopper, but also a personal city guide (Louis Vuitton already publishes its own collection of high-end guides) to be used from planning the trip to the actual journey. Even better, the app was to rely on the Chinese-made WeChat application (500 millions users, roughly 85% from mainland China) connected to an actual human able to guide the wealthy tourist on a real-time basis, either with messages or voice contact. In such case, relying on peer recommendations made no sense, but a highly personalized service did. A couple of selected partners were in the loop.

Regardless of the market segment, a notorious brand coupled to a set of mobile services is a potent combination. In the case of Airbnb, the “full stack” company — a concept coined by Andreesen Horowitz’ Chris Dixon —  is likely to be a master tool for securing the position of a brand in its market.

frederic.filloux@mondaynote.com

News Media Should Drop Native Apps

 

by Frédéric Filloux

When it comes to the most basic form of news delivery, facts keep piling up in a way that makes native apps more and more questionable. Here is why it’s worth considering a move back to mobile sites or web apps…  

One of the most shared statistic on mobile use is this one: Applications account for 86% of the time spend b’y users. This leaves a mere 14% for browser-based activities, i.e. sites designed for mobile, either especially coded for nomad consumption, built using responsive design techniques that adapt look and feel to screen size, or special WebApp designs such as FT.com.

This 86/14 split is completely misleading for two reasons: the weight of mobile gaming, and the importance of Facebook.

Take a look at this chart form Flurry Analytics:

358_time_spent

 

If you combine gaming, Facebook and Social Messaging, roughly 60% of time spent on mobile is swallowed by this trio. As for Facebook, it reaped the top four slots in downloads for non-gaming apps worldwide:

358_2

Mobile consumption will concentrate even more as messaging and free direct communication (red dot in the chart) combine into the fastest growing segment by far. Mobile carriers have reason to be scared. Consequently, Facebook will tighten its grip on the mobile ecosystem as it commands the two dominant direct communication apps. If this wasn’t enough, there are two other fast-growing usages: Video streaming (+44% downloads last year) and Travel & Transportation — Uber, AirBnB, CityMapper — (+31%.)

The main characteristic of these services is they couldn’t be designed outside of a full-fledged application: They require key phone features not easily accessible through a browser such as radio modules, GPS,  image rendering (for maps, graphics), camera, etc.

By comparison, news-related applications do not requires a lot of phone resources. They collect XML feeds, some low resolution images and render those in pre-defined, non-dynamic templates. They use a tiny fraction of a modern smartphone’s processing power.

In fact, for news media, as the following matrix shows, native apps (iOS, Android and soon Windows) become a questionable proposition:

358_Matrix_3b
Summing up, apps for news distribution are technically justified for speed, ability to send notifications, inApp purchases, and the hypothetical use of phone sensors. That’s much money and a lot of complications for a small number of features. (On this subject, see the previous The Future of Mobile Apps for News column.)

Unless you are fighting for the prime phone screens, say the first three swipes, or if you are determined to provide key visual or functional differentiators, going for a set of native apps must be carefully weighed. Depending on the level of sophistication and required features, developing a native application costs between $50,000 and $100,000, for each environment, plus dedicated SDKs for marketing, analytics, etc., plus a 30% fee paid to the app store if you go for a paid or subscription model, plus hassles for approval of the smallest update in the messy iOS App Store…

But if you are already big on social and SEO, a mobile site, lightweight, clearly focused on a small feature set can be quite effective. Disappointing as it may be, HTML5-based web apps are all but dead (the difficulty lies in finding good developers and in managing them.)

Among all players, Google has the ability to overhaul the mobile ecosystem. If it comes up with an SDK or a framework aimed at creating simple and effective apps, indeed with limited performance but with enough features to accommodate news delivery, this could become an industry game-changer.

frederic.filloux@mondaynote.com

358_banner_Prez_d