About Frédéric Filloux

Posts by Frédéric Filloux:

The not-so-quaint charm of the email newsletter

 

In spite of today’s obsession with social networks, the email newsletter remains a potent vector for the dissemination of news and for driving traffic back to websites. It comes with one condition, though: reintroducing a human touch. 

Today, producing a newsletter looks so easy: Select RSS feeds from your site, fire a plug-in to extract selected headlines and areas, insert the feeds in a template and send the whole thing via a router interface. Done.

Many sites do it on auto-pilot. And the result of such automated treatment is crude newsletters throwing together a bunch of headlines and snippets. On the surface, the output does reflect the content of a site, but it actually fails to reveal any editorial choice other than the basic home page hierarchy. An opinion piece, an in-depth profile, or an investigative report will be processed in the same mechanical way: headline, nutgraf, a couple of links and nothing further.

Based on my personal use, such work ends up in a special designated folder I created on my main Gmail account for each publication I subscribe to. After a while, I stopped looking at those robotized emails. To make things worse (for the senders), Google does the filing for me — unbeknownst to me, actually. A couple of months ago, Gmail created several tabs, one of them titled “Promotions”, that collect all newsletters, including the ones I willingly subscribed to. Google chooses for me the emails should I read first. Great. I don’t understand why this arbitrary filtering didn’t trigger any outcry, both from subscribers and publishers of legit newsletters (I happen to be both). Needless to say, the opening rate of emails falling into the infamous Promotions folder is significantly altered. All at the pleasure of Google and its algorithms.

Coming back to the newsletter itself, we can detect the beginning of a shift away from robotized email towards the written-by-humans form.

Again, I’ll refer to Quartz, the business site launched a year ago by the Atlantic Media Group (see a previous Monday Note series here). Their email newsletter is called “The Daily Brief”; it is 800-words long, no images, cleverly written and edited, sent to about 45,000 subscribers worldwide, in three editions (US, Asia, Europe and Africa.)

Here is how it looks on mobile devices:

qz phones

The structure is simple: Five main headers containing five to seven items, each summing up what the story you might click on is about. The headers are: “What to watch today”, “While you were sleeping”, “Quartz obsession interlude” (it refers to Quartz’ proprietary revision of the old beat structure), “Matter of Debate”, and “Surprising discoveries”. A good mixture of news, fun, serendipity, thoughtful items. The links do not always send back to qz.com, they can lead anywhere. Sounds pretty simple at first. But, as Quartz editor Kevin Delaney recently told me, the Daily Brief is the result of a thorough editorial process. The email newsletter is touched by no less than four people, including two seasoned editors, Gideon Lichfield, Quartz global news editor who spent 16 years at the Economist, and Adam Pasick, the Asia editor and a 10-year Reuters veteran. Newsrooms who assign junior writers to expedite email newsletters should think again… Quartz is one of the few media I know to actually devote sizable resources for such a “simple” news product (also read this analysis on MailChimp, Quartz email router). But many are now considering the formula: The Wall Street Journal recently launched its “10-Points” email newsletter, built on the same principles as Quartz’s Daily Brief.

wsj-10points2

Sophisticated email newsletters are not new. For years, bloggers affiliated or not with large media organizations have been using them to promote their work and attract readers, gaining significant traction in the process. To name but a few, Andrew Sullivan’s Daily Dish on Politics, or Andrew Ross Sorkin’s Dealbook (part of NYTimes.com) have become full-fledged news brands. I asked Juan Señor, partner at Innovation-Consulting, who worked on many newspaper modernizations, for his opinion on the matter:

Conceptually, our take and that of other newspapers investing in newsletters or news briefings – as we call them – is that you have to move from commodity news to selling intelligence. In an age of abundance you have to sell scarcity. The laws of economics prescribe that the more abundant a product is, the less valuable it is in price. The more volume I have, the less value I can extract from it.’ 

Juan adds two critical factors needed to create a valued product: Timing — sending a news briefing at the right time to maximize its impact — and the multi-device format.

In spite of their age, email newsletters remain a relative primitive stage. Let’s talk first about the user interface. A newsletter begs to be read both on mobiles and on a desktop. You can no longer decide for the reader which screen size h/she will read your stuff on. Responsive design is mandatory. But applying responsive design techniques is way more complicated for newsletters than it is for websites. Even large medias such as the NYT are providing single formats newsletters. (I will humbly admit that, while the Monday Note blog switched to responsive design a while ago, I’m still struggling to do the same for our newsletter.) While I want to send a newsletter from a series of blog posts in a single stroke, I’m still waiting for the WordPress plug-in that will let me do that through a wide range of email routers. In the same fashion, I would welcome add-ons to the most popular word processors that would output good-looking, responsive html emails.

Another thing about email design: It must be conceived to be read offline. I live in a 4G city (Paris) but I still get poor 3G or even EDGE service in too many places (French carriers are said to slow down network speed in order to accelerate the switch to 4G). Therefore, the ability to read complete content offline beyond headlines is, in my view, a basic feature. Going a bit further, I would dream of newsletters pre-loading multiple layers of reading, allowing the reader to jump from the main page to one or two levels down — without requiring a connection.

Deeper improvements to newsletters will come from the usual combination of analytics and semantics. A well-crafted engine will detect what parts of an email newsletter I read the most, what subjects I’m more inclined to click on. Then, the system will adapt the content of my newsletters in order to increase my propensity to open and to engage (i.e. to click on links.) This will make the old-fashioned newsletter an even more powerful website traffic vector.

frederic.filloux@mondaynote.com

 

News: Mobile Trends to Keep In Mind

 

For publishers, developing an all-out mobile strategy has become both more necessary and more challenging. Today, we look at key data points and trends for such a task. 

#1 The Global Picture
— 1.7bn mobile phones (feature phones and smartphones) were sold in 2012 alone
— 3.2bn people use a mobile phone worldwide
— Smartphones gain quickly as phones are replaced every 18 to 24 months
— PCs are completely left in the dust as shown in this slide from Benedict Evans’ excellent Mobile is Eating the World presentation:

ben-evans

The yellow line has two main components:
— 1 billion Android smartphones are said to be in operation worldwide (source: Google)
— 700 million iOS devices have been sold over time, with 500 million still in use, which corresponds to the number of iTunes accounts (source: Asymco, one of the best references for the mobile market.)
— 450 million Symbian-based feature phones are in operation (Asymco.)

#2 The Social Picture 

Mobile phone usage for news consumption gets increasingly tied to social networks. Here are some key numbers :
— Facebook: about 1.19bn users; we don’t exactly know how many are active
— Twitter: 232 million users
— LinkedIn: 259 million users

When it comes to news consumption in a social environment, these three channels have different contributions. This chart, drawn from a Pew Research report, shows the penetration of different social networks and the proportion of the US population who get their news from it.

300_pew

One of the most notable data points in the Pew Report is the concentration of sources for social news:
— 65% say to get their news from one social site
— 26% from two sites
— 9% from three sources or more (such as Google +, LinkedIn)

But, as the same time, these sources are completely intertwined. Again, based on the Pew survey, Twitter appears to be the best distributor of news.

Among those who get their news from Twitter:
— 71% also get their news on Facebook
— 27% on YouTube
— 14% on Google+
— 7% on LinkedIn

Put another way, Facebook collects more than half of the adult population’s news consumption on social networks.

But a closer looks at demographics slightly alters the picture because all social networks are not equal when it comes to education and income segmentation:

If you want to reach the Bachelor+ segment, you will get:
— 64% of them on LinkedIn
— 40% on Twitter
but…
— only 30% on Facebook
— 26% on G+
— 23% on YouTube

And if you target the highest income segment (more than $75K per year), you will again favor LinkedIn that collects 63% of news consumers in this slice, more than Facebook (41%)

Coming back to the mobile strategy issue, despite Facebook’s huge adoption, Twitter appears to be the best bet for news content. According to another Pew survey, the Twitter user is more mobile :

Mobile devices are a key point of access for these Twitter news consumers. The vast majority, 85%, get news (of any kind) at least sometimes on mobile devices. That outpaces Facebook news consumers by 20 percentage points; 64% of Facebook news consumers use mobile devices for news. The same is true of 40% of all U.S. adults overall. Twitter news consumers stand out for being younger and more educated than both the population overall and Facebook news consumers

 And, as we saw earlier, Twitter redistributes extremely well on other social platforms. It’s a no brainer: any mobile site or app should carry a set of hashtags, whether it’s a stream of information produced by the brand or prominent bylines known for their insights.

 #3 The Time Spent Picture

Here is why news is so complicated to handle in mobile environments. According to Flurry Analytics: On the 2 hours and 38 minutes spent each day on a smartphone and an a tablet by an American user, news accounts for 2% as measured in app consumption, which accounts for 80% of time spent. The remaining 20% is spent in a browser where we can assume the share of the news to be much higher. But even in the most optimistic hypothesis, news consumption on a mobile device amounts to around 5 to 6% of time spent (this is correlated by other sources such as Nielsen). Note that this proportion seems to decrease as, in May 2011, Flurry Analytics stated news in the apps ecosystems accounted for 9% of time spent.

This view is actually consistent with broader pictures of digital news consumption, such as these two provided by Nielsen, which show that while users spend 50 minutes per month on CNN (thanks to is broad appeal and to its video content), they only spend 18 minutes on the NYT and a mere 8 minutes on the Washington Post:

300 nielsen

All of the above compares to 6hrs 42min spent on Facebook, 2hrs on YouTube or Yahoo sites.

In actionable terms, this shows the importance of having smartphones apps (or mobile web sites) sharply aimed at providing news in the most compact and digestible way. The “need to know” focus is therefore essential in mobile because catching eyeballs and attention has become increasingly challenging. That’s why The New York Times is expected to launch a compact version of its mobile app (currently dubbed N2K, Need to Know, precisely), aimed at the market’s youngest segment and most likely priced just below $10 a month. (The Times also does it because the growth of digital subscriptions aimed at the upper market is slowing down.) At the other end of the spectrum, the NYT is also said to work on digital magazine for iPad, featuring rich multimedia-narrative on (very) long form such the Pulitzer winning Snow Fall (on that matter, the Nieman analysis is worth a read).

This also explains why the most astute digital publishers go for newsletters designed for mobile that are carefully – and wittily – edited by humans. (One example is the Quartz Daily Brief; it’s anecdotal but everyone I recommended this newsletter to now reads it on a daily basis.) I personally no longer believe in automated newsletters that repackage web site headlines, regardless of their quality. On smartphones, fairly sophisticated users (read: educated and affluent) sought by large media demand time-saving services, to the point content, neatly organized in an elegant visual, and — that’s a complicated subject — tailored to their needs way.

#4 The ARPU View

On mobile devices, the Average Revenue per User should be a critical component when shaping a mobile strategy. First, let’s settle the tablet market question. Even though the so-called “cheap Android” segment  ($100-150 for a plastic device running an older version of Android) thrive in emerging markets, when it comes to extracting significant money from users, the iPad runs the show. It accounts for 80% of the tablet web traffic in the US, UK, Germany, France, Japan, and even China (source: Adobe.)

The smartphone is more complicated. A year ago, many studies made by AppAnnie or Flurry Analytics showed that the iPhone ecosystem brought four times more revenue than Android. More recently, Flurry Analytics ran a story stating that the average app price for Android was $0.06 vs. $0.19 for the iPhone and $0.50 for the iPad.

The gap is closing as Android terminals attracts a growing number of affluent users. Still, compared to iOS, it is notoriously difficult to carry paid-for apps and services in the Android ecosystem, and Android ads remains cheaper. It’s likely to remain the case for quite a while as iOS devices are likely to remain much more expensive than Android ones, and therefore more able to attract high-end demographics and the ads that go to them.

How this impacts a smartphone strategy: Publishers might consider different business models for the two main ecosystems. They could go for fairly sophisticated apps in the iOS world, served  by a well-oiled payment system allowing many flavors of In-App add-ons. By contrast, the Android environment favors a more “go-for-volume” approach; but things could evolve quickly as the Android share of high-end audience grows and as the PlayStore gains in sophistication and gets as friction-free as the AppStore.

frederic.filloux@mondaynote.com

What to do with $250m in digital journalism? (II)

 

In a previous Monday Note, we looked at an ideal newsroom, profusely funded by Pierre Omidyar and managed by whistleblowing facilitator Glenn Greenwald, a structure that combines the agility of a tech startup with the highest of journalistic standards. Today, we look at the product and the business model.   

Profit or non-profit? Definitely for-profit! First, because the eBay founder’s track record (see this The New Inquiry article) shows a fierce appetite for profitable ventures. And second, because there no such thing as a free and independent media press without a strong business side: financial vulnerability is journalism’s worst enemy while profit breeds scalability. How to make money, then, with a narrow niche such as investigative journalism? Can Omidyar’s venture move beyond the cross-subsidy system that powered legacy media for decades? This weekend, in a FT.com interview, Henry Blodget justified the deluge of eye-grabbing headlines spread over Business Insider by saying “The dining and motoring sections pay for the Iraq bureau”. . .

For this, Omidyar can look at a wide set of choices: he could devise click-driven contents built on the proven high volume / cheap ads equation. Or he could opt for what I’ll call the Porsche Model, one in which the most visible activity (in this case sports car manufacturing) brings only a marginal contribution to the P&L when compared to its financial activities: in 2009, Porsche made $1bn in profit from car sales and almost $7bn betting on Volkswagen stock. More realistically, an endowment-like model sounds natural for a deep-pocketed investor like Pierre Omidyar. Most US universities are doing fine with that model: a large sum of money, the endowment, is invested and produces enough interest to run operations. One sure thing: If he really wants to go against big corporations and finance, to shield it from pressure, Omidyar should keep its business model disconnected from its editorial operation.

Investigative journalism is a field in which the subscription model can work. In France, the web site Mediapart offers a credible example. Known for, among many others feats, its investigation of the Budget Minister’s hidden Swiss bank account that led to its resignation, Mediapart maintains a newsroom of seasoned reporters working on hot topics. In five years, it collected close to 80,000 subscribers paying €9.90 per month; the web site intends to make €6m ($8m) in revenue and a profit of €0.4m ($0.5m) this year. Small amounts indeed, but not so bad for a market one fifth the size of the US. Scaling up to the huge English-speaking market, and assuming that it will go for a global scope rather than a US-centric coverage, the Omidyar-Greenwald venture could shoot for 500,000 to 800,000 subscribers within a few years, achieving $40m to $60m in yearly revenue.

On the product side, the motto should be Try Everything – on multiple segments and platforms.

Here is possible product-line structure:

298 graph

Mobile should primarily be a news updating vector. In a developing story, say hearings on the NSA scandal, readers want quotes, live blogging, snapshots – all easy to grab while on the go. Addiction must be the goal.

Newsletters deserve particular attention. They remain an excellent vector to distribute news and a powerful traffic driver. But this requires two conditions: First, they must be carefully designed, written by human beings and not by robots. Second, they must be run like an e-commerce operation: a combination of mass emailing and heavy personalization based on collected navigation data. For an editorial product, this means mapping out granular “semantic profiles” in order to serve users with tailored contents. If the Omidyar-Greenwald project lives up to its promise, it will deliver a regular stream of exclusive stuff. A cleverly engineered email system (both editorially and technically) stands good chances  to become a must-read.

User profiling must allow the creation of several verticals. Judging who will join the venture from the first bylines (see article in CNet), the coverage intends te be broad: from national security to White House politics, sports issues (a sure click-bait), civil liberties, military affairs, etc. This justifies working on audience segmentation, as not everyone will be interested in the same subject. The same goes for social web extensions: the more segmented, the better.

Web TV. If you want to go beyond kittens or Nascar crashes, providing TV contents on the web is more difficult that it appears. But “programs” available in Scandinavia show that, for developing stories, Web TV can be a great substitute for conventional TV as it allows simultaneous coverage of multiple events. Nordic viewers love that.

Fact-checking. Since the Omidyar-Greenwald project is built. t on trust and transparency, it should consider launching the equivalent of politifact.com, a fact-checking web site operated by the Tampa Bay Times, which landed a Pulitzer Prize in 2009. A vertical fact-checking site on national security, privacy and data protection issue would definitely be a hit.

Other languages. Going after the Chinese market could be hard to resist. According to Internet World Stats, it is by far the largest single market in the world with 538 million people connected to the web in 2012. For a media venture aimed at lifting the veil on corruption, China offers strong potential in itself. As far as evading censorship, it should be an appealing challenge for the squad of hackers hired by Omidyar-Greenwald.

A print version? Yes. It sounds weird, but I strongly believe that a well-designed weekly, large format (tabloid or Berliner), distributed on selected, affluent markets, would complete the product line. Print remains a vector of choice for specific, long-form readings, ambitious news scenographies with high impact photographs, for an in-depth profile or a public interest story.

Global Thinking. Its potential for worldwide reach is one of this venture’s most interesting factors. It will be of limited interest if it doesn’t embrace a global approach to public interest journalism in large democracies but also in countries that are deprived of a free press (a long list). Creating a high standard, worldwide affiliation system to promote investigative journalism everywhere, regardless of the economic and political constraints, should definitely be on the founders’ roadmap.

frederic.filloux@mondaynote.com

What to do with $250m in digital journalism? (1)

 

Pierre Omidyar, Ebay’s founder and now philanthropist, pledged $250m to a new investigative reporting venture. Starting a project of this magnitude from scratch isn’t an everyday occurrence, leading us to wonder how it could look like? (First of two articles) 

For a digital journalism project, 250 million dollars (€185m) is a serious investment. So far, it’s unclear whether this is a one-time investment, merely initial funding (Omidyar’s share in eBay is approx. $8.5bn), or just yearly running costs. To put things in perspective, The New York Times’ 1300 people newsroom costs around $200m per year, including $70m for international coverage alone, i.e. reporting abroad and maintaining 24 foreign bureaus manned by 50 reporters. But, by most measures, the scope of NYT operations is at the far end of the scale.

A more realistic example is the funding of the non-profit media ProPublica (see a previous Monday Note on the subject). According to its 2012 financial statement (PDF here), ProPublica has raised a little more than $10m from philanthropic organizations and spends less than that for a 30 persons staff. No one disputes that, journalistically speaking, ProPublica is a remarkable publication; it faithfully follows its “Journalism in the Public Interest” mission statement, collecting two Pulitzer Prizes in so doing.

Great journalism can be done at a relatively minimal cost, especially when focused on a narrow segment of the news spectrum. On the other hand, as the New York Times P&L shows, the scope and size of its output directly correlates to the money invested in its production – causing the spending to skyrocket as a result.

Since we know little of Pierre Omidyar’s intentions (interview here in the NYT and a story outlining the project), I’ll spare Monday Note readers my usual back-of-the-envelope calculations, and I’ll stick to a general outline of what a richly funded news ventures could look like.

Staffing structure. Once again, ProPublica shows the way: a relatively small team of young staffers, coached by seasoned reporters and editors. For this, Omidyar draws the hottest name in the field, namely the lawyer-activist-Guardian blogger Glenn Greenwald, who played a prominent role in the Snowden leaks (more about him: his blog on The Guardian; a NYT Magazine profile of Greenwald’s pal Laura Poitras, another key Snowden helper).

Greenwald_guardian

Multi-layer hierarchy is the plague of legacy media. The org chart should be minimalist. A management team of five dedicated, experienced editors is sufficient to lead a 24/365 news structure. Add another layer for production tasks and that’s pretty much it. As for the headcount, it depends on the scope of the news coverage: My guess is a newsroom of 100-150, including a production staff (I’ll come back to that in a moment) can do a terrific  job.

No Guild, no unions, no syndicats à la française, please. Behind their “fighting for our people” façade, they cynically protect their cushy prebends and accelerate the industry’s demise. As a result, the field is left open to pure players – who are keeping people in stables, content-recycling factories.

Beyond that, avoiding any kind of collective bargaining allows management to pay whatever will be necessary to hire and retain talent, without relying to fake titles or bogus hierarchy positions to justify their choices. In addition, above-market salaries should discourage ethically dubious external gigs. Lastly, a strict No-Kolkhoze governance must be enforced from the outset; collaboration and heated intellectual debate is fine as long as it doesn’t emasculate decisions, development, innovation – and speed.

A Journalism 2.0 Academy. I strongly believe in the training of staffers, journalists or not. Hiring motivated young lawyers, accountants, financial analysts, even scientists, and teaching them the trade of journalism is one the best ways to raise the competency level in a newsroom. It means having a couple of in-house “teachers” who will compile and document the best internal and external practices, and dispense those on a permanent basis. This is what excellence requires.

A Technology Directorate. On purpose, I’m borrowing jargon from the CIA or the FSB. A modern news organization should get inspiration from the intelligence community, with a small staff of top level engineers, hackers, cryptographers, data miners, semantic specialists. Together, they will collect data, protect communications for the staff and their sources, provide secured workstations, laptops and servers, build a mirroring infrastructure as a precaution against governmental intrusion. This is complex and expensive: It means establishing encrypted links between countries, preferably on a dedicated network (take advantage of Google’s anger against the NSA to rent capacity), and putting servers in countries like Iceland — a libertarian country and also one of the most connected in the world. While writing this, I ran a couple of “ping” tests, and it turned out that, from Europe, the response-time from an Icelandic server is twice as short as from the New York Times!

Besides assisting the newsroom, tech staff should build a secure and super-fast and easy-to-use Content Management System. Most likely, the best way will turn out to be a WordPress system hack – as Forbes, Quartz, AllThingsD, and plenty of others did. Whatever the setup ends up being, it must be loaded with a powerful semantic engine, connected to scores of databases that will help enrich stories with metadata (see a previous Monday Note on the subject The story as gateway to knowledge). By the same token, a v2.0 newsroom should have its own “aggrefilter”, its own Techmeme that will monitor hundreds websites, blogs and twitter feeds and programmatically collect the most relevant stories. This could be a potent tool for a newsroom (we are building one at Les Echos that will primarily benefit our news team.)

Predictive Analysis Tools and Signal-to-Noise detection. In a more ambitious fashion, an ideal news machine should run analytics aimed at anticipating/predicting spasms in the news cycle. Pierre Omidyar and Glenn Greenwald should acquire or build a unit like the Swedish company Recorded Future (more in this story in Wired UK), which is used by large corporations and by the CIA. Perhaps more realistically, building tools to analyze and decipher in realtime the internet’s “noise”, and being able to detect “low-level signals” could be critical to effectively surfing the wave.

That’s all for today. Next week, I’ll address two main points: Designing modern news products, and ideas on how to make (some) money with this enthralling venture.

frederic.filloux@mondaynote.com

The Age of the Platform

 

Before deciding what should comes “first” in digital, publishers must figure out the right production workflow. Each and every player must plot its very own path away from the now aging notion of publication to the broader platform model. 

Last week, I spent a few of days in Berlin at the European INMA conference. Among many interesting moments, there was our visit to the Axel Springer group, the number one print publisher in Germany that also operates scores of publications in 44 countries. In 2012, Springer had a revenue of €3.3bn and an EBIDTA of €628m; 40% of its revenue comes from digital, thanks to 160 different online properties and 120 applications. Attaining this level required an aggressive growth strategy: since 2006, Springer launched or acquired new digital activities at the stunning rate of one every two weeks!

Like most modern news outlets, Springer is obsessed with having everyone in the company work without distinction between digital and print. Its latest initiative involves the definitive transformation of the venerable daily Die Welt into a multimedia news factory. To achieve this, the company bets on the radical architecture of its brand new newsroom. Of course, Die Welt is not the first to bet on the physical setting of the workplace to accelerate changes. Among others, the UK’s Telegraph did the same several years ago (it didn’t go smoothly at first but, in the end, the effort paid back.)

Here is the floor plan of the Die Welt’s newsroom that will enter in operation within a couple of months (I reconstructed it from a picture and briefing notes) :

die_welt_newsrm_plan

The open space resembles a sound-proof cathedral on the ground floor of the Axel Springer building in the center of Berlin. It will operate from 5am to midnight. The star shape reflects the news products’ diversity and time imperatives; the closest the workstations are from the center (where on-duty management sits), the faster the treatments are supposed to be: mobile staffers will stay close to the top editors as people in charge of building pages for the daily will dwell at the outer edges. This newsroom is mostly a production center; it actually accommodates only half of the Die Welt 300+ editorial staff as reporters and some staff writers will be located in a separate room. Note how all individual offices are gone while the periphery is filled with meeting rooms of various sizes and shapes that staffers use as needed.

Management gurus often say a radical alteration of physical settings is a key instrument of change. I can’t agree more. Interestingly enough, a firm like Innovation Media Consulting I’ve known since the Nineties as mostly an art direction company now works with architects and workflow specialists to induce changes in the way newsrooms operate.

But a super-modern floor plan is only part of the equation. In last week’s Monday Note,  I addressed the need to make the story the kernel of a cluster of high value products. Both are merely components of a much deeper change, that is the creation of a true News Platform. Anglo-Saxon newsrooms enjoy several advantages over Southern Europe (for instance) ones. Since the beginning, their journalism is built on a clear separation between writers (or reporters) on one side, and editors on the other. Anglo-Saxon journalism comes embedded with a separation between the writing and the editing of journalistic material — that is not the custom in a country like France in which most interns sees themselves as potential heirs to Joseph Kessel. More seriously, here, the principle of heavy editing is much less accepted than in the US, UK or Germany where the process results in much better structured articles, and most powerful storytelling for long-form reporting. In addition, in those countries, newsrooms with top editors entirely dedicated to their role of managers are better equipped to address the needs of morphing news organizations. For the most part, these factors explain why, in the Anglo-Saxon world, the News Platform transformation is way ahead of anywhere else. Axel Springer’s management concedes that this radical news flow structure is the result of a process that started years ago — that’s why it has been smoothly accepted by the staff. Everyone now sees it as the indispensable platform to produce across all major vectors now used by the readers – mobile, tablets, web and print – with greater efficiency along with consistent quality,.

frederic.filloux@mondaynote.com

The story as gateway to knowledge (and revenue)

 

In digital journalism, the article is no longer an end in itself. Quite the contrary, it’s an entry point to the depths and riches of the web, and a significant contributor to the revenue stream.

Last week in Paris, I met the representative of a major US tech firm in charge of content-based partnerships. This witty, fast-thinking young engineer toured European capitals for an upcoming web + mobile platform, meeting guys like me in charge of digital operations in large media companies. Our discussion quickly centered on the notion of article in the digital world. Like many of his peers (I can’t  name them otherwise you might triangulate with whom I spoke), he looked at the journalistic article in an old-fashioned way: a block of text, augmented with links here and there, period.

This no longer is how it works — or how it should work.

There are many forms of digital journalistic contents. They range from the morning briefing you’ll eat up on your smartphone while inhaling your breakfast, or during your commute to work, to the long-form piece aimed at lean-back reading, preferably on a tablet and with a glass of chilled chardonnay. In between, there is the immense output of large media outlets that create good original content, hundreds of pieces every day.

If we draw a quick matrix of contents vs time and devices, chances are it will look like this:

usages, devices

As the graph shows, in a ideal world, a news stream should be broken into multiple formats to fit different devices at different times of the day. Of course, the size of the bubbles depicting usage intensity varies by market.

Three notes: the smartphone appears as the clear winner with high usage, spread all over the day; tablets enjoy the largest scope of contents (plus the highest engagement). As for the PC, it has been evicted as a vector for mainstream, general news. Still, thanks to its unparalleled capabilities and penetration as a productivity tool, the PC retains the most of the business uses. Consequently, the news read on a PC, largely in the context of a professional use, carry a greater value — as long as the article is linked to three different functions:

First as an audience concentrator from multiple sources, see here:

traffic drivers

Second, by building a system in which the article becomes an entry point to the web’s depths, i.e. to the trove of publicly and freely available databases. To get an idea of the open web’s riches, see the image below and click this link to dive into it:

debpedia_colored

 

This two-year old graph was designed by University of Berlin computer scientists. All these datasets are up for grabs by editors and publishers willing to expand their contents. Every single piece of news can be greatly augmented by hundreds of datasets orbiting around the DNpedia Knowledge Base (part of the Wikipedia Project.) According to its official description, the English version of DPpedia describes 4 million objects, including:

  • 832,000 persons
  • 639,000 places (including 427,000 populated places)
  • 372,000 creative works (including 116,000 music albums, 78,000 films and 18,500 video games)
  • 209,000 organizations (including 49,000 companies and 45,000 educational institutions), 226,000 species
  • 5,600 diseases.
  • When extended to the 119 available languages, the number of objects rises to 25 million.

The third way to raise the value of editorial contents is to use the article as a promotional vehicle for a broad set of ancillary products that media organizations should develop:

article upsell

(Needless to say, in this chart, Church and State must remain separated: the article is to be a journalistic product, aimed primarily at informing the public; the “promotional” aspect being only secondary.)

Until now, connecting to multiple datasets and up-selling extra products weren’t priorities for most legacy media. The main reasons are well-known: insufficient technological culture and investments  — which left the field totally open to pure players that made a modern, productive use of both datasets and new commercial channels. Things are changing though. Slowly.

frederic.filloux@mondaynote.com

The Quartz Way (2)

 

Last week, we looked at Atlantic Media’s business site Quartz (qz.com) from an editorial and product standpoint. Today, we focus on its business model based on an emerging form of advertising. 

The Quartz business model is simple: it’s free and therefore entirely ad supported. Why? Doesn’t qz.com target a business readership that shouldn’t mind spending nine dollars a month? “It was part of the original equation: Mobile first, and free, embracing the open web”, explains publisher Jay Lauf, whom I met in Paris a couple of weeks ago. Jay is also an Atlantic Media senior vice-president and the group publisher (he once was Wired’s publisher).

jay-lauf-head-shot

Jay Lauf, Publisher (Photo: Quartz)

According to him, launching Quartz was the latest iteration of a much grander plan. Four years ago, Atlantic Media held a meeting aimed at defining their strategy: “What we will do, but also what we will not do”, says Jay Lauf. The group came up with three key priorities: #1 being a growth company (as opposed to passively manage the shift from print to digital). That idea was greatly helped by Atlantic’s ownership structure controlled by David Bradley. #2 “Digitally lead for everything”, which was not obvious for a ancient publication — Atlantic Monthly was created in 1857. #3 Atlantic must focus on “decision makers and influential people”.

Today, the goals set four years ago translate into a cluster of media brands reaching every month a highly solvent readership of 30 million people:

  • The Atlantic, the digital version of the eponymous magazine.
  • The Atlantic Wire aimed at a younger generation mostly relying on social media.
  • The Atlantic Cities, that focuses of urban centers and urban planning.
  • The National Journal that itself includes several publications, mostly about politics and society.
  • Government Executive Media, which operates a number niche publications covering the federal government (including its use of technology)
  • Atlantic Media Strategies, an independent division offering a full catalogue of advertising and marketing solutions. These range from analytics, social media campaigns and content creation, such as this one with General Electric in which a dedicated site features America’s economic futures – according to GE.

quartz_graph

All this brings us to Quartz’s business model. It relies entirely on native advertising also known as branded or sponsored content (see a previous Monday Note What’s Fuss About Native Ads?). Quartz’s implementation is straightforward: a small number of advertisers, served with high yield campaigns.

Below is yesterday’s screenshot of Quartz’s endless scroll, featuring regular displays of branded content (in this case Boeing):

qz_scroll_ads

Most of the time, the content is made or adapted especially for Quartz with a variable involvement of its advertising division (the branded content operations are kept segregated from the editorial department.) Quartz staff involvement goes from collaborating on the ad content to setting up HTML5 integration. On purpose, Quartz maintains a staff of copywriters and graphic designers assigned to assist brands in their communication. While ad spaces are clearly identified, their content is never completely dissociated from surrounding articles. Quite often, it reflects the newsroom’s “Obsessions“. Such precautions, plus the Quartz layout, warrant good click-rates and high prices. Quartz people are discreet about the KPIs, but sources in the ad community said that CPMs for its native ads content could be roughly ten times higher than traditional display ads.

Atlantic Media’s weight and bargaining power helped jumpstart the ad pump. A year ago, the site started with four brands: Chevron, Boeing, Credit Suisse and Cadillac. Today, Quartz has more twenty advertisers from the same league. Unlike other multi-page websites, its one-scroll structure not only proposes a single format, but also re-creates scarcity. (Plus the fact that Quartz does not have any mobile apps greatly simplifies the commercial process.) Still, it can be a double-edged sword: scarcity could indeed translate into high prices, but it also limits the number of available slots, therefore capping the revenue stream. Quartz’s publisher and head of sales made a tough choice — high rates vs. high volume — and so far it seems to work fine as the site is close to break-even ahead of schedule.

How far it can go remains to be seen. Quartz is a relatively small operation (50 people altogether, including 25 journalists producing 35-40 stories a day and a nice location in NYC’s Soho district.) My guess is it shouldn’t burn more than $10m a year. By extrapolating from the site’s audience, profitability sounds in reach of Quartz’s current “value model”. But the asymptote — factoring ads rates, number of slots, advertisers’ “dimension”, and traffic — could also be near and therefore constrain Quartz’s ability to scale up. That’s why the publication is now entering the crowded sector of conferences with its “Quartz Live”, featuring its customary exclusive attendance and editorial-rich ways. Will Quartz escape the temptation to launch paid-for products? Its journalistic content leaves open many opportunities in that field. For example, a mixture of semantic-assembled, high-end briefings, tailored to carefully profiled segments of its audience could generate a nice revenue stream, or ebooks and long-form features.
To be continued next year…

frederic.filloux@mondaynote.com

 

The Quartz Way (1)

 

Quartz, a web-only business publication, just turned one year old. On both editorial and business dimensions, Quartz features all components of a modern media venture. Is this a formula for the long run? To answer the question, in the first of two articles, we take a closer look at the editorial product.

Quartz (qz.com) is the kind of media most business writers would love to be part of. It’s smart, fun, witty, basic and sophisticated at the same time. Like Jony Ive design at Apple, its apparent simplicity is the combined product of deep thought and of a series of bold moves by its owner, the Atlantic Media group, publisher of the eponymous monthly. From all standpoints, content, organization or even business model, Quartz came up with innovations (see the Monday Note I wrote for the launch in September 2012).

Ten days ago, my phone interview with editor-in-chief Kevin Delaney, started with a discussion of his newsroom of 25 writers and editors. On Tuesday September 24 at 9pm Paris Time, Quartz had this piece at the top of its infinite scroll:

Quartz illustr

Editorially, this epitomizes (in a way) what Quartz is about: topics addressed through well-defined angles (in this case, the idea that if Amazon hit large book retailers hard, it didn’t have much impact on small independent bookstores.) The story was short but right to the point — taking the opposite side of the now worn tale of Amazon devastating the book-selling landscape. To illustrate his piece, instead of using yet another photograph of Jeff Bezos haranguing a crowd, the writer picked this weird image of a girl showing off at a bookstore event.

Yes, at Quartz, journalists are the ones who get to select the pictures that go with their article. Most of the time, this yields better audience numbers.

Actually, explains Kevin Delaney, the staff is supposed to produce a complete package, ready to be processed by editors, with links, headline, photos (lifted from Reuters, Getty, AP or sometime the Creative Commons trove) properly cropped and adjusted. Everything is done within a WordPress interface, chosen for its versatility, but also because most journalists already know to use it. As for headlines (the task usually handled by editors), the Quartz newsroom relies on team chats to quickly and collaboratively work on pieces.

kevin_delaney
Kevin Delaney (photo: Quartz)

The same goes for graphics like in this snapshot of Tweeter’s IPO prospectus, a part of the magazine’s comprehensive coverage of the upcoming event. To further encourage the use of graphics and charts in stories, Quartz engineering director Michael Donohoe (a NYT alumni) ChartBuilder, a bespoke, easy to use tool.  [Correction : as pointed out by Quartz'global news editor Gideon Lichfield, ChartBuilder has been developed by David Yanofsky, one of Quartz journalist/coder/data hackers...] As an internet-native company, Quartz threw its software in the open-source world (see how it looks in Github) — an unthinkable move in the close-to-the-vest legacy media world…

While listening to Delaney describing his organization, I couldn’t help but mentally itemize what separates its super-agile setup from traditional media. A couple of months ago, I met the digital management of a major UK newspaper. There, execs kept whining about the slow pace evolution of the news staff and the struggle to get writers to add links and basic metadata (don’t even think about pix or graphics) to their work product. By and large, most legacy media I know of, in France, UK and the United States, are years behind swift boats such as Quartz, Politico or the older but still sharp Slate.

I used to think the breadth and depth of older large newsrooms could guarantee their survival in a digital world plagued by mediocrity and loose ethics. But considering great pure players like Quartz — which is just the latest offspring of a larger league — I now come to think we are witnessing the emergence of a new breed of smaller, digital-only outlets that are closing the gap, quality-wise, with legacy media. In the context of an increasingly segmented and short-on-time readership, I can only wonder how long the legacy newsroom’s strategic advantage of size and scope will last.

Quartz editorial staff has nothing to do with the low-paid, poultry farm newsrooms of many digital outlets. Most of the 25 journalists and editors (out a staff of 50) were drawn from well established brands such as Bloomberg, The Economist, Reuters, New York Magazine or The Wall Street Journal (Kevin Delaney, 41, is himself a former WSJ.com managing editor). “Our staff is slightly younger than the average newsroom, and it is steeped in the notion of entrepreneurial journalism”, says the Quartz editor-in-chief. “With Quartz, we had many opportunity to rethink the assumptions of traditional media”.

The original idea was to devise how The Economist would look like if it had been born in 2012 rather than in 1843, explains Delaney. It would be digital native, mostly for mobile reading, and focus on contemporary economic engines such as digital, globalization, e-commerce, the future of energy, debt, China, etc. Instead of abiding by the usual classification of business news that looks like a nomenclature from the Bureau of Labor Statistics  (Industry, Services, Markets, Trade, etc.), Quartz opted for a sexier taxonomy; its coverage is based on an evolving list of “Obsessions“, a much more cognitive-friendly way to consider the news cycle than the usual “beat” (read this on the matter). As an avid magazine reader, Delaney said he derived the idea from publications like New York Magazine.

The challenge is connecting this categorization to audience expectations… Hence the importance of the social reverberation of Quartz treatments. They translate into stunning numbers: according to Kevin Delaney, 85% to 90% of its traffic is “earned” and social referrals account 50% of the site’s traffic. In other words, the traffic coming from people typing http://qz.com in their browser accounts for only 10-15% of the volume. To put things in perspective, on a legacy media site, social traffic weighs about 5% — in some rare cases 10% — and around 40% to 50% of the pages views are generated via the home page.

Since the site is nothing else but an infinite rolling page of stories, there is no classic jumping board home page. Another obsession of Quartz founders: “We wanted to minimize friction and encourage readers to share our stories. We designed the site first for tablets, then for mobile and as a classic website, in that order,” insists Kevin Delaney. No apps in sight, but a site built in HTML5 and responsive design that adjusts to screen size. At first, the no-app choice sounded weird for a media aimed at a mobile audience, but considering the rising costs and complexity of building, managing, and maintaining native apps on multiple platforms, a single HTML design was probably the best approach.

I’m not through talking about Quartz. Next week, we’ll examine the venture’s business aspects, its bold ways of dealing with advertising.

frederic.filloux@mondaynote.com

News: Personalized or Serendipitous?

 

Every digital news designer faces the question: should the traditional serendipity of contents be preserved or should we go full steam for personalization? It turns out Google is already working on ways to combine both — on its usual grand scale.

Serendipity always seemed inseparable from journalism. For any media product, taking readers away from their main center of interest is part of the fabric. I go on a website for a morning update and soon find myself captured by crafty editing that will drive me to read up on a subject that was, until now, alien to me. That’s the beauty of a great news package.

Or is it still the case? Isn’t it a mostly generational inclination? Does a Gen Y individual really care about being drawn to a science story when getting online to see sports results?

Several elements concur to the erosion of serendipity and, more generally, curiosity.

First, behavioral among digital readers are evolving. These extend far beyond generations: Regardless of her age, today’s reader is short on time. At every moment of the day (except, maybe, in the loo or in bed at night), her reading time is slashed by multiple stimuli: social teases, incoming mail, alerts or simply succumbing to distractions that lie just one click (or one app) away. That’s one of the tragedies of traditional news outlets: When it comes to retaining the commuter’s attention, for instance, Slate or The Washington Post are in direct competition with addictive products such as Facebook or Angry Birds…

Second, the old “trusted news brand” notion is going away. Young people can’t be bothered to leaf though several titles to get their feed of a variety of topics; that’s why aggregators thrive. The more innocuous ones, such as Mediagazer, mostly send traffic back to the original news provider; but legions of others (Business Insider, The Huffington Post…) melt news brands into their own, repackage contents with eye-grabbing headlines and boost the whole package with aggressive marketing.

Below, see how BuzzFeed summed up the New York Times story on the NSA monitoring social traffic: 80 words in BF that capture the substance of a 2000 words article by two experienced journalists who collected exclusive documents and reported from Washington, New York and Berlin. buzzfeed nyt

(Note that BuzzFeed is serving a more appealing headline and a livelier photograph of general Keith Alexander, head of NSA.) How many BuzzFeed glancers did click on the link sending back to the original story? I’d bet no more that 5%. (Anyway, judging by the 500 comments that followed it, the NYT did well with their article.) This trends also explains why the Times is working on new digital products that take into account both time scarcity and the Gen Y way with news.

This leads us the third reason to wonder about personalization: the economics of digital news. In the devastated landscape of online advertising, it became more critical than ever to structure news content with the goal of retaining readers within a site. That’s why proper tagging, use of metadata, semantic recommendation engines and topic pages entries are so important. More pages per visit means more ads exposure, then more revenue. Again, pure players excel at providing incentives to read more stuff within their own environment, thus generating more page views.

Coming back to the customization issue, should we turn the dial fully to the end? Or should we preserve at least some of the fortuitous discovery that was always part of the old media’s charm?

Let’s first get rid of the idea of the reader presetting his/her own preferences. No one does it. At least for mainstream products. Therefore, news customization must rely on technology, not human input.

Last week, I spoke with Richard Gingras, the senior director of news and social products at Google (in other words, he oversees Google News and Google + from an editorial an business perspective). Richard is a veteran of the news business. Among many things, he headed Salon.com, one of the first and best online publication ever.

gingras

According to him, “Today’s news personalization is very unsophisticated. We look at your news reading patterns, we determine that you looked at five stories about the Arab Spring and we deduct you might like articles about Egypt. This is not how it should work. In fact, you might be interested in many other things such as the fall from grace of dictators, generation-driven revolutions, etc. These requires understanding concepts”. And that’s a matter Google is working on, he says. Not only for news, but for products such as Google Now which is the main application of Google’s efforts on predictive search. (Read for example With Personal Data, Predictive Apps Stay a Step Ahead in the MIT Technology Review, or Apps That Know What You Want, Before You Do in the NYTimes).

The idea is to connect all of Google’s knowledge, from the individual level to his/her social group context, and beyond. This incredibly granular analysis of personal preferences and inclinations, set in the framework of the large macro-scale of the digital world, is at the core of the search giant’s strategy as summed-up below:

google infos2

On the top of this architecture, Google is developing techniques aimed at capturing the precious “signals” needed to serve more relevant contents, explains Richard Gingras. Not only in the direct vicinity of a topic, but based on center of interests drawn from concepts associated to individuals’ online patterns analyzed in a wider context. In doing so, Gingras underlines the ability of Google News to develop a kind of educated serendipity (term is mine) as opposed to narrowing the user’s mind by serving her the unrefined output of a personalization engine. In other words, based on your consumption of news, your search patterns, and a deep analysis (semantic, tonality, implied emotions) of your mail and your posts — matched against hundreds of millions of others — Google will be able to suggest a link to the profile of an artist in Harper’s when you dropped in Google News to check on Syria. That’s not customized news in a restricted sense, but that not straightforward serendipity either. That’s Google’s way of anticipating your intellectual and emotional wishes. Fascinating and scary.

frederic.filloux@mondaynote.com

 

Memo #3 to Jeff — Data & User Profiling for The Washington Post

 

For customer-related technologies, the financial and intellectual backing of Jeff Bezos, and his Amazon experience can give The Post a huge competitive advantage. Here is what should be at the top of the to-do list. 

Every digital manager must plan to tap into Amazon’s fantastic engineering firepower. (Even though Bezos bought the newspaper out of his own pocket, the first thing he’ll do — if he hasn’t already — will be drafting some of his techies as “advisors” to The Post.) The key point being: the influx of engineering brainpower must not be limited to the digital side of the house, or to the newspaper’s IT infrastructure. It should impact all activities: editorial, marketing, subscriptions and paid-for products. Let’s dive into details.

Turbo-boosting the editorial. Let’s start with the basics: What characterizes media outlets playing in The Washington Post’s league? It is their ability to line up top journalistic resources to cover stories that matter, in-depth, with multiple angles and treatment modes (text, features stories, photographs, graphics, multimedia storytelling, live blogging, opinions, etc.), while deploying the best expertise on topics covered. These are the five items that make the difference between the bulk of pure players and true legacy media.

In many ways, the above is anti-economic, it is loaded with inherent inefficiencies — dry holes, dead ends, waste of time on promising leads —  that drive nuts “quant zealots” obsessed with KPI’s and productivity measurements. At this point, the difference between great newsroom managers (i.e. editors) and average ones lies in their ability to make some room for “managed inefficiencies”. An editor’s key, delicate duty is weighing the purpose of resource-intensive tasks such as flummoxing the competition, pursuing a worthy story, or launching a months-long journalistic project aimed at a Pulitzer prize. Unfortunately, weak leadership, balking at tough choices and yielding instead to a sorry attempt to spread an even level of (dis)satisfaction among constituencies causes inefficiencies to grow like weed.

The foremost goal of technology-enhanced news content is smartly weaving together all components of a topic. The idea is to keep the reader aboard by encouraging multiple levels of reading, with different angles for a subject, calls to essential archives or to other forms of journalism such as blogs or infographics. In this field, Amazon is light-years ahead of the news industry. By raising the number of editorial treatments seen by the reader, almost twenty years of Amazon’s e-commerce recommendation engine refinements will undoubtedly benefit The Post.

Another key item will be the level of news personalization. What should a Post reader see mostly? News that matters to him or her, or everything the paper’s staff collects? How to define mostly? Fully tailored contents based on past navigation? Stated preferences combined with the preserved serendipity that together make the core of news construction? This is a deeply involved problem — and the subject of a future Monday Note.

Reader profiling. All digital publishers dream of knowing exactly what reader sees what content, where, at what time of the day and on which vector: web, smartphone, tablet. The finer the granularity, the better. Slicing and dicing readership in segments of age, professions, residence, income, interests yields three types of uses:

  • increasing news content stickiness by serving customized content as mentioned earlier
  • smarter customized advertising, as opposed to dumbly drowning users into a flood of ads for months by using data collected during the shopping season. This practice, known as “retargeting”, is one of the internet “seven plagues” and the most potent repellent to advertising
  • channelling the reader to the catalogue of ancillary products any news outlet should operate. For example: once a reader is identified (even anonymously) as working in the legal field, for a media group struggling to fill the last seats of its conference on privacy laws, why not show this loyal reader a one-time only, 50% discounted ticket, valid for 24 hours only? Simplistic as this example might seem, its large scale application is far from trivial: it requires super-accurate analytics, the deployment of “event engines” that will trigger the display of the right offer, at the right time, to the right segment of the population. Fortunately, this is the kind of work Amazon geeks are particularly good at.

For The Washington Post, the benefits are numerous. Research shows that serving the right ad to the right profile can raise its value by a factor of 1.5x to 2x. And the performance of ancillary products (conferences, business events, news-related ebooks or professional products, education packages, etc.) will become easier to measure.

Impact on paywall and subscription models. Paywall theory can be summarized as follows:

  • deploying a wide range of tactics all aimed at significantly raising the number of news contents items (not necessarily articles) a reader watches every month. Let’s make no mistakes: the main dial is under the newsroom’s control, marketing wizardry won’t do the trick
  • finding readers most likely to convert to a paid-for subscription and, week after week, serving them (I write serving, not bombarding) offers they can’t refuse: an extended test-period, or a news-related bonus that reflects the breadth of the company’s line of products.

As with most theories, practice is much harder. A paid-for system is a long-term, investment-intensive, staffing-critical effort. Two legacy media did it particularly well: The Financial Times and The New York Times. The former built a subscription base that now surpasses the paper’s; the latter added $100m a year in revenue that did not exist three years ago. Most paywall strategies underperform for two reasons: first, an error in predicting the editorial contents’ ability to retain readers beyond a free threshold of 10, 15, or 20 stories a month; second, a failure to build the data-driven infrastructure that is mandatory for any paid-for product. The Washington Post does relatively well with the first test. For the second, the backing of Amazon tech brains will give it the best chances to succeed.

frederic.filloux@mondaynote.com