How AI and Chinese-like business models could reshape digital media

In January 2018, Facebook announced that it was going to distance itself from news by de-prioritizing it from the News Feed. Just one month later, media companies started feeling the impact. LittleThings, which was reaching a 50 million audience seemingly overnight thanks to Facebook, was no more. Mic, a millennial news publisher, suffered after the social network canceled a key video programming deal with them and was eventually sold to Bustle for only $5 million. As 2018 comes to a close, it’s becoming clear that unless new business models can be adapted into the U.S. market, media companies that rely on a solo distribution channel/business model will face an uphill battle.

A pivotal moment

Google and Facebook are well known to have an iron grip on digital ad revenue. Logically, many publishers tried to match those platforms’ expectations by making wholesale shifts in their business models, but more often than not couldn’t transition and diversify fast enough. As content could suddenly live on a social platform instead of a publisher’s website or a dedicated place, many prioritized growth and scale over monetization. The business model would be figured out later, and a whole new crop of media was born. However, under such conditions, as underlined by rich antoniellomedia [became] a game of musical chairs with too many players and too few chairs”. This meant that only media companies that could build extremely strong brands would prosper (Complex, Vice, Vox…). Their focus on diversifying revenues and building a recognizable brand identity over the years would give them a leg up in the industry. Other players like Buzzfeed also managed to diversify revenues through partnerships with companies like Walmart through Tasty and others lifestyle brands. Yet, even the latter continues to struggle monetizing quality content on Facebook and other social platforms. While Jonah Peretti suggests Facebook should share revenues on paywall, the thought of the platform sharing its revenues on the newsfeed is highly improbable and doesn’t make sense for them as a “social” company. Now a new wave is coming, many news media companies including Business Insider and Buzzfeed News begin relying on paywall in order to monetize quality content, which could be part of the answer.

Are paywalls the answer?

A decade ago, the idea of restricting access to your content was anathema to newsrooms, which at the time were trained by the ad model to maximize traffic. Since then, the adoption of paywalls has accelerated since publishers realized digital news couldn’t survive on advertising alone. Paywalls are an amazing tool to finance great journalism and can be a powerful business model if your brand is as strong as the New York Times or the Washington Post. Loyal audiences are more than willing to pay for the quality of those media sources, explaining why average subscribers are usually in their 40s. Paywalls, however, have a hard time attracting a younger audience. Gen Z and Millennials don’t understand why they should have to pay for content from one media source when they are used to having virtual access to an infinite amount of sources right at their fingertips. They would rather opt for an OTT-like model similar to Netflix or only pay for the content they specifically need. This makes paywall a limiting business model that favors older generations but lacks the support for new players to unleash their potential.

New York Times’ paywall

How China solves this issue through innovative business models

Historically, innovation in Silicon Valley consisted of having a strong business model and an impressive product. For example, Airbnb created a unique product, but it was its business model that really changed the game and allowed the company to scale at an unprecedented speed. The idea of a host allowing a guest to pay immediately had never been seen before. However, despite the hardships they encountered, media platforms didn’t innovate their business models. As shown in the graph below from Andreessen Horowitz, in the last decade, large consumer Internet companies in the media space (Facebook, Spotify, Netflix…) have not developed very diverse ways to monetize. They are either advertising-driven or transaction/subscription-driven. As mentioned by connie chan in a recent articlerevenue is heavily concentrated in one of these two business models, which in turn drives how the companies think about creating product. For example, advertising-driven companies focus on increasing engagement and time in app; transaction-driven services optimize for the lowest friction before checkout.”

Most revenues come from advertising or subscriptions (source: a16z)

With most media falling into one of those two categories, it might appear that there is no place in between, but there is another way. China has created a new business model by thinking about content consumption from a mobile-first perspective, which not only provides diversified revenue streams for content creators but also allows users to make better, more flexible purchasing decisions. For most popular products, Chinese users are able to consume content in a serialized and snackable way designed for mobile users. They can tip an author, buy a book chapter by chapter or subscribe to a podcast they like for a few dollars. Importantly, in most Chinese apps, from e-commerce (Taobao, Alibaba…) to media (Toutiao, Zuiyou…) there are elements of social interaction, gamification, and powerful discovery elements that create an experience driving sales and building a strong network effect.

In her article, connie chan uses the example of podcasting to draw a powerful comparison between China and the US.

The podcast Serial was downloaded 40 million times during its first three-month season in October 2014. This wildly successful spin-off of This is American Life likely perhaps earned up to one million dollars in ad revenue, according to the industry standard of $25 per 1,000 downloads. Revenue might even have been lower, as rates were negotiated well before anyone knew how successful the podcast would become.

Despite being deemed ‘podcasting’s first breakout hit’, revenue couldn’t cover the cost of producing a second season. The host/executive producer Sarah Koenig started asking listeners for donations during episode nine.” […]

In China, many podcasts charge the end customer instead of relying solely on advertising. Some podcasts are part of the platform’s paid subscription service, but users always have the option to pay for podcasts individually.

Beyond the scale and business models, what else makes these Chinese podcast platforms different from their Western counterparts? Similar to book platforms, these apps hook users through social elements and gamification. Ximalaya has a leveling concept, where you ‘level up’ by listening or spending money in the app to earn coupons and even limited-time membership to use their paid subscription service.”

Imagine how powerful a platform could become if it properly leveraged all those tools with the amount of digital media content produced every day in the US every day. This is what we at Snipfeed are building. For too long, media sources have relied on platforms to provide reach and scale, but these platforms weren’t designed for their content and wouldn’t help them to solve the ARM formula (acquisition, retention, monetization). In order for a vastly diverse amount of news sources to exist, this formula needs to be solved on mobile first, so that new generation of users who barely use their computers can enjoy and financially support creators.

Tencent Revenue distribution (source: a16z)

Mobile first

The success of those Chinese platforms also resides in the fact that they are mobile-first. For many in the US, mobile first is still pretty unclear. The first answer people give you when you mention being mobile first is “everything is already on mobile, there are so many apps out there”. But mobile first is much more than that, it is about creating new experiences for mobile only. As mentioned by Connie Chan in a conference last year, for Chinese companies, “Mobile First” really means utilizing “features that are mobile specific, that are specific to mobile hardware.” In other words, it is about “designing for mobile first, knowing all the different sensors and all the different abilities that smartphones allow you to have she says. They utilize literally every single part of a smartphone.

The best example of such a platform is Toutiao, which launched as smartphone use was taking off in China. Their mobile penetration increased from nearly nothing in 2010 to 65% by 2014. Moreover, many of the largest content providers had not yet developed mobile apps or mobile-friendly sites, meaning that true mobile-optimized information and entertainment was rare. Which is broadly still the case in the US today. Popular news apps like Apple News do a great job, but the format didn’t really evolve in the last 10 years. On most sites, you still open a link that leads to an article largely similar to a print newspaper. Toutiao, on the other hand, created a new format from scratch, which also explains why over 120M people use it each day. The average user spends more than 74 minutes each day in Toutiao , which is more than the average user spends on Facebook, and more than twice what they spend on Snapchat.

Drake is definitely mobile first✨

AI + innovative business models=monetization at scale

One big difference between China and the US is also the way Chinese platforms leverage AI to empower publishers. From Bytedance to Tencent, the last two years gave birth to a number of AI consumer apps. What’s most interesting about platforms like Toutiao isn’t that people consume such varied content all in one place… it’s how Toutiao serves it up. Without any explicit user inputs, social graph, or product purchase history to rely on, Toutiao offers a personalized, high quality-content feed for each user that is powered by machine and deep learning algorithms.

AI, in this case, is extremely interesting from a business model perspective. First regarding ads, while most mobile ads are still banner ads creating painful experiences, AI makes products extremely good at identifying what its users want to see. It is fitting, then, that the business model maps perfectly to that strength. Toutiao, for instance, generates revenue by matching relevant ads to users, using the same proprietary technology behind their content targeting. As underlined by Anu Hariharan, “this has three important benefits:

First, it reduced the impact of monetization on the user experience — and may have actually improved the experience! Users typically consider ads as intrusive and degrading to their experience, but ads aligned with user preferences are less so.

The second is that it increased the rates that Toutiao could charge advertisers. One of the key problems in advertising is identifying how to selectively place your ads in front of the highest potential customers, and advertisers spend countless hours and enormous sums of money trying to target effectively.

Third, since the primary use case is to read and view content, users are more receptive to seeing relevant targeted ads and therefore there is more inventory available to advertisers.

The combination of all these three factors results in much better CTRs (Click Through Rates) on Toutiao vs. competitors.

AI can also help users find what they want to see and what they are more likely to pay for. By creating homogenous cohorts, machine learning makes it easy for content creators to reach the right audience. This model works particularly well for soft-news like sport or music, but not necessarily for politics. Staying in your own bubble has obvious harmful effects when it comes to politics, however, at Snipfeed we thwart this phenomenon known as ‘echo chambers’ by making sure users see at least 30% of content outside their “comfort zone”.

AI can also be used to spot preventively potential fake news and clickbait in order to create next level experiences for users (Snipfeed developed Fakebuster to do so). But overhaul such a model is highly beneficial for media themselves continues Anu HariharanStrong incentives via revenue sharing enabled writers to make money from very early on. In 2014, Toutiao rolled out incentive programs to attract more content creators to the platform. These ranged from offering office space, tools, minimum guarantees per month if they hit certain key milestones (e.g., # of articles, read rates) to sharing revenue via monetization. Toutiao began monetizing via ads since 2014 and this enabled revenue sharing opportunities with their content contributors.”.

At the end of the day, we are just at the beginning of a content revolution driven by mobile and AI. Once these tools are better understood, their technologies will give birth to new platforms that will redefine the industry. Quibi founded by Jeffrey Katzenberg and Meg Whitman, for instance, will launch next year in the hopes of redefining TV on mobile. At Snipfeed, we believe we can leverage those complex technologies to create a whole new experience with news, and raise awareness worldwide in these critical times. By collaborating with publishers and content creators, we believe we will create a sustainable platform that reaches mobile audiences in droves in a new, personalized and truly unique way.