Follow by Email

Search This Blog

Nov 21, 2018

Publishers need better technology platforms

I have been literally preaching for years that good, quality content (in the eyes of the target group) is "just" a precondition, and that the single most important source of competitive advantage for digital publishers will be technology. The fact that international publishers with a wealth of titles and platforms often fail to establish a common tech platform (CMS, video players, recommendation algorithms, data collection...) on which the brands can build their own offerings in order to capitalize on scale is sad enough, but maybe initiatives across different publishing houses may help out here. We can see increased cooperations in the German market (like ZDF, Discovery and ProSieben joining forces in on-demand video), so maybe these will go further in future. This article by Jesse Knight, former CTO/CIO of Vice, makes that case in a compelling way:

Nov 16, 2018

AI in football match analyis

Any football match (some call it soccer, but to me, it's football, and Amercian football is "handegg") in a major league produces millions of data points that are tracked, summarized and visualized. Heatmaps alone are based on tracking each player's movements on the pitch at any given time, the position of the ball is tracked, passes made, shots etc. All this data is mostly used to entertain us, the fans, and help coaches in their post-match analysis.

Chelsea is now working with scientists on a super-ambitious tpoic: evaluate decision making on the pitch. AI scientists try to determine what would have happened if the player would have made a different decision in any given situation. This is so complex (because it effects the decision subsequently made by others, too) that I am really sceptical about the (at least short term) outcome, but it shows the ambitions and possibilities of AI in sports. At some point in the future, we might have this "scenario analysis" in real time while we are watching football.

TikTok vs. Lasso

If you don't know the names in the headline yet: TikTok is a short video creation app and social network, specialized on lip syncing and visual effects. It is owned by ByteDance from China, a content & technology company valued at around 75 bn USD last August, and absorbed & integrated a service called into TikTok. Musically had/has 8.5 million users in Germany alone, reportedly around 200 million globally.

If this sounds rather hard to believe for you, don't take it from me or some journalist, take it from Facebook: They launched a competitor to TikTok, called Lasso, and this probably not because they aren't successful. Here's a comparison between the two:

Nov 14, 2018

Effect of the native video player on Reddit

Reddit and its impact are underestimated, especially in Europe. In Germany, it ranks 30th among the most popular websites, beating Focus, Welt, ZDF,, and others, and its daily 11:25min time spent even puts it in the top10 (according to stats for Germany) In August 2017, they launched their native video player - and here are some stats:

  • 1 billion "native" video views per month (non-native: embedded video like YouTube)
  • 50% of all video views go through the native player
  • 45% more daily time spent with those who watch embedded video
  • 39% more (on top of those 45) in daily time spent with users who watch native video
  • 70% of video views through mobile apps

More here:

How YouTube recommendations work

This study by Pew Research Center does not analyze the algorithm itself, but derives insights from what is actually shown to users, and the results are interesting. In short: YouTube recommends slightly longer, and progressively more popular (higher view count) videos in relation to the one you started with. And only (depending on age) 15-25% of users say they don't use the recommendation feature - while roughly 2/3 of users in all age brackets use recommended videos occasionally. Also, there's a short tail of recommendations: During >170,000 random five-step walks through YouTube, 72% of recommended videos in this analysis were only shown one time, and 26% of videos were shown 2-10 times, but 2 percent of videos were shown 11-100 times and a tiny number of videos (<1%) were shown over 100 times as recommendations.

More super-interesting findings (51% of >4500 surveyed YouTube users say they go there to learn how to do stuff they haven't done before) and stats here:

19% of NPR's live audio streaming comes from smart speakers (last year: 4%)

Lengthy, but good article on recode with many stats, research and insights about smart speakers and how they are used. It supports my thesis that "voice in, voice out" is just one scenario good for short interactions and information. Voice in, microwaved food out is what Amazon tries with their 60 USD microwave, voice in, video out is what the Echo show tries to do, and voice in, website out is what a Siri search looks like. "Voice in" is probably one of the biggest bets in digital currently, and whoever owns the listening entity, also owns the logic working up any "out" that consumers may adopt. The platform race will be decided between Amazon, Apple, Googe, Alibaba and maybe one or two surprise companies, but anyone who wants to be successful on these platforms needs to understand and serve the logics that are applied behind the voice command. And maybe there will be "voice only" offers that make so much sense with voice, they may make some apps and websites obsolete.

Another source for statistics and numbers - 2017 survey on smart speakers/voice by NPR [PDF]:

Nov 8, 2018

E-Commerce goes offline - true multichannel ahead

Amazon opened another cashier-less Go convenience store in San Francisco last week, and has announced a few more for 2019. While this seems to be only remotely connected to online purchases on their platform, the end-game in retail will probably be a merger of the two - physical outlets that complement the online experience and the other way round. This seems to be Alibaba's strategy with their "new retail" concept, spending more than 10 billion USD since 2016 on the acquisition of physical retail stores. And they expand their own "Hema" supermarkets aggressively, where your can pay by facial recognition and get groceries delivered in 30 minutes if you live in a 3km radius around it. They have also implemented an initiative to digitize convenience stores and small shops while using them as delivery hubs - something that Zalando is now pushing in Europe as well, calling it "connected retail". It seems certain that the future of retail is true multichannel - not just serving a number of platforms and transaction types, but interconnecting them, and it's stunning to see how the online retailers and e-commerce companies are pushing these types of innovation instead of the traditional retailers.

Amazon: Groceries / Go stores

Zalando: Connected retail

Alibaba: New retail

Alibaba: Hema Supermarkets

Nov 1, 2018

Netflix is bigger than Sky in Germany

Finally a few numbers on German PayTV / OTT:
Sky as a traditional Pay TV service had 4.8 million subscribers by the end of June 2018. A recent study by Ampere analysis suggests that Netflix reached 5.1 million subscribers in Germany, and Amazon Prime Video tops the list with 9.9 million subscribers - although it is unclear how many actually use the service as it is bundled with the e-commerce offer. It would be interesting to have numbers on viewing hours, but we're not that far yet.

Hard to believe for me, but based on these 10 million Prime customers maybe reasonable: 54% of German households have two OTT subscriptions, 12% more than four (!).

The big comeback of direct traffic

With search & social as the dominant traffic sources for most publishers, direct traffic declined almost everywhere. But since 2017 - when Facebook changed the algorithm in a "news/media unfriendly" way - referrals crashed. For publishers relying on social traffic this was a bloodbath, but others have experienced a rise in direct traffic. With efforts like video formats (check the NYT for example or Bleacher Report & others) and a big comeback of newslettering (even with millenials and gen Z), direct traffic has become the most important traffic source for many publishers again.

Chartbeat confirms this with an interesting observation (on 4000 news pages) - when Facebook was down for 45 minutes on August 3, 2018:

  • direct website traffic went up 11%
  • direct app traffic went up 22%
  • search referral went up 8%
  • overall traffic to publisher sites went up 2.3%

HIstory & risks of the like button (and ultimately of big tech)

Interesting piece about like buttons as a votig mechanism and how Facebook made it big - and then broke it, relying on algorithms to manage the consequences. Today, very few sites have an embedded like button the way they used to, and it's "share or nothing", while on the platform, we are now dealing with "reactions" that still amplify fake news and are hard to manage. I found the history part more interesting in this one:

A bigger & broader view about the consequences of such features (but also news feeds etc.) in this piece from BuzzFeed News, taking Brazil's terrible election results as an occasion to think about how big tech companies helped to radicalize the world:

Stories are the new news feed

A few years ago, we talked about the three transforming megatrends "mobile, social, video". Now, if you want to visualize these, you just need one picture - a (Snapchat, Instagram, Whatsapp, Messenger, TikTok, Facebook-)Story. Full screen vertical video in snippets rules everywhere. The question is how well a business can be built around this format - that's the key message from the Facebook Q3/2018 results: