Follow by Email

Search This Blog

Jul 16, 2018

Content Creation based on Data: Amazon vs Netflix

This article is a few weeks old, and at its core is a great Ted Talk from 2016 that compares how Amazon and Netflix used data to create content. The talk is just 12 min long and if you add 5 min for that article (with a lot of newer info), your time is well spent. In a nutshell, Amazon based their decisions on actual reactions of viewers who saw pilots (which they produced after holding a competition), while Netflix did not use (limited) reactive data to existing pieces of content, but reactions & insights about any content watched on Netflix - and created theirs from there. Amazon's result was "Alpha House", Netflix's was "House of Cards".

Article on Forbes incl. Sebastian Wernicke's Ted Talk:

Talk only:

Hands On with the BBC News Lab

At the end of the article, they reveal that they are hiring, and it's a smart way to maybe attract developers: show them what is actually done at BBC News Lab. They describe their work as being divided into two categories: new tools to help journalists produce better content, and new news formats for BBC audiences. Interesting insights:

Jul 10, 2018

Meet, the Chinese E-Commerce Giant

When I started monitoring & including WeChat in talks a few years back, I came across the "Kingmaker Strategy": Buy a significant share in a fast growing (potential) leader in a segment, and leverage that investment by (sometimes exclusively) integrating that company in your own service. That's what Tencent did with (wouldn't integrate Alibaba as they have that big rivalry) on WeChat, and now that company has grown to a 65bn USD valuation. This is an interesting article on Wired about how they use automation and robotics, and what their approach to ecommerce is. 

FC Bayern on Social Media 2017/2018

A few numbers and stats on FC Bayern during the last season, where they - in my opinion - upped their social media game significantly. The top post achieved close to 600,000 likes on Instagram alone, and it was - cat content. A cat selfie in front of the team. Unthinkable a few years back, but FC Bayern understood that stuff like this does not damage the brand, but rather increases relevance in target groups that you probably won't get on TV anymore (also because in most markets, Champions League football is on Pay TV anyway - and you'll need to provide some kind of exposure for sponsors, and make new fans on new channels). Some interesting numbers:

Jul 9, 2018

Health & fitness apps success factors

Great overview by Stanwood's Hannes Kleist of the German market in fitness and health apps - which are the genres, which monetization models work, how a paywall or freemium model should be implemented, and how to find the right pricing, based on data. Read this:

[fun post] Bizarre correlations & statistics

If you want to make fun of statistics that don't entirely convince you, this is a great source - it shows bizarre correlations in data with nearly identical curves, like "US spending on science" and "Suicides by hanging and suffocation", or between "US crude oil imports from Norway" and "Drivers killed in collision with railway train". Not all data analysis that brings up statistically relevant / significant correlations makes sense, obviously.

The Kindle edition of the book where these are from is just 7 bucks (German store):

AI acquisitons 2013-2017

Great overview by CBSinsights that shows how we are in an "acquisition binge" for AI startups - and nine companies are buying the majority of those. Alphabet/Google, Facebook, Microsoft, Apple, Amazon, Intel, Twitter, Salesforce - and a monitoring form called Meltwater snap up most of them. In this article you will find a list of their acquisitions 2013-2017, and a number of charts on all other major M&A activities in this field. Great for a quick overview.

Alibaba has an AI copywriting tool

Partners sell products on Alibaba's sites, so they upload images, product descriptions and advertisements like banners or 20-sec-video for marketing. In order to make that easier and more efficient, just like any other site, Alibaba offers tools (compare Facebook's ad planner). But Alibaba now claims they have a deep learnig/natural language processing tool that can write 20,000 lines of copy per second - and pass the turing test. You can even choose if your copy should be promotional, fun, heartwarming, functional or "poetic". Offered in Chinese only (currently), this shows us where we're headed - additionally to this, they have tools that automate banner-resizing to different formats and AI-powered video tools to easily generate 20 second ads for their platform. So the direction is: basic creative idea - human. Everything else: machine.

Is IGTV an attack on YouTube?

I am a firm believer in vertical video - basically in every format. From Snapchat over WhatsApp Status to Instagram Stories, short form works exceptionally good, and IGTV now opens that format for long form video, too. This article on Hacker Noon makes the case that in a few years' time, IGTV may do to YouTube what it made with Millenials to TV: to become a potentially dispensable part of their lives. Very interesting though that what is missing - and should be added by Instagram - is a smart search feature as the structure of content being tied to accounts that you follow may cover a "social" way of distributing content, but leaves the whole area of intentional viewing (YouTube's stronghold) mostly open. Will be interesting to see the further development here.

Ben Evans about Machine Learning

If an article about machine learning begins with dismissing sentences like "data is the new oil", I am interested. Benedict Evans, Partner at Andreessen Horowitz, offers his perspective, and as always, it's valuable. You should read it, but if I try to summarize, Ben makes a case that ML is currently a disruptor to what we used to call "data mining". While data analysis is the goal-oriented examination of data, data mining describes the analysis of data sets without asking a question - trying to find correlations, anomalies and potentially significant insights in data. So ML can help in both areas, give you better anwers to questions you already have, and deliver answers to questions that you didn't ask. And it can broaden your data set as ML makes it possible to include new data types in an analysis, like audio, video or images.
Worth the time to read through it:

Teens, Social Media & Technology 2018

Pew Research Center with stats & numbers from the US market on generation Z. 95% of US teens have a smartphone, and 45% say they are online "almost constantly". That's coming from 24% in 2015 and is a dramatic increase. In platforms used overall YouTube leads (72%), followed by Instagram and Snapchat, while "most used platform" is led by Snapchat. Nothing that you wouldn't know if you have an n=2 group of teens in your personal environment, but it helps to see these impressions verified on scale by research.

Twitter's turnaround

I recently made a quick research for a client about news apps and news aggregators, and I was surprised to find Twitter in the "news" section of app stores instead of social networking. At a time where Facebook positions itself away from news, Twitter seems to embrace that opening (although the article below says it's a coincidence). "Live" has always been boon and bane for Twitter - an algorithm clearly delivers more relevance than a purely time-based newsfeed, on the other hand these obstruct any "live"-news mass distribution as we can clearly see with the World Cup (where I get "goal" posts a day later on Facebook or Instagram). BuzzFeed News wrote an interesting overview about Twitter's comeback:

Aibaba vs Tencent

Or "Ma vs. Ma". Comprehensive piece from Forbes about the two Chinese giants, how they compete with each other and how they differ from their American peers. It includes some interesting info about the domestic Chinese market and a few great charts/stats, about their operations and investment strategies. And personally I wasn't aware of Alibaba's "integrated retail program". where even 7-11-sized shops and bodegas get modern inventory software, sensors that monitor foot traffic and camera generated heatmaps to optimize their physical stores. Great read:

Jul 7, 2018

Facebook sends less and less traffic

Lengthy article from Slate - including some very interesting numbers and stats from their own publication - about how Facebook's change in treating news, concentrating on a "trust ranking" among publishers and emphasizing meaningful interactions has impacted their business. Their traffic from Facebook dropped 87% from 2017 to 2018. They come to the conclusion that Facebook "retreats from the news business".

Full article:

Digiday's take (with less statistics):