Weekly Digest #4 [Scuttlebutt, AI, Personality, Privacy, WeChat & Big Data]

A snapshot of article summaries that we enjoyed reading ( Nov 26— Dec 1)

“Science may provide the most useful way to organize empirical, reproducible data, but its power to do so is predicated on its inability to grasp the most central aspects of human life: hope, fear, love, hate, beauty, envy, honor, weakness, striving, suffering, virtue.”
— Paul Kalanithi (When Breath Becomes Air)

Scuttlebutt Investing: The term coined by Philip Fisher refers to rumors or gossip and is derived from the occupation of sailing. It involves completely immersing yourself in a subject/industry/company for some length of time to play out every scenario and run down every question till you become a master on it. A common criticism of scuttlebutt method is that it is highly anecdotal. An actual practitioner of this method requires to combine both anecdotal and empirical methodology. It requires in-depth analysis of financial statements with a high quality of due diligence and then talking to people with an educated mindset on the topic. It is also important to look at information that disproves the hypothesis about the company. Interaction with people on the other side of the fence and understanding their point of view. Making a rational impartial judgment of the company based on all facts — good and bad. Talking to people who are not in our circle of family and friends develops a broader worldview in order to get a realistic picture of a company based on different perspectives. Empathizing with the end customer of the company also helps in the process. The objective of scuttlebutt is not to get inside information but get information which cannot be found through reading. Think broadly about the people who can be helpful and engage them. Keeping a vow of confidentiality also helps in understanding the actual color of the company. Scuttlebutt research provides : 1. Deeper insights into the company which cannot be obtained through reading (secondary research) 2. Primary research which is important to build the courage and conviction to hold when the market moves in the opposite direction.

Artificial Intelligence & Machine Learning: There is confusion surrounding artificial intelligence(AI) and machine learning(ML) as some people use the terms interchangeably while some refer to them as separate parallel technologies. Machine learning is a subset of AI. ML relies on defining behavioral rules by examining and comparing large data sets to find common patterns which can solve classification problems. When an algorithm compares with previously labeled human data to find common patterns, it is called supervised learning. Another type of ML is unsupervised learning which provides the algorithm with unlabeled data and lets it find patterns by itself. Reinforcement learning is a type of ML algorithm which has a set of rules and constraints and it learns by itself on how to achieve the goals. Similarly, artificial intelligence (AI) is the science and engineering of making computers behave in ways that, until recently, we thought required human intelligence. AI today is symbolized by Alexa and Siri, movie recommendations by Netflix and algorithms of hedge funds. These technologies are augmented intelligence technologies that enhance our abilities and make us more productive. Historically, the AI industry went through many ups and downs as there was great potential in the technology during the early decades. But undelivered promises caused a general disenchantment with the usage of the term AI. After 2012, machine learning, deep learning, and neural networks did great progress and were adopted in a number of fields. Deep learning fields of natural language processing, etc. took great leaps. For those struck with limitations of software, these developments seemed magic. A wrong description of neural networks mimicking the human brain, warning of an apocalyptic future by Elon Musk and fear of technological unemployment have reignited hype and excitement around AI. In order to live up to this hype, people use both the terms of AI and ML interchangeably however advanced AI claimed by these companies is usually a variant of ML or some other technology.

Personality — Dark Core: Personality and intelligence are fundamental constructs in psychology that influence our behavior. Spearman made two discoveries about human intelligence: 1. People who score high on one test of intelligence also tend to score high on other tests of intelligence. 2. You can measure intelligence with enough reliability and validity as long as tests have cognitive complexity. A new research suggests that these principles not only apply to human cognitive abilities but also to human malevolence. People who display ethically, morally and socially questionable behavior in their lives possess “dark traits”. The dark traits (d-factor) is the basic tendency to maximize one’s own utility at the expense of others, accompanied by beliefs that serve as justifications for one’s malevolent behaviors. Utility doesn’t refer to utility maximization which is independent of others. The proposed d factor is comprised of 9 traits: narcissism, Machiavellianism, moral disengagement, narcissism, psychopathy, psychological entitlement, sadism, self-interest, and spitefulness. If someone scores high on one trait, one can score high on other traits. The key prediction being those who score high on d factor will not be motivated to increase utility of others without benefiting themselves, and will not derive utility for themselves from utility of others. All dark traits were substantially correlated with each other. Those scoring high on the D factor were likely to display unethical behavior and more likely to keep money for themselves. The D factor had positive associations with self-centeredness, dominance, impulsivity, insensitivity, power, and aggression. It had negative associations with nurturance, perspective taking, internal moral identity, sincerity, fairness, greed-avoidance, and modesty. Thus, the D factor of personality has emerged to be an interesting conversation in research circles.

Privacy Paradox: A need for more privacy while consuming more data is the core of this paradox. How we wish to protect our privacy when that protection itself is counterproductive to the efficiency of the platform. Platforms thrive on more data to produce better results and privacy regulations act as a valve to control data flow. The Cambridge Analytics scandal highlighted deep issues on safeguarding privacy. However, there has been no clamor by consumers to protect their privacy. The EU has a strong view of penalizing the platforms with antitrust regulations to solve this problem. An MIT research suggests that it is not a good idea. Privacy regulations impact smaller firms that need access to data in a much harsher manner than the bigger firms due to transaction costs. Privacy regulation requires firms to obtain one-time permission from consumers. As the impact of transaction costs falls largely on smaller firms, the inherent nature of privacy regulations suggests that they are anti-competitive as per the study. This reveals two paradoxes with respect to the relationship between data and privacy. The first paradox being — as we desire to get more privacy protection, we increasingly use more services that utilize more of our data. The second paradox being information asymmetry. Data user might not know what data subject wants or knows and vice versa. As a result, the privacy trade-offs are temporary. Disclosure of data might provide immediate benefit but cause harm later. Based on personal data, many applications infer your social activities to sell you ads for your benefit. Strong regulations on privacy may cause consumers to be offered inferior products and services or even potentially damaging ones. Hence dealing with privacy becomes complicated as it involves a mixture of collective behavior, psychology, marketing, and computer science. The Privacy Paradox stems from the law of unintended consequences. These unintended consequences are large enough to realize there is a lot of work and thought required in this area.

The rise of WeChat: Unlike AOL and Yahoo, Tencent is one of the few companies to have evolved its messaging product portfolio from a web/desktop product to a mobile messaging platform. This led to the creation of a new category of “messaging as a platform”. Lesson 1: Build your own competition: Tencent started WeChat as a messaging and photo sharing app. It gradually moved to voice and became the first app to offer voice, text, and photo in a combined app. Followed the principle — if you don’t disrupt yourself, someone else will. Lesson 2: Design for groups. An individual user’s behavior is different from how they behave in groups. To identify this, WeChat observed how users behaved among groups of friends and strangers in everyday life. These features (People Nearby) boosted new user acquisition. Lesson 3: Extend features from users’ inner desires. Attention to communication nuances & cultural behaviors to enhance the quality of user experience. Photo sharing in private friend circles and personalized conversation with celebrity accounts were features that fit with users. Lesson 4: Big ideas come from solving your own problems. Founders should be an avid user of their product and solve their own problems. WeChat Red Packets started from Tencent’s practice of every manager providing the staff a red envelope with some cash after Chinese new year. WeChat Red Packets sent in groups facilitated the growth of WeChat Pay in closely knit social circles — friends and family who connected their bank accounts based on trust. Strategic partnership with Didi opened a broad universe of payments from phone bills to utility payments. Lesson 5: Monetize subtly. Monetization and user growth are not mutually exclusive and WeChat used monetization as a lever to improve product experience. Limiting native ads, effective ad placing, social gaming and adoption of a coupon sharing model helped in user growth. Lesson 6: Measure what you value and not what you are supposed to. Unlike social products, WeChat does not only measure growth by a number of users or messages sent but it also focuses on measuring how deeply the product is engaged in every aspect of daily life. Lesson 7: Don’t play favorites with features. Most product managers focus on building sticky features and drive retention metrics. WeChat as a product focuses on utility oriented qualities like ease of use and high functionality. It limits the marketing content in the app. A clean UI with 4 tabs despite a wide range of features and services is a key feature of the app. The WeChat team built product features that answered cultural needs and emphasized group interactions. A simple tool that could extend in the hands of every mobile user. WeChat platform has completely revolutionized how the Chinese communicate and socialize online, how they pay each other making it a successful runaway startup success.

Big Data — Do’s & Dont’s: We extract value from Big Data and apply it to every aspect of a business. However, some key points to note. First is ‘Not to leave the exercise to data scientists alone’. The days when technology could be left to the chief information officer are long gone. Maybe, in the next iteration of AI, analytics will be able to tell what to sell, who to hire, how to market to customers and even how to manage the business — but they still won’t be able to tell what business to be in and why. Top leadership has to stay on top of what the data is telling them, lead the analytics team in addressing real business concerns, and work constantly to make sure all the data is being targeted at intelligent investment. Second thing to avoid is ‘Getting Lost in Translation’. Companies need business analysts who can read information and see how to use it to spot market opportunities, identify problems and come up with solutions. They need business ‘translators’ who may not be full-fledged data scientists but are sufficiently proficient in analytics to take numbers & apply them for the benefit of the business. The third point is to ‘Avoid Drowning in a Sea of Data’. There is a natural urge to want to capture every atom of the business’s legacy data, then cast around wondering what to do it. Instead of trying to capture all historical information, decide the business priorities, identify what data is likely to be useful in addressing it, and add to it gradually. The fourth aspect is ‘Knowing Where to Start’. One must begin by identifying the most promising sources of value to the business. The next step is to identify as many use cases as possible and look at applying new data and techniques to generate new insights. The fifth point advises on ‘Democratizing the Data’. One of the most common reasons for lack of uptake of data analytics is that the people who can put it to best use lack meaningful access to it. Companies must make data accessible to as many as possible, work on consensus for data validation to build an agreed source of ‘truth’ for the business and develop an egalitarian culture whereby everyone is allowed to ‘play’ with the data without fear or favor and try to generate new ideas. Sixth point talks of ‘Being Ready to Trade Secrets’. Traditionally, companies of all shapes and sizes guarded all market and business information closely as a way to retain know-how and competitive edge. This is no longer necessary as there are plenty of industries where data sharing, for example as part of an industry bloc, increases the comprehensiveness of the data and enables individual businesses, who may operate in different segments of the market, to enhance their offerings and create greater value. The seventh point deals with ‘Cultivating a Culture’. To get the most from the data in terms of generating ideas that will lead to new innovations, one needs to foster a test-and-learn culture: leaders set out the vision, employees are encouraged to identify opportunities and develop proofs of concepts and use data to analyze the results. This is a learning process, so it also means a no-blame culture. The eighth point talks about ‘Not Being Afraid to Kill Your Darling’. No matter how attached one is to an idea of what the next big product launch should look like, one should be prepared to walk away as soon as the data tells it otherwise. Data can do many things but — if you’ve got the analytics right — it can’t lie. The Ninth and last point talks about ‘Not Doubting the Doubters’. Data Analyst’s mindset is relentlessly self-critical and questioning. This is at complete odds with the characteristics required of a leader or manager who needs to exude confidence. The trick is to balance the two; to recognize that nothing is guaranteed in future events but still be able to motivate people to implement the decided strategy. One needs to see the world bottom-up, from the perspective of the data analyst, but communicate top-down to drive the process of change.

Follow us on twitter @GlanceThrough for daily tweet summaries of articles we find interesting

The link to the original articles is hyperlinked in the title of each summary.