Control in theory vs. practice – consumer product research

Every once a while, I observe myself using my iPhone and realize how much control I would want on my phone in theory/if I was asked and how little I need in practice every day.

It speaks to the challenge of getting predictive insight from conversations with users in consumer products where so much of the behavior is subconscious.

By asking questions that draw attention to a particular action or need, we unintentionally move it from the subconscious to the conscious.

And, in doing so, we lose its predictive value.

Schrodinger’s cat.

Solar’s future is insanely cheap

Clean Energy Investor Ramez Naam shared an excellent post on his blog recently titled “Solar’s Future is Insanely Cheap.” He made 4 points in this post –

1. Solar cost dropped 5x over the course of the last decade – that was a whopping 25% of what the International Energy Agency forecast in 2010. It was also 50% less than his own optimistic forecast from 2011

2. To understand this drop, we must understand Wright’s law – the cost of a technology drops exponentially as a function of cumulative scale of production. So, as we produce more of a certain technology, we learn how to better optimize its production.

In solar energy’s case, every doubling of cumulative production has resulted in a 30%-40% decline in prices.

3. Even with very conservative forecasts, it is likely that even medium cost solar plants will be cheaper than the cheapest fossil power plants within a decade.

This doesn’t mean the journey will be straightforward – but, we continue to make progress at a pace that exceeds any previous expectation.

4. Of course, the presence of cheap solar energy isn’t going to be a panacea until we make significant progress on cheap energy storage. We will need to combine solar with the likes of wind, hydro, and nuclear power to de-carbonize places that get little in the way of sunlight. And, we’ll need all the solar we can get as we transition toward Electric vehicles.

But, we’re making a lot of progress and there are plenty of reasons to remain optimistic.

3 books that might change your mind – 2019 edition

Here are 3 books I read this year that changed how I view the world –

1. Alchemy by Rory Sutherland (my review, Amazon): This was the book of the year. It’s impact on me was as follows – every time I hear someone say “that makes sense and should work” or something similar, I stop in my tracks and remind myself that things that the idea that things that make sense should work is a falsehood.

Rory Sutherland’s work has pushed me to be more curious and more accepting of ideas that aren’t logical. After all, logical solutions work better for logical problems vs. psycho-logical problems. Thus, Alchemy has put in a reminder as strong as any that things that work don’t need to make sense and that a dash of alchemy is often what we need to solve problems.

In that sense, its impact on me was profound.

2. The Diet Myth by Tim Spector (my review, Amazon): Tim Spector’s work drilled in a simple lesson – there is no such thing as the perfect diet because it is an interaction between the person’s gut microbes and the food. Everyone reacts to different things differently.

His advice as a result was simple – focus on natural, plant based, foods, Milk and food with living bacteria (yoghurt, cheeses), etc., are recommended. You won’t go wrong with diet that worked for your grandmother. Everything works great in moderation.

There were many other notes in the book that were eye-opening – e.g. his notes on junk food and the research on antibiotics.

Nutrition research is hard to do and books on nutrition are generally riddled with flawed science and incredibly loose correlations (I browsed a book the other day that estimated that eating xg of Beef increased life span by y years based on 2 bar charts). Tim Spector’s book does a great job rising above the pseudoscience with a thoughtful take on the research. It is a great buy and a must read.

3. Ready Player One by Ernest Cline (my review, Amazon): Ready Player One was fascinating because it took me from being a virtual reality non-believer to plausibly consider a future where we might just live in a virtual reality simulation.

It also had a different sort of impact. As of this writing, I’ve reviewed 252 books on my book review blog. The number of science fiction books I read until I read Ready Player One in the beginning of 2019 was 0. And, despite the fact that this was a quiet year in terms of books for me (hence the 3 books recommended versus the usual 5), I read 4 works of science fiction and enjoyed reading all of them.

Thank you, Ernest Cline.


More resources:

Past editions of “5 books that might change your mind” – 2018, 201720162015201420122011

Categorized book recommendations of all 230+ books (mostly non-fiction) I’ve read in the past decade – RohanRajiv.blog

Personal IPOs and stock market tickers

Eugene Wei had a post on his excellent blog yesterday where he drew parallels between individuals building brands on social media and a company going public.


One way to understand the impact of these public social networks on humanity is to think of this as the era in which humans took their personal thoughts and lives public at scale. Billions of humans IPO’d, whether we were ready for it or not, explaining why the concept of a personal “brand” became such a pervasive metaphor.

In another era, most of us lived in social circles of limited scope. Family, school, coworkers, neighbors. We were, for the most part, private entities. Social media companies quickly hit on the ideal configuration for rapid network growth: take the interaction between any two people and make it public. Conversation and information-sharing became a democratic form of performance art.

One reason social networks quickly converged on this as the optimal strategy and configuration is that the majority of people on any social network merely lurk. By making the conversations of the more extroverted, productive nodes public, you sustain the interest of that silent majority of observers with what is effectively crowd-sourced (read: free) content. The concept of 1/9/90 is that a stable equilibrium can be achieved in a large network if the shouting class, the minority which entertains the much larger but silent majority, is given enough quantifiable doses of affirmation (likes) to keep the content spigot flowing. As these large public social networks grew, even many who were previously modest began taking the stage on social media to karaoke to the crowd. Live fast, die young, and leave a viral post.

Just as there are many advantages to being a public company, becoming a public figure carries all sorts of upside. Once your ideas and your self are traded publicly, anyone can invest and drive the value of those goods higher. If you’ve ever written a viral blog post or tweetstorm and gained thousands of followers, if you’ve had a YouTube video picked up by traditional media and found yourselves interviewed on the local news, you’ve felt that rush of being a soaring stock. Social networks not only provide public liquidity for anything you care to share on them, but they also continued to tweak their algorithms to accelerate the virality quotient of their feeds. In a previous generation, Warhol quipped the duration of sudden fame was 15 minutes, but social media has made that the time it takes to become famous.

The problem is that, like many private companies who find the scrutiny of public markets overly stringent, many of us were ill-equipped for “going public” with what were once private conversations and thought. It’s not just those who made enormous public gaffes and got “canceled.” Most people by now have experienced the random attack from a troll, the distributed judgment of the public at large, and have realized the cost of living our lives in public. Most celebrities learn this lesson very early on, most companies put their public-facing executives through PR training, but most humans never grew up under the watchful gaze of hundreds of millions of eyes of Sauron.


It is a powerful analogy and it made me wonder about Instagram’s recent decision to remove “like” counts. If an Instagram profile is the equivalent of a stock ticker, a “like” may have been the equivalent of the stock price.

And, as there’s only so much good that can come from constantly monitoring short term fluctuations on your stock price, removing “like” counts may turn out to be a master stroke.

The Walkman decision

There’s a great story about a decision Sony made when they shipped the legendary Walkman.

Against the advice of market research, Sony’s co-founder had asked the engineering team to build a portable music player that would ease the boredom on long flights. The engineers then came back with what could only be termed a product manager’s dream – for nearly the same amount of effort and cost, they were able to add an additional feature – a record button – to this cassette player.

But, to their dismay, Chairman Akio Morita asked them to remove the record button.

By reducing the device to serve a single use case, he eliminated any potential user confusion. In the same way McDonalds removed cutlery from restaurants to make it clear how they wanted customers to eat their burgers, Sony released the Walkman with a lower range of functionality to give them the highest chance to change customer behavior.

And change behavior they did.

(H/T: Alchemy by Rory Sutherland)

The story of Sandwich

John Montagu was a consummate card player who didn’t like meal interruptions while playing his favorite game of cribbage. So, it is said that he asked for veal stuffed between two pieces of bread to make it easy for him to eat while playing.

As John Montagu was also the Earl of Sandwich, others started asking for the “same as Sandwich.” And the rest, as they say, is history.

For when we find ourselves stuck in discussions about building for the “average” user, it is worth reminding ourselves that the sandwich, like many innovations, happened on the edges thanks to a passionate early adopter with a weird request.

(H/T: Alchemy by Rory Sutherland – a fantastic read)

Skyspring Hotel New York

Someone we know in India received an job offer from an acquaintance in the Middle East to work at a Skyspring hotel in New York. They needed to front $600 for the visa and he’d received instructions via an email.

They come from a modest background (his mother is a cook) and she asked my mother to help check if this was a real job – $600 is a lot of money after all. The “pay upfront to get this job” rang all kinds of alarm bells. But, it was hard to dismiss it outright since it came from someone they knew.

And, it didn’t help that Google had a very convincing looking card show up on search.

Of course, it all unravels the moment you spend more than a minute investigating. The hotel has no trace on Tripadvisor or Booking.com and the phone number doesn’t work. The email has a few typos (why do scammers not get that right?), was sketchy on details of the work visa, and it came from a questionable looking “@consultant.com” email address.

All in all, it was more sophisticated than the traditional Nigerian prince scam and it could have fooled someone who wasn’t discerning. I think the Google card was the most convincing piece of the scam and I couldn’t find a way to flag it on my phone (I found it on my laptop and did so).

It did get me thinking about how important it is to design products with scam/bad actor use cases in mind. It isn’t enough to just think of the happy path.

Stopping shoplifting

Bloomberg shared the story of the company behind the product that claims to be able to detect shoplifters by monitoring fidgeting, restlessness, and other suspicious body language. The goal is prevention – if the person is approached, the chances are high that the crime never happens.

On the one hand, this is awesome. If we can use technology to stop folks from committing crimes, that is a win.

On the other hand, it does make me wonder where this road will take us.

For example, will the data about the identified shoplifters go to a centralized database? Will that database be shared with other retailers to stop crime together? Will law enforcement make a case that the data should be shared with them? Will we then use the data in the database to move beyond behavioral signals to demographic signals?

It isn’t hard to envision why these steps wouldn’t logically follow the first. What happens to someone who makes a bad decision to steal a loaf of bread because he’s going through a tough time? Given how quickly he will be identified and caught, how hard will it be for him to pick himself back up after he commits that first crime?

Many questions. No simple answers.

Collective notification log

NYT reporter Katie Rosman shared a screenshot she found of a teacher who had her students turn up their phone volumes in class and create a collective record of notifications they received.

I wonder what this chart would look like if we, as co-workers and family members, did this exercise at important meetings and family meals.

And, perhaps more importantly, what if we made it a point to do it periodically?

(H/T: Greg’s “Cofounder Weekly” newsletter that brings together a collection of interesting/fun tweets on tech/entrepreneurship)