Swipe to Unlock
01 Oct 2022| Book | Swipe to Unlock: The Primer on Technology and Business Strategy |
| Author | Parth Detroja, Aditya Agashe, Neel Mehta |
| Published | September 19, 2017 |
Great quote on agency
When you grow up, you tend to get told that the world is the way it is and your life is just to live your life inside the world, try not to bash into the walls too much, try to have a nice family, have fun, save a little money.
That’s a very limited life. Life can be much broader, once you discover one simple fact, and that is that everything around you that you call life was made up by people that were no smarter than you. And you can change it, you can influence it, you can build your own things that other people will use… That’s maybe the most important thing. It’s to shake off this erroneous notion that life is there and you’re just gonna live in it, versus embrace it, change it, improve it, make your mark upon it…
Once you learn that, you’ll never be the same again.
— Steve Jobs
Data science/ AI features (like recommendation systems) can create both differientiation and “stickiness”. Also data flywheels.
First, a great recommendation system is a selling point, helping Spotify stand out from rivals like Apple Music. That’s because just having a huge music library isn’t enough. In business terms, music is a commodity — any song sounds more or less the same whether it’s on Spotify or Apple Music or anything else — and anyone with enough money can get the licenses to build a giant library. So, if all the music streaming services can have effectively the same selection of music, Spotify needs something to differentiate itself from competitors. And Spotify’s recommendation system certainly fits the bill — it’s widely considered better than Apple Music’s. And since collaborative filtering gets better with more users, Spotify (which already has tons of users) can continue furthering its lead.
The second reason is that personalized recommendations make users more likely to stick with the service. The more you use Spotify, the more the algorithms know about your tastes, and hence the better they can recommend music to you. So if you use Spotify a lot, your recommendations would be pretty good, and you’d lose a lot by jumping to Apple Music, which doesn’t know you at all. So this high “switching cost” makes you less likely to jump. (More generally, any personal data you put into an app — like making Spotify playlists — raises the switching cost, since you’d have to recreate that data in any new app you jumped to.)
Three kinds of APIs: feature, data and hardware (although this doesn’t seem exhaustive)
Three kinds of APIs The first kind of API, which we’ll call “feature APIs,” lets one app ask another specialized app to solve a particular problem, like calculating driving directions, sending text messages, or translating sentences.
The second type of API, which we’ll call “data APIs,” lets one app ask another app to hand over some interesting information, such as sports scores, recent Tweets, or today’s weather.
The final kind of API, “hardware APIs,” lets developers access features of the device itself. Instagram taps into your phone’s Camera API to zoom, focus, and snap photos. Google Maps itself uses your phone’s Geolocation API to figure out where in the world you are. Your phone even has sensors called accelerometers and gyroscopes, which fitness apps use to determine which direction you’re walking and how fast you’re moving.
Gather data and test/ experiment, rather “debate endlessly”
Not sure which marketing catchphrase will get people to buy? Instead of endless debating, just run an A/B test! Not sure if a red or green “sign up” button will get more people to click? Run a test! (If you’re curious, the red button got 34% more clicks in one experiment.) Not sure which Tinder profile picture will get you the most swipes? Tinder even lets you run A/B tests to figure out which of your photos, when shown as your main profile picture, gets you the most right swipes.
The “consumerization of the enterprise” trend
What BlackBerry failed to realize, though, is that people really enjoyed using their iPhones, with their bright colors and touchscreens. And instead of selling phones to corporate IT managers, as BlackBerry did, Apple sold iPhones directly to consumers — that is, average people like you or us.
The result? With greater access to iPhones, people started carrying two phones: BlackBerries for work but iPhones for personal use. Soon, businesses realized they could save money and keep employees happier by just letting employees use their personal phones for work purposes. Slowly but surely, iPhones started creeping into BlackBerry’s treasured enterprise market — a perfect example of the trend known as “consumerization of the enterprise.” BlackBerry realized that for smartphones, everyday users, not businesspeople, called the shots.
Flywheels can work both ways: both positive and negative reinforcing cycles
As BlackBerry sank into third place behind iPhone and Android, it got stuck in a vicious cycle called the “chicken-and-egg problem.” Developers wouldn’t build apps for the BlackBerry platform if it didn’t have any users, and users wouldn’t buy BlackBerries unless there were enough apps. Imagine if no one would walk into a party unless there were already enough people there — nobody would ever come in. BlackBerry tried hard to entice developers to their platform, even offering $10,000 to anyone who made a BlackBerry app in 2012. But it still didn’t work.
Example of a positive reinforcing cycle/ flywheel. Also, the power of free in consumer to jumpstart flywheels.
Google’s strategy begins with getting as many people as possible to use Android. Clearly, making Android free is working: Android powers over 80% of the world’s smartphones.
Google forces any phone manufacturer that uses Android to install all the core Google apps, like YouTube and Google Maps, by default. In the US, Google even forces phone manufacturers to put the Google search bar no more than one swipe away from the home screen. By getting more people on its apps, Google can get more data, show more ads, and thus make more money.
Multiple reinforcing forces are even better
A second source of income — which is smaller but still considerable — is app purchases. In most countries, Google insists that phone manufacturers prominently feature Google Play, Android’s app store, on the home screen of any Android phone.
Third, making Android more prevalent could help Google keep more ad revenue for itself. Whenever an iOS user clicks on ads in Google Search, Apple keeps a sizable chunk of the ad revenue that would otherwise go to Google. Plus, Google pays Apple an estimated $12 billion per year just to make Google Search the default search engine on iOS. So you can see why Google would prefer if people ran Google searches on Androids instead of iPhones.
KaiOS case study: going down market
KaiOS realized that, while Firefox OS’s business strategy was flawed, its technical underpinnings were sound. Since Firefox OS was open source, KaiOS picked up the old code and created a new operating system that was still web-based but now worked for touchscreen-less feature phones. And instead of trying to fight for Android’s slice of the pie, KaiOS was expanding the pie, going for the market — feature phones — that Android would never get into. (While low-end Androids might be cheap, they still can’t match the Jio Phone.)
KaiOS’s second smart move was realizing that the people demanded apps like WhatsApp and YouTube. So it partnered with Google to create custom-built versions of apps like Google Search, Google Maps, YouTube, and the Google Assistant for KaiOS — these apps were presumably had a better user experience than just loading the mobile website. The Jio Phone grew like crazy, selling 40 million units in its first year and a half on the market. And KaiOS has done great too, with over 85 million phones sold in over 100 countries.
The importance of power users/ “whales” in freemium strategy
Freemium apps usually use one of two strategies to make money: in-app purchases or paid subscriptions.
As you’ve probably experienced, most people are willing to pay exactly zero dollars for software: one study found that only 6% of all iOS app downloads were for paid apps. Yes, even $1 is too much for most people. But while most people won’t pay anything, the small fraction of people who use a certain app the most are often willing to pay a lot of money. Economists call this the 80-20 rule or the Pareto principle: 20% of your customers will generate 80% of your revenue, and 80% of your customers will generate the other 20%.
The key for app developers is to find that 20% of people who want to pay money (called “whales” in the industry, probably because they’re rare but huge) and squeeze as much cash as possible out of them. These whales are big: the average paying user of a mobile game spends $86.50 on in-app purchases every year. Some whales are of positively Moby Dickian proportions: in 2015, the mobile title Game of War: Fire Age earned almost $550 per paying user.
In short, the freemium strategy is this: give away the app for free to draw in a massive number of users, find the “power users” who love the app, and charge them — either once or via a recurring subscription — for extra features.
Ads = selling your attention. Can lead to surprisingly large businesses
First, apps and websites can charge advertisers a small fee every time someone views an ad, a strategy called Pay-Per-Impression, or PPI.. Since so many people view ads, apps and websites usually charge in increments of 1,000 views; that is, the pricing for an ad campaign could be $5 for 1,000 “impressions.” Because advertisers often pay per thousand views, Pay-Per-Impression is more often called Cost-Per-Mille, or CPM. (Mille comes from the prefix milli, as in millimeter.) Alternatively, apps and websites can charge advertisers whenever someone actually clicks an ad, which is called Cost-Per-Click, or CPC. CPC is less frequently known as PPC, or Pay-Per-Click.
In short, ad-driven companies — for the most part — aren’t selling your data. As PCWorld put it, it’s more accurate to say that they’re selling you. And this strategy is so successful that it can build billion-dollar companies that charge users nothing. That’s the remarkable part of the app economy.
Rise of content as marketing/ advertising
When you think of ads, you might think of banner ads: those flashy, animated rectangles that show up on webpages or at the bottom of apps. They’re still popular on websites, although fewer and fewer major apps have banner ads nowadays, probably because ads are annoying and take up valuable space. Plus, people rarely click on banner ads anymore, so they aren’t very profitable. In fact, people only intentionally click on banner ads 0.17% of the time — that’s roughly 1 out of every 600 ads you see.
Sponsored content, also called “native advertising,” means ads that blend in with normal content, making viewers more likely to take the ads seriously instead of ignoring them. Sponsored content is growing especially quickly in the world of journalism. Advertisers can pay to include legitimate-looking articles (which are really just ads) amidst normal material on websites like the New York Times, CNN, NBC, and the Wall Street Journal. Newer media companies like BuzzFeed also love native advertising.
An increasing amount of “journalism” is becoming dressed-up advertising. For example, the New York Times once ran a story about why the traditional jail system doesn’t work for female inmates. It was a well-researched and engaging story, but it was all an advertisement for the Netflix series Orange is the New Black.
Platform/ marketplace companies
Amazon, Uber, and Airbnb are all free to download, never charge you for using the app, and show you few, if any, ads. So how do they make money? “Marketplace” or “platform” apps like these, which connect buyers to sellers (or riders to drivers, etc.), earn commission by sneaking some fees into the purchases you make. It’s like how the government makes money using a sales tax or how a real estate agent charges commission whenever they help you buy or sell a house.
The travel-booking service Wanderu, for instance, helps you find the best bus tickets and refers you to the websites of bus lines like Greyhound and Megabus to purchase them. Wanderu doesn’t charge shoppers anything, but it takes a small commission from the bus lines for the privilege of sending customers their way.
To go back to the analogy, cloud computing is like Uber for computers. Instead of owning your own car or computer, you can get your files or transportation on-demand from anywhere with an internet connection.
SaaS and susbcription businesses
Moving to a subscription model for Photoshop (also known as a software-as-a-service model) proved to be a smart business move for Adobe. For one, they could earn more consistent revenue, since they started getting subscription fees every month instead of having to wait for a huge new release, which only happens once every few years. It also helped Adobe fight piracy, since the monthly license check means that Adobe gets to decide who can and can’t use the software. And third, since Photoshop now regularly connects to the internet, Adobe can constantly push updates and bug fixes instead of having to wait until the next big version is ready. This keeps customers happy and helps squash security issues faster. (This development model is called “agile development.”)
Adobe’s revenue jumped 70% within a year. Why? For one, a subscription lets you get continuous updates for no extra cost. Second, it makes Photoshop more accessible for new users. You now get a month-long free trial, and the first year costs $240, compared to $700 to buy the last boxed version. Third, Creative Cloud lets you store your Photoshop files in the cloud for no extra cost, making it easy to edit on any of your devices. So despite the original controversy, moving to a subscription service was huge for Adobe: it helped double Adobe’s stock price and boost revenue 70% in just one year.
SaaS vs IaaS vs PaaS
There’s a third kind of cloud service that falls somewhere between IaaS and SaaS: platform-as-a-service (PaaS, pronounced like pass). These platforms often include extra useful features like databases, advanced analytics, and entire operating systems. Basically, PaaS makes it even easier for developers to build websites in the cloud. PaaS examples aren’t as famous, but one is Heroku, a service that lets you just send in your app’s code and then automatically sets up the website, with minimal setup required. (AWS, which is IaaS, also makes it easy to set up a website, but PaaS makes it even easier.)
What’s the difference between SaaS, IaaS, and PaaS? Let’s use an analogy: food. SaaS is like a restaurant: you just tell the waiter what food you want, and they’ll bring it to you. IaaS is like renting a kitchen: you have the space, but you have to bring your own ingredients and utensils and cook the food yourself. PaaS sits in between SaaS and IaaS: you give someone your ingredients and recipe, and they’ll prepare the food for you.
“Big data” trend
We humans generate a mind-boggling amount of data. As Google cofounder Eric Schmidt put it in 2010, “every two days now we create as much information as we did from the dawn of civilization up until 2003.” That’s five exabytes, or five trillion GB, of data every two days. That’s like if every person on Earth filled up a 512GB iPhone every single day. (And mind you, this quote was made way back in 2010!) This is an enormous amount of data. It’s huge, colossal, titanic. Or, as technologists call it, “big.” Companies are using big data to reinvent technology and themselves, to the point where one analyst said that “information is the oil of the 21st century.”
The power/ potential of data/ AI solutions
Target identified a group of roughly 25 buying habits that, when analyzed together, would let the company assign each shopper a “pregnancy prediction” score. With this “predictive analytics” solution, Target can predict pregnancy with 87% confidence — and sometimes, they can even predict the approximate date of delivery. As the example from Minnesota shows, Target might even know soon-to-be moms better than their own parents!
Techniques like this helped Target’s “Mom & Baby” section grow quickly and boosted Target’s overall revenue. But the challenge for retailers like Target is capitalizing on these customer insights without coming across as “creepy.” As you might expect, many couples were startled when Target sent them targeted pregnancy coupons shortly after they realized they were expecting. Some were so startled, in fact, that they stopped visiting Target altogether. Thus, Target started getting subtler with its promotions. The company would still send prenatal vitamin coupons, but they would be nestled between, say, a charcoal coupon and a lawnmower ad to make these targeted ads appear “random.”
Technology and its dominance of global markets
One of the greatest understatements of the 21st century is that tech has come to dominate the business world. Apple, Amazon, Facebook, Microsoft, and Google’s parent company Alphabet all routinely top the list of the world’s most valuable companies.
As Salesforce cofounder Parker Harris put it: “all business leaders need to be technologists… and every enterprise needs to become an app company.”
Labor-enabling vs labor-replacing technologies
Economists group technologies into two categories: labor-enabling and labor-replacing. Labor-enabling technologies help workers be more productive. For example, consider PCs and the internet — they’ve made it far easier to write essays, find information, or talk to coworkers. Then there are labor-replacing technologies, like the self-driving cars and industrial robots we mentioned before. As the name implies, labor-replacing technologies can remove the need for human workers. These opposing forces are in a constant tug-of-war.
Who wins? The results might be unexpected. For instance, think about the ATM, which became popular in the 1970s. Most customers no longer needed to talk to a teller inside a bank’s branch office. Many people assumed that this would eliminate the teller job altogether. Right? Wrong. Thanks to the ATM, banks needed fewer human tellers in their branches. But this made branches cheaper to operate, which led to banks opening more branches. And that, in turn, led banks to employ more tellers. The upshot? The number of tellers in the US grew from 300,000 to 600,000 from 1970 to 2010. In other words, the ATM actually created teller jobs instead of replacing them.
Technology and potential for job automation
There’s pretty strong evidence that automation will take away lots of jobs. A 2013 Oxford study found that half of all American jobs were at risk of automation by 2033. People with lower skill levels would be particularly hurt. President Obama’s Chief Economic Adviser found that 83% of jobs paying under $20 an hour were at risk of automation, versus just 4% of jobs making over $40 an hour. Furthermore, 44% of jobs requiring less than a high school education were likely to be automated, compared to just 1% of jobs requiring a bachelor’s degree.
But there’s also data to indicate that robots aren’t taking our jobs. Throughout the mid-2010s, American unemployment was low (less than 5% in 2017), workers were staying at their jobs longer, and wages were slightly rising. That hardly indicates that robots are massively stealing jobs. More broadly, automation could move people away from doing more manual tasks to doing more tasks that require the human mind. For instance, manufacturing plants envision having fewer assembly-line workers in the future — but more engineers, coders, and managers. Technology has created entire industries, like IT and software development. Automation wouldn’t just create science, technology, engineering, and math (STEM) jobs, either: self-driving cars will still need mechanics and marketers, for instance.
Some scholars take the opposite approach, arguing that white-collar jobs are as endangered as blue-collar jobs, if not more so. The writer Kai-Fu Lee argues that there’s not much economic incentive to automate low-skill, low-wage jobs like baristas; instead, companies looking to save money would rather eliminate skilled, high-paying jobs like financial analysts. There have indeed been a few notable cases when artificial intelligence has managed to match professional humans’ job performance. An AI bot called Amelia has proved surprisingly good at customer support for banks, insurers, and telecom companies; she uses humanlike facial expressions and gestures and tries to empathize with callers. And with every customer she handles, she gets even better. Even higher-skill jobs have been getting automated. One Japanese insurance company replaced 34 human agents with IBM’s Watson AI, and large numbers of people in the mortgage financing industry have seen their jobs get destroyed by automation. AI can’t yet replace doctors and lawyers, but some AIs have gotten good at the kind of research paralegals often do, and robots can now do some medical operations — both at a fraction of the cost of their human counterparts.
Definition of conversion
Conversion: Whenever a user does something that the business wants; the precise action can depend on the company’s goals. Conversions could include joining the mailing list, signing up for an account, or buying an item.
Definition of a funnel
Funnel: A metaphor for how the pool of potential customers shrinks before they make a particular “conversion,” like buying a product. For instance, suppose an e-commerce website gets 1,000 visitors, but only 500 search for something, 100 put something in their cart, and 50 make a purchase.
Definition of customer lifetime value (LTV)
Lifetime value (LTV): How much money a customer will bring you, directly or indirectly, over the duration of your relationship with them.
For instance, if a college bookstore thinks that students will spend $500 a year on textbooks over 4 years of school, each student’s lifetime value would be $2,000. Generally, companies will only try to acquire a customer if their lifetime value is higher than the cost of turning them into a customer (known as customer acquisition cost, or CAC).
Market penetration and segmentation
Market penetration: How much of a target market a product or industry actually reaches. For instance, there are about 30 million teenagers in the US, and if a teen-focused social network had 6 million teenaged users, it’d have 20% penetration of the teenager market.
Market segmentation: Breaking down a huge, diverse market into smaller, more specific ones. For instance, a company could segment its market by gender, location, interests (also known as “psychographics”), and income (part of so-called “behavioral” segmentation).
Product management vs product marketers
Product Manager (PM): PMs sit at the intersection of business, design, and engineers. Based on what the customers and business need, PMs decide what products (apps, websites, or hardware) to make and what features the products need to have, then work with engineers to build and launch the products. Think of them as the conductors of the orchestra: they help all the various parts work together to make a great piece of music (or software, in this case).
Product Marketing Manager (PMM): A slightly more marketing-focused version of Product Managers, they’re more focused on launching and marketing products instead of building them.