mobilized: An Insider's Guide to the Business and Future of Connected Technology - S.C. Moatti (2016)
Chapter 4. The Mind Rule: The Best Mobile Products Learn as We Use Them
To be successful, mobile products must adapt constantly. They do it fast, and slow.
Fast: Mobile companies relentlessly learn from their users and adapt to their rapidly evolving needs. Adapting is a matter of survival.
Slow: Mobile companies must reinvent themselves as they break new ground. They have no choice.
By constantly analyzing and optimizing information, mobile products work like an extension of our mind.
Mobile companies use two types of tools to learn from their users and get better at what they do: scientific and artistic.
Case studies discussed in this chapter: Facebook, Lyft, Nokia, WhatsApp, Viber, and Yelp.
On my first day working at Facebook, I got a booklet titled, “Facebook was not originally created to be a company.”
That sounded almost like a question. What, then, was it created to be? The answer was on the overleaf: “It was built to accomplish a social mission—to make the world more open and connected.”
Great mobile companies start with big visions. Not technology visions. Visions for a better world. Human-first visions. They assemble and organize amazing teams. The products they build often look like dreams come true. Overnight successes.
But where does greatness in mobile come from? How is it born?
Think of Uber. Or WhatsApp. Would you say you understand why or how they became so essential to so many people? Do you feel you were warned that they would matter one day? Do you think of them as the result of a lot of trial and error?
Personally, I couldn’t figure out the secret sauce. I understood that these companies had all the pieces in place. They were in the right place at the right time. They had supersmart people backing them. They worked really hard. And then … what? Magic happened? They got lucky? How did they do this? Why them and not another start-up?
I ask myself the same questions about anyone who achieves something greater than themselves. How did they do it? How do people actually go about achieving what most people think is completely impossible?
According to New York Times best-selling author Robert Greene, top achievers apply a skill called mastery to everything they do.
In his book Mastery, Greene says that mastery starts with what he calls apprenticeship. “We love the myth of the artist as creator,” he says.61 But the truth is a lot less glamorous: all great masters relentlessly practice their craft for thousands and thousands of hours. They learn and persist until an opportunity to break through comes along. Then, they become masters.
Take Nelson Mandela, someone who always fascinated me (and many others). The son of a prominent political figure in South Africa, he became a political activist in opposition to the racial segregation laws in place in South Africa at the time, called apartheid.62
Continually evaluating the political landscape, Mandela learned quickly and constantly adapted his tactics against apartheid in the face of changing conditions. Initially an advocate of nonviolent action, he became the commander of a revolutionary army, which led to a sentence of life imprisonment. Prison didn’t stop him. He befriended his jailers, becoming a potent symbol of grace under adversity. People around him praised his character. He refused to compromise with the racist regime in place. His fame only grew bigger.
In 1990, Mandela was freed from prison. The next year, the last remaining apartheid laws were repealed. In 1994, Mandela became the first black president of South Africa.
By overseeing the peaceful transition of his country from apartheid—an unjust system that had been in place since 1948—in only four years, Nelson Mandela became a symbol of freedom and hope. To some, it may seem like it happened overnight. Instead, it took a lifetime.
Along the way, Mandela reinvented himself many times: from an activist to an outlaw, then to a convict, a hero, a head of state. Ultimately, he received a Nobel Peace Prize for his life’s work. This came as a by-product of achieving mastery.
Mobile products obviously are not on the same level as Mandela’s struggle against apartheid, but they operate by the same playbook. They are always learning with and from their users, by breaking down problems into smaller, more manageable steps that can be solved relatively quickly. They operate like our mind, which constantly analyzes and optimizes what it perceives.
When they find something that works, they do more of it. Gradually, each problem they solve brings them closer to becoming masters at their craft. This is when we notice them. Their seemingly overnight success is a by-product of aiming for mastery.
We’ll begin by looking at how mobile companies learn fast, by tirelessly solving problems and seeking impact. Then, we’ll examine how they learn slow, by reinventing themselves and breaking new ground.
Learning Fast: Relentlessly Adapting to Us
Remember when I told you in the previous chapter how I started using Lyft and Uber when I needed to go places? I’m not the only one by far.
Early on, car service company Lyft was facing a challenge every business would love to have: too much demand.
On Friday night in particular, rides are in short supply because many more people need them. Plus, even though they make more money on Friday night, Lyft drivers often take the evening off to have a social life of their own.
For passengers, this dearth of drivers could create the same type of stress I experienced when I couldn’t find a cab, and Lyft learned early on that it needed to find a solution and increase the supply of available rides.
To solve the problem, the company reached out to a team of behavioral economists.63 Recall from the previous chapter that behavioral economists believe most people do not make decisions based solely on reason.
The researchers sat down with a dozen Lyft drivers, trying to understand how they decided when to work and when not to work. They found out that Wednesday morning is the time when drivers make weekend plans, so this is when they made the decision to work on Friday night or take it off.
Based on that insight, the researchers recommended that every Wednesday morning, Lyft should show drivers the opportunity cost of taking Friday night off.
Then they tested and compared different ways to communicate this to the drivers. They learned that drivers were more willing to work on a Friday night after they understood not how much they would make if they did but how much they would lose if they didn’t.
You may be thinking that this feels manipulative. If you do, I understand. However, Lyft’s intent is to meet the demand of its customers and maximize the profit of its drivers. Because of its iterative approach, it optimized its offering until it was able to give drivers enough reasons to meet the demand of its customers on Friday night.
What I find particularly impressive is how Lyft was able to balance and optimize supply and demand in real time and adapt to the needs of both drivers and passengers.
For mobile companies, learning fast like this isn’t optional; it’s a matter of survival. On mobile, user behavior changes at an accelerated pace, as does the technology that caters to it. Every few weeks, people become more proficient with their mobile products. Every 18 to 24 months, they upgrade to a new one.
There is no slowing down. Mobile companies know that they cannot move at a leisurely pace. What worked a month ago no longer works today. What was enough in a given context no longer is in a different environment. What seemed worth selling is suddenly offered for free from a competitor.
They need to learn and get better fast. Whether they’re at the early stages of building and launching a product, in the customer acquisition phase, or in a more mature phase of monetization, they still need to relentlessly adapt.
Take mobile messaging giant WhatsApp.64 It is constantly adding new features—groups, videos, emojis. People love these, so they use WhatsApp all the time. The more they use it, the more WhatsApp learns what they like and the better it gets. The better it gets, the more people use it.
The WhatsApp team solves one small problem after the other, as quickly as possible. It goes through a relentless trial-and-error process to mirror customers’ needs and provide those things people want in the real world—assurances that their message was read, that the person they are trying to contact is checking in, and so on. In developing countries like India, where they often still pay for text messages and data usage, people love WhatsApp because it lets them send messages to their friends for free.
Because it creates a product that keeps getting better over time, it gets more and more people to use it. As of June 2015, it had an astonishing 800 million users.65
Learning Slow: Breaking New Ground
People who accomplish great things learn fast … and by doing so, they improve incrementally until they slowly reinvent themselves and achieve a level of mastery.
Soon after he took his company public, Trulia cofounder Sami Inkinen set out to reinvent himself completely.66 He became a triathlon world champion. In 2014, he and his wife broke the world record for rowing unassisted from San Francisco to Hawaii. Inkinen jokingly says that his biggest accomplishment is that after such an intense experience, he is still married.
I asked him what his secret was for accomplishing these impossible goals. “If you can measure something, you can improve it,” he said. “Every time I try something new, I write down three things I could improve doing next time.”
Successful mobile companies do the same. As they find ways to keep up with the pace of innovation and learn from their users, they strive to break new ground. They target new users or launch new offerings. Each time, they need to learn everything from scratch. This slow journey to mastery takes time.
Companies that target new users often need to create an entirely new experience just for them. Take Jack (not his real name), a recently retired operations manager. He likes spending time with his grandson Tommy, a baseball prodigy. Jack would love to go see every one of Tommy’s games, but they live several hours away from each other, so this is not an option. Instead, Tommy suggests that Jack check out the videos and photos he posts on Facebook.
Jack has never been on Facebook. He has been reluctant to join the social network because he finds the idea of a stream of spammy and irrelevant content stressful and overwhelming. He’s also anxious that the experience will be unforgiving to novices and that he won’t be able to navigate the network easily.
To his surprise upon downloading and trying out Facebook for the first time, Jack isn’t inundated by posts, pictures, and updates he doesn’t care about. He isn’t even asked to fill in a profile. Instead, a brief message on an off-white screen suggests that he connect with friends. It’s a lot easier than he expected.
Jack’s experience is completely tailored to his level of proficiency with Facebook. The company learned from late adopters like Jack and created an entirely new experience to onboard them successfully.
Another way companies reinvent themselves in mobile is by launching new offerings. They often wrestle with this because their large functional organizations get so absorbed in the day-to-day that they lose track of what their users want. Their products stop getting better over time. It becomes an uphill battle to learn new ways of operating and innovating.
Many companies attempt to solve that problem by practicing what’s called selective insulation of their mobile divisions. On the one hand, they make it clear who is responsible for mobile decisions. On the other hand, they encourage everyone at the company to participate in the mobile transformation.
WhatsApp is a great example of selective insulation. When Facebook acquired it in 2014, the WhatsApp team was very small.67 Even after the acquisition was completed, it continued to operate independently.
Leading mobile browser and advertising platform Opera Software uses a similar playbook. Opera’s secret sauce is its unique ability to build tools that are very fast and don’t consume a lot of Internet data on a mobile phone. In the early 2000s, Opera used these ingredients to create a very efficient mobile Internet browser. If you’re old enough, you may remember it. Most people in India use it to this day. In fact, it’s one of the most used smartphone apps in the country.68
In the late 2000s, Opera Software created a new division called Opera Mediaworks and empowered the leaders of that division to completely reinvent the business. They applied a different recipe to the secret sauce and built an entirely new mobile product: one of the largest mobile video networks in the world. Today, it reaches over a billion people.
Companies sometimes need to reinvent their entire business model. Remember Clay, the contractor we discussed in chapter 3, who doubled his business when he started using Yelp? Because he did great work, he got great reviews. Customer references are the best kind of advertising. Besides, he got all of it for free. A great deal for him, of course, but that’s not sustainable for a company like Yelp.
Like any business, Yelp needs to make money. It initially did so by carving out areas on its website where it could display banner advertising. On mobile, however, there is no space for banner ads because the screen is too small. Mobile banners ads are annoying rather than enticing to users.
So Yelp had to reinvent its business model. The company went back to its users and tried to understand what they would be receptive to. They learned that users have no patience when they search on the go; they want instant recommendations rather than have to scroll through dozens of reviews.
Yelp tried promoting select small businesses in its search results by placing them at the top of the list. It worked. Instead of displaying banner ads, it now charges small businesses a fee in exchange for greater visibility and better placement. Users like it better this way, and Yelp brings in revenue.
When mobile products and the companies that create them actively engage in this ongoing learning and reinventing process, everyone benefits.
Learning in Context
Companies that get really serious about learning do so in context. They focus on impact. They don’t care how much effort it takes to reach their goal. They care about having more people use them so they can learn faster and get even better. They also realize that what worked a month ago no longer works today. What was enough in a given context no longer is in a different environment.
Remember real estate marketplace Trulia? It relies on a metric called Net Promoter Score, or NPS, to measure impact. NPS is a widespread metric in mobile.69 It asks users a simple question—“How likely are you to recommend Trulia to a friend?”—and captures their answer on a scale of 1 to 10 (1: not at all likely, 10: very likely). If the rating is anything below 8, it assumes that users are not satisfied.
The higher its NPS, the more likely users are to recommend Trulia to a friend, so the more people will use it.
In fact, growth hacker guru Sean Ellis, who helped companies like Dropbox and Eventbrite go from zero to IPO, talks about the need to constantly reverse engineer a product in order to make it a must-have experience.
“Use NPS to understand what users might like and what they might be disappointed with,” he says.70 “If you don’t nail the first experience of a user, there’s usually no second experience. And then, the ongoing experience is what referrals are based on, so it’s pretty critical.”
The success of a new mobile product often hinges on how it is rolled out. Some companies release very gradually, first to a randomly selected 1 percent of its users, then 5 percent, and only then to all users. Others set up sandboxes where they can get things right first, before making a new offering available to all when it’s finally ready.
Mobile messenger Viber, for instance, releases new features in its smaller markets first, then in its larger ones. While it isn’t widely used in the US, over 100 million people around the world use Viber every month to communicate for free with friends and family. The company was acquired in 2014 by Japanese Internet giant Rakuten for close to $1 billion.71
One of Viber’s largest markets is the United Kingdom. Because of the strategic importance of such a large market, Viber is reluctant to test new products and features there. Instead, it has identified a couple of markets that behave just like the UK, but are smaller. These markets would normally not get a large share of attention from Viber’s executives, but because of their similarity to the important UK market, they have become vital test markets.
What I find interesting about Viber’s approach is that it uses its own markets as sandboxes to continue to improve its offering and, at the same time, keep the majority of its users happy. Again, everyone benefits.
Learning in context means focusing on impact. It’s what Trulia does with NPS. It’s what Viber does with its careful use of test markets. Now let’s look at how to build mobile products that learn in all the right ways to create this impact.
Building for Learning: Mobile Products as Extensions of Our Mind
Mobile products and the companies that create them must learn relentlessly. Quickly, so they adapt to our rapidly changing needs and behaviors. And slowly, so they reinvent themselves by breaking new ground.
To accomplish this, they rely on two types of tools: scientific and artistic. Scientific tools measure everything and power optimization and rapid iterations. After a while, though, they produce diminishing returns, and this is where artistic tools come in. They help mobile companies devise creative solutions to stagnant problems.
Scientific tools wire everything so it can be measured. They collect massive amounts of data that they can then analyze. The technology they use to power this is called big data.
My former boss Ken Rudin ran the team at Facebook that looks after big data72—all the analytics, optimizations, and business insights that come from it. His team constantly keeps up with technology trends such as Hadoop and deep learning in order to remain ahead of the competition.
Rudin created a weeklong training program called Data Camp, which everyone in the company can attend. This program is widely popular because it is fun and people learn interesting things there. More importantly, it spreads a data-driven approach across the entire organization. It shapes the culture to be one of incremental learning and relentless optimization.
Realistically, only a few businesses can afford to build their own scientific tools, so dozens of companies, including Mixpanel, Kochava, and Leanplum, have stepped in to offer them out of the box, as turnkey solutions.
Scientific tools also power methodologies such as lean agile, which are designed for rapid iterations and frequent rollout. This approach is the foundation of Steve Blank’s lean start-up methodology,73 which is designed to demonstrate empirically that there is value in a product or feature.
But scientific tools alone only get so far. After a while, they produce diminishing returns. Another type of tool is needed: artistic tools.
“Big data is as much an art as it is a science,” says Rudin, Facebook’s former analytics czar. “It’s about business needs. If nothing changes, you’ve made no impact.”
Rudin trains his team to ask the right questions in order to make sure they spend their time and brainpower on priorities that matter to the company.
Artistic tools are absolutely critical to the type of breakthrough success a mobile start-up needs if it has any ambition to make it big. If you work at such a place, you must learn to master artistic tools. If you on the other hand work at a large company with stringent constraints, using artistic tools may be too risky.
People often call the successful use of artistic tools by another name: luck. They underestimate artistic tools because they are less pragmatically actionable. They are under the impression that these tools aren’t real.
But there is method to the madness. Artistic tools require thinking outside the box. That’s much harder to do than applying a proven formula. Yet both are crucial to success.
Mobile companies recognize this, so they assign what are called growth teams to learn and master scientific and artistic tools.
“Growth teams integrate product and marketing together,” says Uber head of supply growth Andrew Chen.74 Product teams, they say, are very focused on building the core features that make people want to use a product. Marketing teams focus on attracting new users. The goal of a growth team is to get users to experience the core value of a product as quickly as possible.
Thinking scientifically. Thinking artistically. Emulating both sides of our brains, left and right. This is how mobile products become extensions of our mind.
Building For Learning: Scientific And Artistic Tools
To constantly learn, reinvent, and ultimately reach mastery, mobile companies rely on scientific and artistic tools.75 Scientific tools help set a goal and track toward it. Once their effectiveness levels off, artistic tools take over. They help mobile companies think creatively about overcoming roadblocks to improvement.
Scientific tools come in two types:
Goals, such as a business equation, define desired outcomes. A business’s profit, for example, is a function of volume (the number of transactions) and price. This translates into a simple business equation:
Often, companies break down the number of transactions by customer because they want customers to buy multiple times from them (it’s more effective to get existing customers to buy more than to acquire new ones). So their business equation has three factors instead of two. It looks like this:
Mobile companies set goals that roll up into one of these three factors. Their teams focus on either increasing the number of transactions by getting existing customers to come back; increasing the number of customers by acquiring more customers; or increasing profit per transaction by raising price.
Funnels, such as attribution and tracking, trace every customer interaction and identify which ones convert toward a goal and which ones don’t. Teams use them to identify gaps and try to fill them. They optimize conversion in a very systematic and methodical way.
Once their scientific tools start producing diminishing returns, mobile companies will turn to artistic tools, which are high risk/high reward. Artistic tools allow mobile teams to think outside the box and be creative when systematically looking for breakthrough opportunities.
There are three types of artistic tools:
Hooks, which make the top of the funnel wider, peak people’s curiosity. Think about the Facebook poke feature or the Zillow zestimate. (Who doesn’t want to poke a friend back or know how much their house is worth?) Hooks attract customers that would normally not have been interested. They provide a meaningful incentive for people to enter the funnel.
Shortcuts, which cut a slice at the top or bottom (or both) of the funnel, remove an entire step from a user’s experience. Amazon’s personalized recommendations are a good example. These recommendations, labeled “other books you may like,” appear immediately after a purchase. The step of searching for similar books is skipped all together.
Layers, which open up a whole new layer in the funnel, create new channels that allow mobile companies to reach their customers. For example in the early days, Airbnb sourced a significant amount of its listings from Craigslist. While this wasn’t legally allowed by Craigslist’s terms of service, Airbnb used this layer to grow its inventory meaningfully.
Mobile products that use scientific and artistic tools to constantly analyze, interpret, and optimize our environment act as extensions of our minds.
Building Nokia and Facebook for Learning
I implemented the learning fast/learning slow approach at both Nokia and Facebook.
Let’s start with Nokia. Every morning, my entire team would get together for no more than 15 minutes. We’d review our goals, how we were tracking against them, what needed to be done next, and who would own the action items. We wouldn’t focus our attention beyond what had to be done that day because we knew we would meet up again the following morning to touch base.
After two weeks of intense iterations, we’d update our product. We’d collect data about how our users received the changes we made. If they liked it, we’d plan to do more of it. If they didn’t, we’d plan to course correct. Based on these early signals, we’d come up with a new set of goals for two weeks and would go back to our daily optimization.
The results of this approach were astonishing. Everything we tracked looked like the proverbial hockey stick on a growth chart. The team was so excited about the impact it was making that it worked harder and better together. It made our users happier and they would use our product more. It created a virtuous cycle. The more people used our product, the better it got. This approach is an example of scientific tools in action. It is often referred to as agile methodology when it is used in engineering organizations.
But at Nokia, the sales and marketing teams were also on board. The entire organization was working in unison. So instead of seeing results only at a technical level, like more predictable and higher quality product releases, we also saw a business impact: more user growth, better retention, increased revenue. It wasn’t just the technology that was constantly being optimized and reinvented; it was the entire business. This approach is definitely unorthodox. It’s an example of using artistic tools to foster communication and synergy.
Thanks to our outstanding results, we reached the top 1 percent of the app store. Millions of people downloaded our app. Our business unit eventually came to the attention of executives at Nokia’s headquarters in Finland. They contacted me directly, curious about the secrets behind our success.
We invited them to come to San Francisco and see our team in action. It was January. We knew it was an offer they couldn’t refuse: in the winter, Finland is buried under six feet of snow while San Francisco is mostly sunny.
After this visit, they invited me to discuss ways we could integrate our independent business unit into the larger organization. This eventually led to our mobile product being preloaded on every Nokia smartphone. It was a huge accomplishment for our team because at the time, two out of every five smartphones were sold by Nokia.
Nokia empowered small, nimble teams like ours, letting them learn fast and grow independently. If they were successful and the time was right, they brought them into their mainstream business units as a way to reinvent themselves.
Now, let’s look at Facebook. In 2013, my team conducted a test to determine if Facebook users would be open to receiving push notifications.
We selected a random 1 percent of users who had opted out of receiving push notifications. We showed them a screen immediately after they sent an instant message that explained to them the benefit of push notifications. For instance, they could know immediately when their friend responded to their message. It was an example of a hook, one of the artistic tools we described earlier.
What was especially tricky about our test was how much it required from the users to be successful. Those who were interested in receiving push notifications would have to go through five separate steps in order to update their settings. None of the five steps could be skipped. And we couldn’t do it for them.
Typically, mobile companies lose 90 percent of their users with each step, so we were expecting hardly anyone to go through such a painful, five-step process. It was a long shot.
To our surprise, almost one in four users went ahead and changed their settings. As soon as we found out, we expanded the reach of the test and implemented the execution machine of goal setting, tracking, improving, rinse and repeat that scientific tools power.
What I found interesting about this particular test is that it revealed a change in user preferences. As recently as six months before we ran the test, a majority of Facebook users were skeptical about push notifications so they opted out of receiving them. They weren’t quite sure what they were or whether they had any value. Some felt a bit assaulted by them.
Things changed quickly when users realized that push notifications allowed them to know instantly when a friend had posted a photo of them, or sent them a message, or tagged them in a post. All of a sudden, people wanted push notifications. What was appropriate earlier was no longer the norm now. We had to adapt and give users another chance to decide whether or not to receive push notifications.
Remember and Share
Where does greatness in mobile come from? We can learn about it by looking at greatness in human beings. Some extraordinary individuals achieve what most people think is completely impossible. They reach their level of mastery by learning fast, to constantly adapt and improve, and slow, to reinvent themselves and break new ground. Mobile products, too, learn fast and slow.
Fast: On mobile, user behavior and technology changes at an incredibly rapid pace. Mobile companies understand that the more people use their product, the better it gets. They make sure their product scores high on the Net Promoter Score (NPS) index.
Slow: To keep up with the pace of innovation on mobile, companies need to reinvent themselves by reaching new users, launching new products, or inventing new business models. They have no choice.
Mobile companies rely on two types of tools to learn from their users: scientific and artistic. Scientific tools let them test relentlessly until they find something that people want. Then, they do more of it. But at some point, this produces diminishing returns. Artistic tools let them think out of the box and create disruptive changes.