Summary of Build by Tony Fadell

The ideal size of company to work with is 30-100 people. Big companies teach you the wrong things, like how to get promoted, and getting promoted at Google is not a real accomplishment. It shows only that you knew how to hack the system.

Don’t work as a consultant with McKinsey or one of its competitors. You work only with Fortune 5000 companies led by risk-averse CEOs who call you to make decisions they should be making, and you get the job by promising a magical solution to scared companies. You then end up shouting advice from the rooftop and walking away, leaving others to implement it, which will never help you build a company.

Be a CEO of one. Don’t count on your manager or founders to make your project successful, thinking they’ve planned it and you just need to do what you’re told. They may not have planned it. Talk to teams that are closest to the customer, like support and marketing, to understand how you product is being perceived by users, which can be completely different from the story you hear internally.

Failure

College gives you the wrong mental model for failure. It tells you that you should feel bad if you fail, because only bad students fail in college. Good students pass every course every semester, maybe with high marks or with okay marks, but they pass. If you’ve failed, you’re useless. But this is a bad model of the real world. Your 20s should be devoted to finding what you like. Take up work you like — if you love quantum computing, or space exploration, or synthetic biology, work on that! Join a community of like-minded people. It doesn’t matter if 99% of people think synthetic biology is dumb. If you think a company is awesome, work there even if they offer you the lowest position, because you can prove yourself and work your way up. Start a startup. 90% of your 20s should be characterized by failures, since that’s how we discover what we care about. We can’t skip the pain and go straight to the conclusion. You should have butterflies in your stomach. If your experience kicks your ass, the kick will propel you into a new phase of life.

Don’t be too afraid of failure — Steve Jobs had many failures like the G4 Cube:

People don’t remember them — they remember his huge successes like the Mac, iPhone and iPad. So, you should have a healthy failure rate, and don’t obsess about your failures.

If your startup fails, what you’ve made still matters. It still counts. You’ll walk away having created something you’re proud of. You’ll have learnt and grown, and developed relationships with people you’ll work with later.

People

More important than the company you work at or what you’re building is who you’re working with. Seek out mentors. If you’re interested in self-driving cars, and Elon Musk doesn’t respond to your tweets, message someone two levels below him. Ask an intelligent question. Share something you’ve built. Help with a problem he’s facing. Be kind. Don’t be spammy, like messaging and asking for funding or a job. Doing it right may result in a friendship. Or an introduction. Or a job. And any job working with your heroes is a good job.

People who have high standards and criticize work that’s only good enough are sometimes confused for assholes. They’re not, unless they criticize the person (“You’re a bad engineer!”) rather than the work (“This is a bad app!”)

Don’t work with people (or companies) you don’t trust. Things will only get worse. And you’ll be miserable. But don’t constantly keep quitting before you’ve accomplished anything — it will be obvious in your CV.

You should always keep networking. Not in a fake way like approaching conference attendees to pitch your services, but forming a genuine human relationship. Network constantly, even when you’re happy with your job and have no intention of leaving. People who don’t find their options limited when the time comes. One executive who was Steve Jobs’ protege and worked directly with him building the iPhone eventually decided to leave Apple. You’d think he’d have every advantage in the world, that everyone would want to work with him or fund him. But when he left, he realized that he had never spoken to anyone outside Steve Jobs’s inner circle. He had no clue how to raise money, or whom to talk to, which delayed his progress. Your only connection to the world should not be through your employer.

Decisions

When you make a decision, you can frequently get data to validate, invalidate, or improve your decision. You should look at this data with an open mind.

But data is often not available when you’re doing something new. When building the first iPhone, nobody could validate that building a smartphone without a keyboard would work. In fact, all the data said that phones have to have physical keyboards. When collecting data, you might ask the wrong questions, or ask them the wrong way. Users don’t say what they think. Users can’t make decisions, only say yes or no to decisions you’ve already made. People collecting the data might add their spin to it. Data can tell you what’s not working but not how to fix it.

You can’t A/B test your way to a successful design. To A/B test, you first need to decide what A and B are. You can’t randomly throw things against the wall to see what sticks.

Some people don’t understand the limitations of data. They take the idea of data-driven decision-making too far: they’re afraid to make a decision not backed by data. This attitude is crippling. They’re unable to move forward because of analysis paralysis. If you make an opinion-based decision, and it doesn’t work, you need to take responsibility for it, and people sometimes insist on data-driven decisions out of a fear of being held accountable, which is not healthy.

We should be informed by data, not controlled by it. You should always give yourself the option of saying “Yes, the data says X, but I’m making an opinion-based decision.” It’s also not a dictatorship, so leaders should explain themselves. And listen to dissenters. Don’t argue. Hear them out and, if agreement is not reached, say that this is in our users’ best interest and we’re not going to reach the right decision by consensus. Then expect the dissenters to disagree and commit.

If you’re the leader and your boss asks for more data, explain that you’ve collected all the data available. Understand his concerns. Then explain that your decision is in the best interest of the users.

Some bosses have a personality type that makes them want to be nice. So they keep saying “We need more data” over and over again as a non-confrontational no. This wastes a huge amount of time and is frustrating.

Marketing

Sometimes a product team builds a product and then throws it over the wall to a marketing team and tells them, “Here’s the product. Now market it.” This doesn’t work. The marketing team needs to understand not just the what but also the why. Why the product was built. Without this, they’ll make up some stories like “It will do X” when the product was designed for Y. And customers who buy the product will be disappointed to find out it doesn’t do X. They may dismiss the product as a whole as useless, not realizing it was designed to do Y, resulting in it failing. To prevent this, marketing should be involved in the design process, right from the beginning. When the Nest thermostat was being prototyped, the marketing team and the CEO Tony prototyped boxes with different messages to see which ones work. They had a primary, secondary, and tertiary message on each box. The team figured out whether these messages would appeal to someone walking down the aisle in a shop who stopped for a second to look at the box. This way, the value proposition on the box, on the website (“Buy this because…”), and in press interviews matched what the product actually did.

If you try to align the marketing and product teams close to launch rather than at the prototyping phase, you may have to throw away the product you built and rebuild it from scratch to do what people want. Or launch a product that doesn’t do what people want, making it fail.

Some companies have a PM who focuses on product development, leaving marketing to a separate marketing team. This doesn’t work for the above reason.

When presenting a product to a user, present the why before the what:

Plant the virus of doubt, by reminding them of how something they use daily is annoying. Make them feel the frustration again. Then, when they’re primed, present your solution. A story should have both logic (to justify the action) and emotion (to get people to actually take action now rather than maybe after a few months).

Sales

Salespeople traditionally earn a commission at the end of every month for the sales they made that month. This makes them not give a shit about the company’s mission or about whether a customer they brought in will be happy and continue using the company’s services or churn. All they care about it earning their commission. They become hyper-competitive. Salespeople go on vacations where they get wasted and puke in hotel lobbies, or are dragged out of the company holiday party by cops. Customers are treated transactionally, like an ATM machine you walk up to and withdraw money from. Once people feel treated this way, and not as people, it destroys their relationship with the company. After all, you don’t have a relationship with an ATM machine. This is hard to recover from. Have you gone to a shop and seen the salespeople fight with each other to attend to you? Companies like The Wolf of Wall Street show sales culture accurately.

To fix this, salespeople should be remunerated with a competitive salary, plus commissions that vest over a schedule, like 15-20% immediately, then a second tranche after a few months, and so on. If one of the customers the salesperson brought in churns, subsequent tranches won’t be paid. Increase the commission to compensate for the vesting. This will make the right salesperson realize that he can actually make more than by traditional commissions. If the salesperson accepts a commission in the form of equity, increase the commission even more.

Any new contract should be signed off by customer success, to prevent overselling.

Product-Building

There’s a lot of focus on user experience, but you need a step back and look at the customer experience (CX) end-to-end. For example, if an advertisement promises a benefit the product doesn’t deliver, to the customer, that’s all the same. If you’re building a device (like a thermostat) with an app to control it, the app is as important as the device.

The installation experience is part of the CX — if it takes too long, the buyer’s non-technical wife may want the product returned. If the user is stuck during installation and reads a support article, which is unhelpful, the customer experience is again bad. To shorten installation, Nest included the appropriate screwdriver in the box, which was costly, but improved the critical first experience of the user. And it paid for itself in reduced support costs and five-star reviews.

Within a company, people often think in silos, where each team does their part right, but for a good customer experience, all these teams need to work together end-to-end: marketing, installation, hardware, app and support.

Sometimes people engage in wishful thinking, like Jeff Bezos when he built the Fire Phone to make it easy to scan a barcode and order an item from Amazon. But you could do that with just an app. People often come up with ad-hoc justifications (scanning) that actually don’t justify the decision (build a phone). Scanning may be a good feature, but it doesn’t justify building a phone. Even people as accomplished as Jeff Bezos fail at this.

Managing

Managers should provide intrinsic motivation, like telling people why building a high-quality product matters, by explaining from the customer’s point of view. Managers should also provide extrinsic motivation, like a deserved promotion or raise, a party or a kind word.

A manager can manage 8 people. Up to 15 for a while.

As you rise up, you spend all your time on people problems, a tangled web of professional relationships, and secret plans by people against each other. You need to appear to listen to your board, but you should ignore them at least to some extent. You have no time to focus on the product you’re building.

CEOs shouldn’t micromanage, which is telling people how to do something. They can and should give input on what and whether it’s excellent or merely good. Tony read support articles to ensure they worked for the customer. When Tony visited the Aston Martin office, it was raining, and he found a guy in the parking lot examining cars to see which were damaged by the rain. It turned out to be the CEO. He’d send back any car that’s less than perfect, for rework, regardless of how many times it went back. CEOs should care. They should give a shit.

Startups

Techie-led startups make a few common mistakes. They build technology, but people buy solutions to problems, not technology. Or the startup builds solutions to problems that people they don’t have. Or that would be a problem years from now. Sometimes techies sell a vision for which the technology doesn’t exist. Or which requires an ecosystem that doesn’t yet exist, like the iPhone in the 80s, which didn’t have a cellular data network, or Wifi.

Don’t chase trends like VR. Tony built digital music players at Philips when the Internet was taking off and making consumer electronic devices was considered dumb. Ignore the hype.

Sometimes market trends doom your startup. For example, Magic Cap built a private network controlled by the the phone company, just as the open Internet was taking off. There’s no way to foresee and eliminate this risk.

Acquisitions

Nest sold to Google. Acquisitions are justified by good intentions and wishful thinking on both sides, which doesn’t pan out:

First, Google acquired Nest to supercharge it with Google technology and thus let them progress years faster than they could as an independent company. For this claimed benefit to pan out, it was critical for Nest to integrate with certain teams from Google. But those teams did not view the integration as critical, and they’d cancel meetings at the last minute, or not bother replying to emails. Sundar gave the excuse that they’re busy.

Second, Nest was told their sales will increase by 30% when Nest thermostats are listed in the Google store. In reality, the Google store team didn’t care to integrate with Nest. Sales actually decreased because of Google’s poor reputation for privacy.

Third, acquisitions can cause organ rejection. Teams within Google were wary of Nest. When there was bad press about Google’s privacy, Nest and Google jointly clarified that Nest does not use the ad-supported model that Google does. Teams within Google interpreted this as a threat. The acquiring company expects the newcomer to switch to the former’s culture. Nest was told to be Googley. But Nest already had a high-performing culture and didn’t want to destabilize the company by changing. But this, in turn, caused Nest not to be welcomed with open arms. 50 to 85% of acquisitions fail because of culture mismatch.

Fourth, acquisitions result in a lot of work that doesn’t benefit users or grow the business, like moving from AWS to Google Cloud, switching email and other accounts, complying with the acquirer’s security guidelines, dealing with the bureaucracy of a public company, and so on.

Fifth, when Nest sold to Google, their costs increased by $10,000 per employee per month. Part of it was higher salaries and benefits. Google also offered facilities Nest didn’t need, like conference rooms with high-end equipment, and brand new buildings. Connecting each employee’s laptop to the IT network cost $10,000/year on top of the cost of the laptop. Everything including finance, HR and legal got more expensive. A Google corporate tax got added to everything.

Sixth, the strategy can change back and forth. Nest was part of Google, but then Google created the Alphabet holding company and transferred Nest to it. Suddenly Google teams stopped integrating with Nest saying, “You’re not part of Google”. Then Alphabet tried to sell Nest and talked to different companies. After Tony Fadell got fed up and left, Nest was reabsorbed into Google. So Nest was originally not part of Google, then was, then wasn’t, and then was. As a second example, Google knew when acquiring Nest that it had high running costs. But some time after acquisition, Google told Nest to cut costs by firing half their employees.

If you sell your company, you’re losing control of it. Don’t kid yourself.

We think that senior executives got there by being highly competent. But Tony found Larry Page, Sergey Brin, and the rest of Google’s executive team to make billion-dollar decisions by saying whatever comes to mind, without having a plan. Or having a plan based in wishful thinking, not thinking through. Or not actually getting their organization to implement their plan. Or coming up with a new half-baked plan before implementing the earlier half-baked plan. Google’s senior management would bullshit Nest by pitching their new strategy as thoughtful, with a smile, when Tony could see it’s not. Famed executives whose faces you’ve seen on a magazine cover are no more competent than us. Sometimes they have the maturity of kindergarteners. Tony observed this in the C-suite at Philips, as a VP at Apple, as CEO of Nest, and as an executive at Google. People who climb to the top of the corporate mountain are frostbitten and oxygen-deprived.