Converted: The Data-Driven Way to Win Customers’ Hearts

Book Preface

Digital marketing is about keeping the faith. The faith that you can seduce someone to buy more, vote differently, or just love your brand—all in a 250 x 250–pixel box as they’re scrolling through websites, purposefully trying to avoid the message you’re sending. It’s about timing, extraordinary scale, and hoping that customers willing to pay full price don’t find that 15 percent–off coupon code first. It’s also a promise—help the customer reach their ambitions—while you get yours: their money.
My world, digital analytics, is the science behind proving that this faith hasn’t been misplaced. That the messy pile of actions and metrics—the impressions, clicks, and conversions—wouldn’t have happened without exposing users to that six-second pre-roll ad on their way to watching another unboxing video on YouTube. It’s not a perfect world. Here’s reality: shoddy, imperfect experiments that need ninety days but only get a dozen, and porous data sets that may represent as much of reality as a coin toss.
In this space, I’ve been an analyst, researcher, inventor, lecturer, programmer, and, most shamefully, the father of many forgettable slides of glossy funnels and Venn diagrams. A witness to and participant in billion-dollar successes and expensive failures driven by ego, ambition, and, much less often, pragmatism. A key player in the executive rallying cry to be more “data driven.”
My latest gig, spanning more than a decade at Google, has given me the privilege to lead more than 2,500 engagements with our biggest advertisers. I have overseen initiatives that have acquired millions of customers, stretched conversion rates by more than 400 percent, and supposedly generated more than $2 billion in incremental revenue. (I don’t know how the economists derived that number, but I like it, so I’m going with it. I have my sins too.) But, like for most analysts I know, it’s the failures that made me second-guess my career choices, if not my sanity. It’s a messy world, and I’m part of it. We all are.
The moments that make you think your finely tuned model might as well be a random number generator? I’ve had my share. I’ve seen researchers remove individual survey results—“outliers”—until the results matched the product manager’s convictions. I’ve worked with executives who demanded absolute accountability for every dollar spent—until it came to buying the naming rights to a college football game. When their sales numbers showed they would have produced better returns had they wrapped their products in hundred-dollar bills and tossed them into the crowd, they questioned the data. And once, I had to partner with a particularly egregious group of consultants who pulled a revenue estimate right out of their asses. I know that because the footnote was literally “Replace these numbers I pulled out of my ass.” They didn’t even proof it. But nobody reads footnotes—apparently not even the board.
Well-intentioned executives and graduate students used to ask me about the secret to building a successful marketing organization. Was having one more success than your failures enough? Or was it found in embracing some Silicon Valley platitude like “Fail Faster”?
I wanted to know. And I have spent my career in search of an answer.
You’ll find me in Google’s Partner Plex, located on our Mountain View, California, campus, surrounded by relentlessly brilliant engineers who make serving up results to forty thousand search queries every second look effortless. While the engineers crank out code, manage systems, and do the heavy math, my team talks and strategizes with customers. We welcome them to our part of the campus with an AI-powered piano that composes its own music, a rainbow staircase of the latest search trends, and a virtual reality rig to paint in 3D. It’s enough to make Willy Wonka wish he had a golden ticket.

We also have conference rooms. That’s where the work gets done. They’re more than your normal meeting space, because they’re purpose-built for the tasks ahead. Stuffed with power and bandwidth, sugar and caffeine—courtesy of microkitchens tucked into a series of pull-out cabinets with “Drink” and “Eat” laser-etched into the sides. The conference tables, built from the same dark, laminated wood, could be branded with “Think.” Here we collaborate with Google’s largest customers on the future of their products and verticals.
Today I hold the title of Google’s chief measurement strategist, but ever since starting out as an analyst, I was always curious to understand how executives made decisions based on the work my team produced, and why two companies often acted differently when shown the same data.
The question presented itself again and again. Why did companies use identical information to compete in entirely different ways? Over time, a pattern became clear. Most companies were focused on a single moment, a single sentence, a single interaction: Hey, take my offer! They were using the data to change creatives, colors, and targeting, changing words and tone across endless experiments. Anything to get the immediate yes—but everything was short term.
It made sense. CFOs demand accountability. Digital advertising gave it to them. They could instantly connect clicks to action. They spent a dollar and then the customer spent ten. It set the strategies and their weekly dashboards. But it was boxing the CMOs in. That one moment was all that mattered and all that was measured.
But to the CMOs, this made sense too. The larger they grew, the more data they collected, the better they could make the most out of each moment. Innovators, disruptors, or whatever new model the venture capitalists were funding would have to go through (and pay for) the same lessons themselves, setting fire to their balance sheets.
Until their competition sprinted ahead.
Some of these companies had come to the Partner Plex knowing they couldn’t catch up by running the same race. They needed a new way to compete. We helped them find it. Instead of optimizing to the immediate, what if you built a business around long-term relationships with customers, using data to understand who the best customers were and what products they wanted to buy, then building around them? What if you could leave your competitors, with all of their data and their short-term thinking, just to poke around in the scraps?
The answer: You can. And it works incomprehensibly well.
The marketing success stories over the next decade will be about more than just clicks and conversions. They will be about people and conversations with customers that build into relationships.