Why Pinterest empowers its growth engineering team

The front desk at Pinterest is ribboned with blond plywood, as is the cafeteria jutting out from reception. Sturdier material has yet to arrive. Or maybe it’s just a new style. The effect is a company that appears, despite its massive establishment, to be under construction. Over 100 million monthly active users, billions of push notifications a month, and Pinterest is still very much a work-in-progress.

John Egan confirms that this is the desired effect. At a suspected valuation of $11B, it would be easy to say Pinterest has arrived. But as an engineering manager on the Growth team, it’s John’s job to see the company as a thing constantly under construction. He must always be identifying places where it can better engage users. But this philosophy also provides a blueprint for the ever-changing state of growth in tech. Growth, too, is a work-in-progress.

For most companies, the growth team is housed in the product or marketing divisions, with assists from engineering. Indeed, Andrew Chen defines growth hacking as the incorporation of an engineering mindset into the goals of product management and marketing. When speaking about the very first growth hackers, he says, “They looked at the problem through a quantitative lens, the way that an engineer would.”

But this paradigm, while enshrining the engineer, doesn’t often empower her. Engineers are often perceived as implementers, instead of strategists. Smart people who can solve a problem with someone else’s solution.

“When you look at a lot of people who talk about growth, many of them tend to come more from a product and marketing side,” John says. “I think there’s a lot of opportunity on the engineering side to come up with creative solutions.”

The drudgery of growth is hard to appreciate until you’ve done it.

Before Pinterest existed, John was starting up in San Diego in 2009—not exactly Camelot for startups, in time or place. But as an engineer fresh out of the university, waiting for his girlfriend to graduate, John wanted to make it work.

So, after clocking out at his day job, John would meet up with an old college buddy and code. They were building a video chat app, ChitChattur. To them, it didn’t matter if they were dislocated from the tech scene. Their hard work would be plain to see and pay off.

After months of development, ChitChattur was all ready to launch. For distribution, John had purchased some ads on Google and Facebook, which he was confident would spread the good word. And then maybe word-of-mouth would soar. 

But not much happened. Nobody cared about the app, because nobody knew.

John learned, as many did in the early days of Web 2.0, that there was more to launching an app than just coding functionality. After his girlfriend graduated, they moved up to San Francisco, where he joined Shopkick, a digital couponing app. And it was at Shopkick that John was given the chance to redeem his growth mistakes.

When Shopkick rolled out its growth team, John was quick to join. His earlier experience of stumbling around with ChitChattur made him want to understand what it takes to build a startup.

“And I really enjoyed growth, because it’s a mix of product thinking, engineering, user psychology, and analytics, which are all different areas that I liked,” he says. It required something different. A sense of thrift, a sense of creativity.

Shopkick is an app that gives users rewards for walking into certain stores. When John was working in there, users could redeem cash incentives for sending a web invite to a friend (and, of course, successfully referring them to Shopkick). But, because real cash was being used, 100% accuracy was required in tracking.

The tracking, however, was very stopgap. Once a referred user clicked through the invite, he would have to enter his phone number to register. Then, the only way to identify that same user as the referred one was to have him re-enter his number when he downloaded the app. It was a bad flow. The mechanism that ensured accurate tracking also resulted in low conversion.

So John proposed something that hadn’t occurred to any of the PMs: Set a cookie when users click through the initial invite.

“When you download the app and first open it, we do a fast app switch back out to Safari, read the cookie, and then switch back to the app,” he says. “That made it possible to attribute an in-app signup to the original invite the user came from. We were able to eliminate the phone number step and it increased our conversion rate by 50%.”

This is the type of insight John means when he says “creative,” a word he often drops in reference to his team at Pinterest. He’s referring to out-of-the-box solutions that only an engineer could dream up. And in growth, which is all about cutting the Gordian knot, engineers are best-disposed to get creative.

Later, when he moved on to Pinterest, John made sure to instill a similar sense of creativity in his own team.

Getting engineers to operate like growth people is as much about culture as it is talent. 

“Because growth is a relatively new area there aren’t a ton of engineers with prior growth experience,” John says. “Usually what we look for is not someone who is necessarily super familiar with growth tactics, but someone who is a strong product thinker and can think through the problems people are encountering. We have an interview that’s almost like a mini-growth PM interview. I ask a candidate about an app they use and we’ll talk through how, if they were put in charge of growth for that app, they would figure out how to improve acquisition, group activation, etc.”

John can usually tell pretty early in the conversation whether someone has the right instincts: “I have a sense of which people have an intuitive focus on things that are top-of-funnel, and not just an amazing experience that only 5% of users get.”

This boils down to product empathy. He wants to discover engineers who not only understand the underlying technical systems, but also the user on the other side. 

“The biggest qualities I look for are being impact-driven and metrics-driven,” he says. “The engineers that are most successful don’t just try to get the job done. They’re driven by trying to maximize their impact. They understand user mindset and user psychology, and figure out what is going to resonate with people.”

And yet hiring the right engineers is only the first step in the funnel. Then, John has to promote the right culture, so his team takes initiative and tests hacks of their own: ”I encourage every engineer to have ownership of their projects and, when they have ideas, to come to me or the PM and talk about it.”

And when an engineer does speak up, it makes a significant difference. 

As a writer, I’d like to think people with my verbal skill set are the best candidates for writing resonant copy. But there’s an engineer at Pinterest who’s making it seriously hard to hold on to this thesis. 

The growth team had been trying to optimize email subject lines to get higher open rates. A cross-functional huddle had categorized 10 different email types, which had to be supported across 30 languages. Coming together, John, a product manager, an engineer, and some members from the writing team heroically brainstormed a series of subject lines.

“The initial results showed only a 1% lift in open rates,” John says, laughing at their futile efforts. Afterward, the engineer on the project hatched a new plan and said, “I think I can do better.” John invited him to try.

“He started to test different data that we could put in the subject lines,” John says. “For example, you could use the board name or we could use the pin keyword, or the pin description or what topic the pin is related to, and figure out which of those components work. Then the second stage, after seeing what works, was to just test out every variation.”

Which copy would make you open an email? John. Hey John. Hey John! John!!!

“There were basically 300 different experiments being run by that one engineer. ”

By automating all of the subject line combinations, the engineer was able to scale the experiment at a much faster rate than the brainstorming huddle. And by optimizing subject lines according to the most successful variations, the growth team was able to see results much faster. Whereas before the team had lifted the open rate by 1%, this time they saw an 11% gain. And if you’re dealing with a magnitude of millions, like Pinterest, that’s a major shift.

Most startups are not dealing with data points at the magnitude that Pinterest is, but maybe a really good, enterprising growth engineer can help them get there.  

When the “elder” growth hackers wrote the playbook, they were inspired by the engineers around them. But coders somehow aren’t the face of growth, even if, in many cases, they’re better-primed to think about strategy than their peers. 

Pinterest is self-aware enough to know it’s a work-in-progress. That’s why it can afford to listen to insights from all throughout the organization. Instead of having PMs tell engineers to implement this or that, the growth team at Pinterest empowers its engineers to think creatively and take the lead in problem-solving. 

Smaller startups, John thinks, could benefit from this kind of thrift.

“Earlier stage, you have more room reserved for big bets,” he says. “You’ll probably invest in more big bets because 1% or 2% gains aren’t that impactful if you’re working off of a small base.The stage Pinterest is at, a lot of the experiments we run are incremental improvements and a good percentage pay off. But we still have a small bucket reserve for out of the box ideas and big bets. ”

This bucket exists for the same reason as the plywood. The architecture is stripped bare. Every morning, as engineers flood in from the Brannan St. entrance, it’s an invitation: If you can conceive a better way, feel free to rebuild it.