Julie Zhou's discipline of growth
“Oh my God. How do you do that?”
This is the slack-jawed question for Julie Zhou when friends learn that, after turning off the lights at Yik Yak’s West Coast office, she hits the gym and pumps 240 lbs. of pure iron. But Julie is certifiably flyweight, weighing in at just over 100 lbs. How can she lift something at that scale?
Her answer, as it turns out, is fairly rote and unexciting: “Every single week, you lift one amount. Next week, you add five more pounds. Then, you add five more.”
And yet there may be a deceptive genius here. Deadlifter by night, Julie effects a similar magic in her work life. She helps startups get swole.
General Assembly’s Introduction to Growth Hacking is a slightly more expensive (and extensive) version of Julie’s answer. At the helm of this class, she draws from her decade-long experience growing companies and teaches how strategic perseverance can result in Herculean feats.
Currently, Julie is heading up Growth at Yik Yak, a location-based social network that is mega-popular on college campuses. Before that, she was Growthmaster at Hipmunk and had worked in product marketing at Google. In addition to General Assembly, she mentors at 500 Startups and is putting together a “practical guide” on the topic.
“A lot of the literature out there on growth and traction tends to be written by people who have observed lots of companies but not worked in lots of companies through various stages of growth,” Julie says. “I’m someone who’s lived it and breathed it and been in the trenches.”
Out of the trenches, Julie has returned with wisdom. But it’s not for the faint of heart.
Do you even growth hack?
Julie’s first trench was Google.
Right out of college, she moved to Mountain View and joined the juggernaut’s Product Marketing organization. As with most 22-year-olds who land the Google gig, it was clear Julie would have a bright future in technology. What seemed less probable was that it would be in hypergrowth. Because, while there’s a lot of room to grow at Google, virtually none of it is devoted to Growth. For Google, the question of expansion is not How? It’s Where to?
What Julie learned quickly was how to win attention inside of the engineering-driven machine.
“Anything marketing wanted to do had to be backed up with data,” she says. “It had to always have some quantitative justification around why company resources should be spent on those efforts versus anything else.”
One of the major projects Julie took on at Google was text ads, an undertaking she at first mischaracterized as “un-exciting.” Her budget was slightly more exciting, but the medium seemed simple: You had a headline, display URL, and two lines of description to get someone to clickthrough. How big could text ads actually grow?
After looking into the data behind text ads, Julie realized just how much power was at her disposal. On Forbes, Julie wrote: “On a canvas of 95 text characters, tiny changes had massive effects –clever capitalization and site links increased engagement rate by three or four times.”
In the case of text ads, her team was then able to transfer these learnings into larger, costlier channels, such as display and video. Experimenting with smaller stakes had been smart, because it gave them the data they needed for bigger restructuring: “Refining the science of our marketing at the text ad level allowed us to avoid costly mistakes in less measurable channels with far more variables – background color, logo size, color matching, vocal tone, billboard angle.”
By investigating the impact of small insights and scaling these low-cost measures into larger projects, Julie was laying the groundwork for what would become her expertise.
“The first four years of my career were just boot camp data-driven marketing as a science.”
2008, 2009, 2010, 2011–this was eons ago. To the extent “growth hacking” existed as a term, it did so in the immediate network of Sean Ellis. But Julie was already honing the skills that would fill out her LinkedIn profile. She was upping user engagement through a deeply analytical approach.
In the years since, the marketplace for this skill has become more competitive, and with that competition has come a degree of sophistication. It’s at the point where OG Andrew Chen is being poached by Uber and Julie is handing out syllabi.
“It’s become more of a discipline of growth, rather than something that hackers put together on their own,” she says.
But back at Google, Julie’s resume was young. To continue growing herself, all she needed was to have the massive budget vanish.
Today, Hipmunk has a solid reputation as a travel booking site. But when Julie joined the startup, there were only nine employees.
“I did not have nearly the budget I did at Google,” she recalls. “In fact, I had very, very little budget and quickly realized that was no way for Hipmunk to grow. Our resources as far as paid advertising were way too small to make a large impact. So I started looking for other ways to grow.”
One of the working definitions of growth hacking is that it works against definition: You have to be creative. Tactics, as Andrew Chen would say, decay.
A Brief History of Growth Hacking might proclaim that the most significant growth hacks occur within a product feature. Think of Dropbox promising to up your storage space when you refer someone. Think of Airbnb integrating your listing seamlessly into Craigslist. Still, a versatile definition allows for more. It lets a truly good growth hacker meet the product where it’s at in its lifecycle—whether that’s in-app or elsewhere—and identify what needs nourishment.
Minor product tweaks, of the sort Julie had specialized in at Google, didn’t seem likely to improve Hipmunk. The company had already developed a great product. It just couldn’t rely on traditional paid growth channels, which were dominated by competition with far deeper pockets. Increasingly, it became clear that hacking growth would mean tapping into channels vs. product.
Luckily, they had done the math.
“Analytics is more important the scarcer your resources are,” Julie says.
While this has been a constant in her evolving career, it perhaps never mattered more than in these early stage startups. Hipmunk might not have been on other companies’ radars, but they were self-aware. They had existing data from their product and could guess what might solve the channel problem.
Disinterested users can be a stubborn bottleneck, but users who got on-app actually liked Hipmunk. In this case, the app’s gatekeepers were non-incentivized intermediaries. The sellers. After their iOS app had been featured in the “New and Noteworthy” section of Apple’s App Store, it experienced strong adoption. In comparison, Android adoption was at a standstill, but not for any difference in quality.
So, Hipmunk knew what a successful channel could look like. And there wasn’t much of hope of mobilizing this channel by reducing friction in-app or providing social proof on the product. These definitions of “growth hacking” were limited. Instead, she had to use an approach more classically biz dev. But unlike a lot of biz dev, it had to be on the cheaper end.
Julie calls these kinds of strategic partnerships “atomic bonding.” Atomic bonding means finding a way to link a larger partner’s success to your own. For Android, the partnership she had to cinch was Google. So commenced a full-court press of emails and LinkedIn messages to old colleagues. Julie wanted to figure out who she needed to talk to at her former employer. Eventually, she learned that to be front-and-center on the Google Play store, she needed to woo a department deep within Google: Developer Relations.
Bearing Hipmunk schwag, she drove down to the campus with her startup’s sole Android developer. Going before Developer Relations, they asked a simple question: What could Hipmunk do to convince Developer Relations of the strength of their app?
Sphinx-like, Developer Relations presented them with a challenge.
Google I/O was fast-approaching and the Nexus 7, a new tablet, would be premiering there. The Nexus 7 had its own crowd to please. It needed to appear handsome with a suite of design-savvy apps. If they so chose, Hipmunk could try to win Google’s favor by optimizing its Android app for the tablet and in accordance with Google’s newest design guidelines.
A larger company wouldn’t have been able to execute on this. A two-week timeframe. Too many moving parts. Too many tickets in the queue. But even if capital at Hipmunk was minimal, there was an upside: “We had competitive advantages in being a nimble, rather exciting startup.”
There’s a certain hustle to what Julie Zhou does. Here, her to-do list changed dynamically with the rules of engagement. So when Google said two weeks, that became Hipmunk’s calendar. And despite forging the relationship, Julie delegated its execution to her developer. Even though she had been assigned the banner of growth, she had to hedge on other players and circumstances.
“Growth is everyone’s responsibility,” she says. “It’s not just the role of marketing, it’s the role of the entire company and every single department needs to be involved.”
Once Hipmunk had renovated their app, the powers at Google decided to feature it on the Play store. It was quid pro quo, no money exchanged. After I/O, Hipmunk’s daily active Android users skyrocketed. By atomic bonding with a much larger mover, Hipmunk had been able to produce the same result as a larger mover.
Still, it’d be foolish to think of this strategy as something turnkey. It’d be more foolish yet to think of it as an endorsement of “It’s who you know.”
Julie’s growth effort required analytics upfront to acknowledge where Hipmunk was prospering and underperforming. It identified how past successes could be replicated in new relationships. Research granted them access to the right decision makers within the leviathan, and once in front of Developer Relations, Julie and Hipmunk had to make a promise to prove their value. Finally, Julie entrusted a department outside of her own to deliver.
From there, growth was a gradual, measured team effort. Its strategy was much more nuanced and tactical than A/B test-and-Just Add Water.
If Julie had pigeonholed herself in product tweaks, the Android app would still be in some deep layer of the Google Play store, waiting for someone to appreciate its splendor. Instead, she recognized what the app needed to grow in particular.
Where does growth belong?
Depending on where your product is in its lifecycle, growth can belong in any number of departments. At Hipmunk, where the product-market fit was already strong, Julie had to do some back-of-the-envelope biz dev. But atomic bonding Google is probably not your growth hack.
So, where should your growth belong?
“There are many companies that are small enough or have strong enough vision that growth can be shared among a lot of different orgs,” Julie says. “Invariably, departments start siloing themselves and losing the big picture as companies get bigger. At that point, it’s worth creating a central growth effort. Many companies do it differently. Some have growth living within product. Others have growth living within marketing. Other companies still have growth as its own org that has its own marketing and engineering design and analytics team.”
Moving on to Yik Yak meant Julie rethinking her skill set. Social apps fill a different role in users’ lives than a retail site.
“At Hipmunk, a lot of the growth opportunity came from the top of the funnel,” she says. “A lot of the growth opportunity at Yik Yak comes after someone has landed on the app and you’re trying to build a habit and relationship with the users.” This means driving the majority of growth initiatives from within the product team.
Getting Yik Yak in front of users was a non-question. Many companies salivate at the idea of having a hot college app, something those coveted millennials won’t shut up about. Yik Yak is that app. Its location-based network makes users feel like a part of a larger conversation that also feels intimate. A viral chain reaction, it has spread across the world.
“Any company, any person, any department who is thinking about driving growth for their product needs to start with analytics and needs to start with understanding how the users behave right now,” Julie says.
As a rule of thumb, Julie runs tests to confirm her hypotheses before sinking too much time into any one effort. One experiment at Yik Yak was around re-engagement notifications. If Yik Yak told its less engaged users they were missing their school’s feed, would they feel the FOMO?
“We had a targeted group and control group, and we carefully measured the return rate for the targeted versus the control group because there were some people who would come back anyway,” Julie says. “There was a massive difference immediately in the targeted users versus the control users. But that was somewhat to be expected if you get a notification.”
The first 24 hours was only the beginning- Julie tracked behavior over a longer period and got encouraging news. Notifications were an effective way to herd yaks. They not only brought users back, but saw them returning with regularity. The task was now to make this data actionable, to make behavior repeatable through scalable automation.
As a kind of lesson, Julie offers two actionable strategies:
You can send a high volume of notifications that manage every point in a user’s lifecycle, based on whether they have taken certain actions or been off-app for a period of time.
You can make every individual type of notification that much more tailored to the user. A blast notification is much less effective than a personalized notification. There’s a lot of personalization with a notification that can make it work better.
A growth hacker might also try both and measure which works better. Both are meaningful changes to how users interact with a product and require significant buy-in and collaboration with product, engineering and design teams. One may prove stronger in the long-term. For a hot startup like Yik Yak, keeping return behavior strong may be the rule of engagement now, but the tactics will have to change over time.
“Yik Yak is very, very strong in the college community, but we want to be equally as strong when our current users graduate and go to cities,” Julie says. “There, the metrics become different.”
Transience is the friend of the growth hacker. Today’s growtherati need to know more than how to transition from one type of company to another. They need to grow their skillsets as the DNA of their company changes.
Julie’s career so far has been a crash course in growth hacking, but that doesn’t mean it provides a template. When she asks her General Assembly class, “What do you think growth hacking is?” they audition answers like, “Being very data-driven,” “being creative,” “getting the most out of very limited resources,” and “leveraging your company’s strengths vis-a-vis your competitors.”
And these are all right to an extent. They’re passable. Necessary, if insufficient. It helps that they’re vague.
“I do hear a lot of, ‘Tends to be short term,’” Julie says. “They think of growth hacking as discovering arbitrage opportunities that exist in the marketplace and once you discover them, they’ll just go away.”
Growth, on the whole, is more of a slog. Her advice for growth hacking is pretty similar to her advice for deadlifting over twice your weight: If you want to surmount massive problems at scale, a combination of different exercises over time is the surest key to success. Shortcuts, like Googling “Get swole in two weeks,” will only yield pain and embarrassment.
And so Julie cautions: “It’s very boring. It’s not like I did this and suddenly became strong.” It’s that incredibly patronizing and yet effective adage: Slow and steady wins the race. This is true in product growth and weightlifting.
True to her credo, the process of weightlifting is data-driven. It’s iterative. It requires that she frequently run tests and track progress according to those.
“With weightlifting, it’s very easy to tell when you’ve made progress. There’s a higher number that you write down in a spreadsheet. It’s very, very applicable to growth. I’m very much a motivation-by-progress person. Here’s the KPI that I want to have by the end of the quarter. Let’s just track how we’re doing every week.”
So, Julie marches in to Yik Yak’s West Coast office and flips the lights on. She asks herself: “Are we better or worse off than we were last week?”