Back to Blog

As Uber Shows, Growth Hacking Should Have Limits

By Christopher Steiner  •  Mar 14, 2017

Christopher Steiner co-founded Aisle50, YC S2011, which was acquired by Groupon in early 2015.

In some ways Uber has operated as a flawless, well-managed machine of a large company: gobbling up markets, expanding its product, catering to the user, identifying weaknesses and expunging them. In others, however, it has operated as if it were a startup strapped for survival. Uber has skirted, and possibly broken, the law in ways that are, admittedly, rather genius, but also clearly anathema to how most large corporations would behave.

Uber's culture and office environment seem almost lawless, according to some accounts, which is certainly more reflective of a kind of 'brogrammer' startup environment than a big company environment. The result is a denigrating patina that won't be scrubbed off the company easily. Three of the top five non-news Google results for 'sexual harassment company' are about Uber.

Uber is hardly the only startup turned large company—worth $60 billion at last funding—to be sullied through some of its growth strategies. This effect is an indirect byproduct of prevailing ethos in Silicon Valley and the startup community in general, when it comes to startup growth. The idea is that startups need to hack, claw and fight their way to growth in any method possible—provided it's not blatantly illegal.

Sometimes this can be great advice.

When I was going through Y Combinator with my co-founder, one of the best office hours we ever had was when Geoff Ralston suggested we simply go around the big companies whose products and marketers we were trying to court—these were companies such as Coca-Cola, Procter and Gamble, Kellogg's and Unilever. He said we should simply create subscriptions to these products with our retail partner without input from the brands.

Taking Geoff's advice, we went from having a dearth of deals to offer our users to a surplus. We received a few complaints from big companies along the way, but it was more likely that they'd take notice of our rogue offerings, learn more about our startup, and eventually become paying clients. Coca-Cola came on as a client in this way.

If a brand asked us to stop—and some did—we took the offer down. But we made them ask. This is exactly the kind of thing that a startup can do, with nearly no fear, that a large company can't or won't do.

When Instacart launched, it did so independent of retailers. Trader Joe's asked the company to stop listing its stores and merchandise—and Instacart complied. Now Instacart has partnered with some of the biggest retail brands in the world. But to get started, it did things that it wouldn't or needn't do now.

Some of the difference, of course, comes down to a company's appetite for risk. A startup naturally has nearly limitless room to consume risk, whereas an established company will be very careful about risks—to outsiders, it often looks like large companies don't take any risks at all. This is a healthy evolution. We don't expect public companies with public shareholders to behave in the same fashion as a startup operating out of a garage.

When startups grow at rocket-like rates, however, they can get caught in between, with a foot still in the garage, and a foot in big company mode. The culture at a company growing as fast as Uber or Zenefits might have been haphazardly constructed, if it was constructed at all, and not built around the multifaceted challenges and responsibilities that fall to a big company.

A startup may have already established itself as a unicorn many times over and possess a roster of enterprise clients, but it still may have programs in place from when it was a nascent company, trying to hack its way to growth.

With a program internally dubbed Greyballl Uber took things even a step beyond tactics that a three-man startup should be carrying out, and it did so long after the company had already found a vector of wild growth. Greyball was designed to thwart law enforcement from finding Uber cars in places the company may have been operating  illegally. For users suspected of being affiliated with law enforcement, Greyball served up fake versions of Uber’s application, and even fake maps with erroneous locations of nearby Uber cars.

It was incredibly well-executed software designed wholly to spoof law enforcement—a rather nefarious turn for a company worth $60 billion. To be clear, this is not something that would be forgiven of a startup, either, as it rather conspicuously breaks the law.

A startup's culture needs to be defined early

Uber has had a difficult time handling some aspects of its startling rise, and much of its issues, from conduct toward women to unethical marketing strategies, come from a lack of culture building early on at the company. Very few companies will experience Uber-like growth, but founders should build their culture as if this kind of growth is coming.

Trying to course-correct on culture, when a company might be hiring 20 people a month, is nearly impossible—and it's the kind of mistake that leads to situations like that at Uber.

It's trite to say that a startup's culture and the behavior of its employees emulate that of its CEO and founders, but it doesn't make it untrue. The founders and how they conduct themselves are more responsible than anything when determining how employees interact with each other and how they treat the rules, ethics and regulations of the outside world. Setting out rules and values is important, but a founder's own actions speak more loudly than words in an employee handbook.

As Paul Graham says, be good.

The culture of Silicon Valley is at fault for some of this

Startup founders are encouraged to do nearly anything to get their company off the ground. It's this kind of coaching that tutors startups to coddle and court their first customers with special care, answering questions, building features, doing whatever is necessary to keep them on board. Founders need to do things at the beginning that they wouldn't do later, when the company is bigger.

In general, these are good lessons. Startups have to scratch their way into existence. A fight to survive, however, leads startups to do all kinds of zany things: scraping leads, spamming unsubscribed email lists, employing black hat SEO methods.

Most founders figure that, once rolling, their companies will behave better. In many cases, this proves to be true. It goes back to risk tolerance: as founders' companies get bigger, they have more to protect, and most founders will be naturally less tolerant of strategies that could sully the company or the brand—or devalue their equity.

But it can be hard to get out of edgy growth-hacker mode for some founders and startups, especially after having been encouraged to try anything that doesn't burn the company down or land somebody in jail.

It's a fine line, and it can be hard to discern sometimes, especially in the murk and chaos of building a company. But some lines, murk or not, are plain as day.

Obvious, but not to everybody: if a growth hack breaks the law, skip it

This seems rather intuitive, yet we are continually served with examples of companies who couldn't stop themselves from building software that was too clever.

It's not only Uber that has made the news recently for this. Zenefits created a secret program dubbed The Macro, which allowed some of its California sales representatives to complete mandatory pre-licensing education courses in less than the legally-required 52 hours by gaming third-party software.

And perhaps worst of the bunch, Volkswagen has forfeited $20 billion to federal government agencies and consumers for creating software to mask its diesel cars' true emissions during tests. Six Volkswagen executives were indicted for their roles.

The total fallout from the Uber's Greyball program remains undetermined. The same can be said for Zenefits. The damage to VW's brand, however, can be measured in the billions—and that's over and above the levied fines and legal damages.

For a startup—really, for anybody doing anything—growth and triumphs can be addicting. But there is no trail that can't be uncovered on the web.

Running afoul of other sites’ Terms of Service—if you're asked to stop, stop

Clearly, there is a long list of startups who have built databases, client lists, and any other number of things based upon the information available on other websites. Scraping other sites usually goes against the terms of service of the site being monitored. But random ToS agreements don't really stand in the way of most startups.

That being said, if a site has grown wise to what a startup is doing, and has asked the founders or somebody else at the startup to stop, the startup should comply. There needn't even be an acknowledgement made of any kind by the startup. But it should cease doing whatever was breaking the other company's ToS. Not doing so goes against the unwritten rules of fair play on the web, and it also makes the startup more vulnerable to legal damages in the future, especially if the startup were eventually to reach a far larger size.

During its rapid expansion, Airbnb drew consternation from some, and kudos from others, for mining Craigslist for listings. Airbnb also automated the process of putting its own properties on Craigslist, with links back to Airbnb, if landlords selected this option when listing their property. It was crafty play that helped Airbnb grow as quickly as it did. At some point Craigslist cut Airbnb off.

Craigslist tried to do the same thing to two companies, 3Taps and Padmapper, who were scraping the sites rentals and listing them on their own sites in different, and arguably better UIs, but both companies continued to work around Craigslist's digital barriers, IP blockers and legal desist letters. Eventually, Craigslist sued both companies, and reached settlements that included money and binding agreements to not take content from Craigslist.

Airbnb, having grown to a business valued at more than $1 billion by 2012, had a lot to lose by ignoring such requests, so it eventually stopped.

Growth-hacking is great, but it should be done with consideration of the future. No founder wants a closet full of skeletons if and when her startup takes off.