What are Some Decisions Made by the "Growth Team" at Facebook that Helped Facebook Reach 500 Million Users?
Posted Friday, May 18, 2012, at 1:53 PM
This question originally appeared on Quora.
Answer by Andy Johns, Product Manager-User Growth, Quora:
This question suggests it wants to know the exact, tactical decisions. As much as I would like to share those tactics in great detail because it's fascinating and readers of this question would find it interesting/valuable, I'm legally bound to a certain amount of confidentiality so I'll only go as detailed as I'm legally able to. That basically means I won't share numbers and I won't comment on things that aren't directly discernable to the general public. Also, the list of optimizations is too long for anyone to remember. I also didn't work on the majority of those optimizations directly but tended to learn about them via close proximity to the people solving the problem they were assigned to solve. The team grew to 30 to 40 people so most of my day-to-day work was contained to a subset of the optimizations that were taking place at any given moment.
There are a few different types of "decisions" though. There are decisions around tactics, decisions around strategy, decisions around hiring, and decisions around priorities and culture. The growth team participated in decisions across all types and all are critical to a growth team.
There are too many to name but most come down to Internet Marketing 101: test, optimize, rinse, and repeat. If you want a list of methods for testing and optimizing various parts of a product or referring channel, then read my answer to What are some top strategies for conversion optimization? You can pretty much guarantee that just about every one of these strategies was applied in one way or another to various parts of Facebook's product. But again, I can't share specifics that aren't obvious, public domain, nor can I share stats. Nor can I speak to the complete level of depth and complexity that the entire team went into to fine tune the parts of the product that enabled growth to happen. Some tactics I also don't want to share because they were so damn effective.
Who was hired to work on growth was arguably the most important decision being made. The team was led by Chamath Palihapitiya. He is by far the best person I have ever worked for/with. He is deeply analytical, rigorously focused on success (or as he would say, "crushing it!"), gifted as a natural leader and motivator, aggressive and unafraid of taking risks, and knows consumer tech at a black belt level. He was the backbone of the growth team and making him the head of growth was a brilliant decision. I remember in my first week on the job I went to lunch with him to get to know him a bit more. It was a formality that all new hires went through. Pretty standard stuff. The conversation we had was far from standard though. I remember asking him, "So what kind of users am I going after? Any particular demographics or regions? Does it matter?" and he sternly responded "It's fucking land-grab time so get all of the fucking land you can get." In other words, don't ask such a stupid question next time. Get the entire planet on Facebook. Clear enough, right!? I knew I liked him from the start.
He went on to bring others onto the team like Blake Ross, Alex Schultz, Javier Olivan and several other really talented people spanning the spectrum of skills from direct acquisition marketing (SEO, PPC, email, a/b testing, merchandising, link building) to deeply technical back end and front end engineers, designers, and data scientists.
Javier Olivan is responsible for building and scaling the international growth side of the team. He helped build the internationalization team that consisted of engineers who built the translations application that allowed our own users to translate Facebook for us. This video that explains the process always pulls at my heart strings.
Perhaps the greatest lever in getting to 500M or more users was making the site available in virtually every language on the planet. As Nico Vera stated in the video, language localization is the "great equalizer". It made Facebook a platform capable of supporting everyone on the planet. I believe Facebook is available in 80 languages now. That became possible because engineers on the growth team built a tool that allowed the users to crowd source translation of the site for us. Growth was not about hiring 10 people per country and putting them in the 20 most important countries and expecting it to grow. Growth was about engineer systems of scale and enabling our users to grow the product for us.
You can think about strategy a few ways. One way is in how we generically framed growth funnels and how that tended to inform growth roadmaps. For example, you could say that growth is broken down into a few fundamental questions:
- How do I increase the rate of acquisition i.e. get more signups?
- What can I do to activate as many users as quickly as possible in their first 'N' days?
- What are the levers for engagement and retention and how can I pull them?
- How do I bring churned users back into the system to "resurrect" them from the dead?
From those question you can go down the path of identifying products that impact a certain metric (e.g. what products help drive acquisition?) and then figure out how to either (1) optimize those channels to produce more value OR (2) build new acquisition channels from scratch to add more acquisition volume. For example, here is a list of things that could take part in the acquisition funnel:
- Contact importing --> sending invites
- Open/click rates of invite emails
- Conversion rate of user post-invite
- Logged out homepage design and how that converts users to signup
- The steps of the signup form
- Account confirmation post-signup
Then you can try and layer your current acquisition channels with brand new channels. Think of acquisition as a stacked line chart. You want to add more stacks to that chart because each stack represents a new source of acquisition. For example:
Strategy can also exist in the form of company acquisitions or strategic partnerships. Techcrunch noted in February of 2010 that Facebook purchased a company called Octazen (More info here). This was a brilliant move negotiated by folks on the user growth team because Octazen was known for having a robust list of contact importing services. A large percentage of Facebook users have address books that are stored in long-tail mail providers like GMX. Not everyone uses Hotmail, Gmail or Yahoo Mail.
Take a look at the list of mail clients that LinkedIn supports contact importing for (notice the blue link at the bottom where you can get the full list which is almost 100 clients long). I pulled this directly from LinkedIn's contact importing flow.
Enabling users to import all of their contacts gave companies like Facebook and LinkedIn the data they needed to make it easier for them to recommend friends/contacts to their users and connect you with the people you know. Plenty of social media services allow contact importing. But how many of them were rigorous enough to enable virtually every email importing service on the planet? I only know two at this point: Facebook and LinkedIn.
Partnerships were also negotiated by the growth team. Sometimes those deals were struck with international entities to enable growth outside of the U.S. In October 2010 Facebook partnered with Yandex to feed the Firehose of status updates to Yandex (More info here). The largest search engine in Russia is now merchandising Facebook products prominently in their UI. Growth definitely comes from that.
Sometimes the growth team was responsible for taking big bets. Facebook Lite was one of those bets. You can read more about it here but TechCrunch got it right. The Lite site wasn't an attack on Twitter or any related competitor. It was a stripped down version of Facebook that was faster. A slow site kills adoption. Google has shown everyone in the tech industry how important site speed is for driving usage and adoption. Amazon has shown that 100ms in additional latency means they lose $1M of revenue per day (More info here). Facebook needed to be fast to grow in areas where broadband penetration didn't allow for site performance that most of us in high broadband markets are used to having. So a small group of engineers buried themselves in a conference room for 4 weeks and built the Lite site and deployed it in India. The product was eventually killed off but it served its purpose at the time and demonstrated that speed was as much a priority for growth as it was for engineering.
Culture and Priorities
It has been documented quite a bit in various interviews throughout mainstream and tech media, but Facebook built a very unique culture. Within the growth team Chamath Palihapitiya was largely responsible for cultivating a culture of balanced aggression where we pushed as hard as possible to make the site grow.
The growth environment was palpable. You walked into a corner of the building and would see flags hanging that represented not only the international nature of the people we hired, but as symbolism for the global adoption Facebook intended to have.
I can't find the picture at the moment but there were two banners that hung above the growth team at the 1601 California offices. The first sign read "GO BIG OR GO HOME" and had a picture of Godzilla next to it (that made it more awesome). The other read "UP AND TO THE RIGHT". That's what we saw everyday, all day while working. It was a constant reminder of our team mission. There were several other messages scattered throughout the offices like these.
The growth team was also given tremendous air cover by the executive team. That meant that we were enabled to be aggressive and take risk in the name of giving everyone on the planet the opportunity to connect with each other on Facebook. Growth had executive support. This picture should give you an example of how close Chamath was with the core leadership in the company, and consequently what that meant for the growth team:
That's Chamath on the left with Zuck, the VP of engineering, the VP of Tech Ops, and the VP of Product. Growth wasn't mitigated to a sub-function of a higher function within the company like a Paid Search team might be a sub-function of the Marketing team within an e-commerce company. Growth was a horizontal layer across product like engineering/ops is a horizontal framework behind product. Not only would someone ask "What's the performance impact on site speed or stability if we build and ship 'X'?" it became common for people to ask "What's the impact on growth if we build and ship 'X'?". The decision to make growth a canonical part of the product, engineering and operational discussion was a really important decision that the executives made.
I've had this sort of question and conversation come up a bunch over the past two years. There is still this very nascent understanding of what user growth actually means and how it contributes to a company in the consumer technology world. The natural instinct is for it to come back to tactics, tactics, and more tactics. Executing on tactics is certainly a necessary part of making growth happen. But getting to 500M users was a function of all of these things coming together. Culture, priorities and hiring set the table for us to construct a team that was committed to a mission of tactics and strategy. A growth team that "crushes it" cannot be built in the absence of culture, priorities and hiring since tactics and strategy are made possible by the former.
More questions on Facebook: