These are my notes from a talk delivered by new Quibb member Andy Johns in August. Andy is currently the head of the User Growth team at Quora with experience at both Facebook and Twitter managing growth. He delivered this to the LA Lean Startup Circle (a great group if you're in Los Angeles organized by Quibbers Patrick Vlaskovitz, Pete Mauro, and others).
Lots of interesting examples from his time at Facebook and Twitter worth reading. Some of the basics Quibbers will be familiar with if you read AC's blog or any of the other Growth Hacking posts in vogue.
With Andy's permission, enjoy:
What is a User Growth team?
A User Growth team is a team with the goal of optimizing the flow of users in and out of your product. It’s the technology equivalent of a finance team: the growth team manages the flow of users just as the finance team managers the flow of money in/out of the company. Additionally, the most effective user growth teams are given the same authority as the biz dev, M&A, or finance teams.
Focusing on growth is like switching from simple to compound interest. While compounding is only an incremental improvement in the short run it leads to massive returns in the long run. For tech, small optimizations now (+0.5% conversion) can have a large magnitude effect later (+1 million users).
Remember that user growth teams can only optimize existing flows. Startups have no use for a growth team pre-P/M Fit. Until then, focusing on growth is like putting $1 in a high-yield compound interest account… it's just a waste of time.
User Growth Accounting… The Formula
This is the most basic formula for user growth:
New user signups
+Reactivations of current users
-Deactivations of current users
=User growth (over any given period)
Managing to this formula is the most basic job of the growth team. It's a simple but powerful formula because it's an easy way to understand if you’re engineering your product to grow itself. If your product is growing itself, these numbers will improve constantly.
The best way to visualize the formula is with stacked plot graph where each figure has its own line (during the talk, Andy showed us real data from a consumer startup he helps out (sans axes) and walked through the graph and highlighted the insights to look for… awesome data porn).
‘Checkers’ Level User Growth Tactics: Twitter
There are two types of growth tactics: basic ‘checkers’ tactics and advanced ‘chess’ tactics. An example of ‘checkers’ tactics from Twitter:
Twitter's first landing page was designed by an outside firm and had a lot of assets: user profile pictures, popular tweets feed, and a search bar. The signup window and signup button were both really small. The page was visually busy, had a ton of latency, and didn’t focus on signing people up. It didn’t perform very well.
Beyond bad optimization, the signup page was a ‘sacred cow’ within Twitter. Everyone coveted it but no single team owned the page. It was really hard to optimize because there was no central authority and people kept coming out of the woodwork to defend different parts (side note: the signup page is your product's single most important page… one team must own it) The growth team fought tooth and nail to optimize the page.
Over time, the team was able to: axe the feed of tweets, shrink to only a few profile pictures, minimize the copy, cut the search bar, and enlarge the signup window so it took 1/3 of the page.
End result: 2-3x uptick in signups within 24 hours.
Additionally, the signup form today does not have a field for picking your twitter handle. Originally, users had to pick a handle in order to click-through. The problem was over time the number of available handles shrank as the product grew (organic friction). To remedy the problem, the growth team moved picking a twitter handle to a secondary page and suggested available names to the customer.
This works because the vast majority of people join Twitter to read tweets, not tweet themselves. They don’t care about how cool or appropriate their handle is. Now Twitter gets them through the door into the product much faster.
End result: 2x uptick in signups within 24 hours.
‘Chess’ Level User Growth Tactics: Facebook
While at Facebook, Andy’s team developed a system for Impression Maximization. The goal was to optimize any given impression based which type of user is looking at the page. They divided Facebook’s users into 3 buckets:
- New Users
- Engaged Users
- Stale Users
Once bucketed, Facebook broke the webpage into modules. You can imagine the Facebook page as a set of zoned, each one filled with an smaller Facebook web app. The user growth team uses the modules to optimize the page for each of these users in order to maximize retention. For example, Facebook will show a new user a whole bunch of tools to help them find their friends and connect. Its all about converting that new user into an engaged user as fast as possible (aka FB’s Magic Number: get a user 10 friends in 7 days and they’re converted… shared by both Andy here and Chamath Palihapitiya at the User Growth Conference).
Other ‘Chess’ Level Tactics
Here are some other ‘chess’ tactics from Andy. I can grab more details from my notes if people are interested:
- Embracing Constraints: Place artificial constraints on your user growth team (or any team) to make them think out of the box. At Facebook, the growth team developed a new automated analytics system after they were told they had to deliver value but couldn’t hire any replacements to fill holes on the team.
- Channel Optimization: Create an optimized experience for each channel your customers come from. When a user arrives from Google, they should have a different experience then if they arrive from Stumble Upon.
- Funnel Friction: Increasing friction isn’t always bad. One company had consistently low LTV leads coming through their landing page so they increased friction by requiring a paragraph description of the lead’s request. The added friction increased overall LTV and saved them time and money dealing with bad leads.
- Understanding Virality: Borrowed from epidemiology, squeeze extra growth out of a product by optimizing for three figures: Branching Factor (additional users acquired from each new user), Cycle Rate (how long for each user to bring onboard those additional users) & Decay Rate (how fast your users loses interest in your product). Look at Draw Something as a cautionary tale (really high Branch & quick Cycle but incredibly fast Decay).
- Early Adopters: Early Adopters will not be your primary customer in the long-term. Once you gain traction, you don’t need to worry about pleasing them anymore. The key to startup success is consciously making a graceful transition from serving your early adopters to serving your average user.