To prevent people from thinking Facebook apps are losing users when there are bugs in the user count data it reports to services like AppData, starting January 16th Facebook will only report user counts in much broader tiered thresholds instead of to the nearest 100,000. However, it will start reporting App Center rankings. The change follows a disastrous misinterpretation of Instagram data over the holidays.
Facebook explains that an app with 1,100,000 monthly active users (MAU) will now be shown as the #300 largest app by MAU and as having more than 1,000,000 MAU.”
Until now, Facebook app-growth tracking services, such as Inside Network’s AppData, could pull from Facebook the monthly, weekly, and daily active user counts to the nearest 100,000 of any app on its platform (except for Facebook’s own apps). The public and a developer’s competitors could use these services to monitor an app’s rise and fall. It provided great transparency to the Facebook app industry, but it also caused some problems.
Sometimes the transparency made it tough for apps to make changes, because if their experiment failed and user counts dropped, the world would immediately know. This may have discouraged innovation.
The other big issue was that if Facebook had an error when reporting user count data, or a tracking service had trouble recording the data, apps could end up showing zero users or the same user count as the previous day instead of their real stats. This could cause an app’s monthly active user count to plummet. Journalists unfamiliar with app-tracking services would sometimes jump to the conclusion that people were abandoning an app when really it was just a data-reporting error.
This came to a head over the holidays when a dubious New York Post article said Facebook’s acquisition Instagram was hemorrhaging users after scaring them with a privacy and ads policy change that it later went back on. Facebook denied the usage dip, and we and other outlets made strong arguments for why the Post’s claim that Instagram had lost 25 percent of its users was wrong and a clear misinterpretation of AppData, which only reports logged-in Facebook users, not all users. The scare temporarily sunk Facebook’s share price by a few points.
Following that fiasco, Facebook yesterday stopped reporting data on Instagram. That was a predictable change considering Facebook doesn’t publicly report user counts for any of the other mobile apps it’s built, such as Facebook for Android, Facebook Messenger, or the new Poke app.
Now Facebook has announced it will roll out a change to what it reports in the API in a move it believes will protect app developers and give them more freedom to iterate with their products. Instead of data to the nearest 100,000 users – such as saying an app has 1,200,000 users – Facebook will only report that an app has more than 1,000,000 users. This should obscure the user count impact of temporary glitches in data and changes to app designs. In consolation, Facebook will now report the App Center ranking of apps.
Unfortunately, the lack of granular data could hurt businesses like Inside Network (where I worked before TechCrunch), which operates AppData. Less precise data might not be able to command as high of a subscription costs from developers who pay for historical data on all apps, beyond the most recent 30 days of data that is publicly available for free.
On the other hand, while Facebook’s own apps are immune to the changes as their stats aren’t reported, the changes could still assist the social network and its biggest partners. If a popular Zynga game suddenly loses users or a data-reporting error would have made it look that way, it will be harder for the public to tell. Since Facebook’s success is in some ways tied to the top apps on its platform, user-count obscurity could help Facebook by hiding failures.
Facebook’s 10-Q reveals some juicy tidbits left out of its earnings call yesterday. The final price it paid for Instagram was $715 million in stock and cash. Despite rumors that it has stopped growing in mature markets, Facebook’s US user count did increase slightly from 168 million to 171 million monthly users during Q3. And the amount of people who accessed Facebook solely from mobile each month grew 24% to 126 million in since June 30th.
The Instagram acquisition finally closed in early September after months of legal review. The 10-Q explains that Facebook bought it for $300 million cash plus $221 million in vested shares and $194 million (aggregate fair value) in unvested shares. That totals up to $715 million. That number is not listed anywhere in the 10-Q, but it’s well under the $1 billion Facebook would have paid if its share price hadn’t slipped.
As for mobile-only growth, Facebook wrote in its 10-Q: “Approximately 126 million mobile MAUs accessed Facebook solely through mobile apps or our mobile website during the month ended September 30, 2012, increasing 24% from 102 million during the month ended June 30, 2012.”
That means the trend is accelerating, as Facebook saw a 23% increase in mobile only users from 83 million in March to 102 million in June. While it made 14% of its ad revenue from mobile in Q3, up from 0% in February, it has to keep up with the change in how users access the site.
The increase in mobile-only users doesn’t necessarily mean people are ditching their desktops. Instead, it’s likely due to an increase in sign-ups by mobile-only users, many of them on feature phones in emerging markets like India and Brazil.
Right now Facebook shows news feed ads to these users through its Facebook For Every Phone app. However, since they aren’t downloading apps from the iOS or Google Play app stores, several of its new ad units like app install ads can’t be shown to feature phone users. Boosting average revenue per user for these segments will be a tough challenge in the years ahead.
Facebook’s key emerging markets kept growing fast in Q3. Brazil’s monthly active users (MAU) hit 61 million, up 13% from 54 million at the end of June, and up 109% year-over-year. India hit 65 million MAU, up 8.5% from 59 million at the end of Q2, and up 62% since September 30th, 2011.
Unfortunately for Facebook, growth in both those markets is slowing, as Brazil grew 146% and India grew 84% from June 30th 2011 to June 2012. One brighter spot, though. Japan grew to 18 million MAU, up 218% year-over-year.
And now onto mature market growth. While some third-party measurement services including ComScore said Facebook’s unique US visitor count fell in May, Facebook responded through its 10-Q, noting “We had 171 million MAUs in the United States as of September 30, 2012, an increase of 8% compared to the same period in 2011″, and an increase of 1.71% from 168 million as of June 30th.
That’s surely not stunning growth, as Facebook is hitting its saturation point in the US. Instead of boosting its bottom line by increasing its user count, Facebook must now look to increase the ARPU in mature markets.
Users in countries like the US, Canada, UK, Germany, France, and Australia have discretionary income, and Facebook needs to learn how to suck it out of their wallets rather than scraping it off of their eyeballs with ads (yuck, I know). Facebook Gifts, its entrance into ecommerce, is squarely aimed at growing ARPU in these types of markets. However, Facebook will have to navigate obstacles like localized shopping preferences and international shipping to bring Gifts abroad.
You can slog through the whole 10-Q below if you want:
Editor’s Note: Nir Eyal is the founder of two acquired startups and an advisor to several Bay Area companies and incubators. Nir blogs about the intersection of psychology, technology, and business at NirAndFar.com. Follow him on Twitter @nireyal.
If you’re like me, you’ve had enough of the Facebook IPO story. For tech entrepreneurs struggling to build stuff, the cacophony of recent press is just more noise. That’s why when my friend Andrew Chen posted an insightful analysis of Facebook user data, I was happy to get back to learning from what the company did right instead of debating what its bankers did wrong.
Chen calculated Facebook’s historical ratio of daily active users (DAU) to monthly active users (MAU) and the stats are startling. Since March 2009, when the earliest data is available, approximately 50% of Facebook users logged in daily.
As other technology companies struggle to maintain DAU to MAU ratios of 5% or less, Facebook’s numbers appear stratospherically high in comparison. But what is equally surprising is the consistency of that ratio over time. Despite periodic user revolts in reaction to changes in the site, the ratio remained strangely stable. In fact, the number has risen over the past year and is now hovering at 58% as of March of this year.
It’s as if Zuckerberg has steered the company by this golden ratio. Which begs the question: is there some wisdom here regarding this ratio as a predictor of Internet success? Obviously, there are no guarantees and starting cutting edge tech companies will always be risky business. But, assuming you have a solid business model, there are good reasons to believe that if there is one metric to focus on while building your business, it’s the percentage of users who come back daily as expressed by this ratio.
As I’ve written previously, I believe a mastery of the mechanics of habit design is increasingly deciding startup winners and losers. Not only because habits cement user behavior in an increasingly cluttered digital world, but because a high-engagement product is also a high-growth product. The two are one and the same. A high DAU to MAU ratio is a great indicator of the strength of user habits and, ceteris paribus, I’d bet on a business with the higher ratio over a competitor every time. Here’s why:
When it comes to web and mobile startups, high DAU to MAU is more important than the size or growth rate of an entrenched competitor. Case in point, Facebook defeated much earlier competitors like MySpace and Friendster, both of which had healthy growth rates and millions of users by the time Facebook got started.
This is because of what I call the “more is more principle.” High user engagement has an exponential effect on user growth. As David Skok points out on his blog, “The most important factor to increasing growth is not the Viral Coefficient, but the Viral Cycle Time.” Viral Cycle time is the amount of time it takes to complete a viral loop and it has massive impact on user growth. “For example, after 20 days with a cycle time of two days, you will have 20,470 users,” Skok writes. “But if you halved that cycle time to one day, you would have over 20 million users! It is logical that it would be better to have more cycles occur, but it is less obvious just how much better.”
Having a greater proportion of DAUs dramatically increases Viral Cycle Time for two reasons. First, daily users initiate loops more often – think tagging a photo on Facebook. Second, more daily active users means more people to respond and react to each invitation. The cycle not only perpetuates; with high DAU to MAU, it accelerates.
Those who talk tech split into two dogmatic camps. Some prioritize growth and accept low engagement, while others believe a company needs to nail engagement before focusing on growth. I believe this is a false dichotomy. If you have only one or the other, congratulations, you’ve got squat.
Let’s first take a look at user growth. Distribution, of course, is critically important and no company can survive without a sound customer acquisition strategy. Not only is growth essential but it is something engineer-driven companies love to work on. In fact, the title of “Growth Hacker” has recently become a badge of honor among Silicon Valley digerati. Tweaking viral coefficients and instantly seeing the results is intoxicating. It’s startup feedback at its finest.
But optimizing growth without engagement has its pitfalls. As Peter Thiel recently told his class at Stanford, the effectiveness of distribution channels tends to follow a power law. Just as businesses tend to have only one revenue stream, they also have only one good growth strategy – the effectiveness of which is 10x the results of other distribution channels. The problem with having only one real way to grow is that the method becomes obvious to others and is quickly copied. For example, in its early days, Facebook capitalized on users importing their email contact list to drive growth. But soon thereafter, so did everyone else.
But having competitors copy you is a high-class problem. It means something is working. Worse yet is discovering a fantastic viral loop that drives growth only to see engagement crater when users realize there’s little long-term value in the service. Ringtone businesses, sheep-throwing Facebook games circa 2008, and today’s social video sharing apps using questionable growth tactics, are just a few of the “leaky bucket” businesses that occur when distribution outpaces engagement.
When it comes to building a big business, clearly a good acquisition channel is mandatory, but not sufficient. Given the power law of user growth, you will likely only have one major way of acquiring customers and it won’t be much of a secret. You’ll need some other competitive advantage.
As opposed to distribution channels, the mechanics driving user engagement do not follow a power law. In fact, it is the nuances of user behavior that make the competition irrelevant, just as it did in the case of Facebook’s early rivals.
Discovering non-obvious user needs and creating accompanying habits is accomplished through deep observation grounded in solid behavioral theory, followed by methodical trial and error. It takes time to create new habits and getting the user to act the way you’d hoped is accomplished by uncovering a thousand tiny insights into the user’s psyche. The process of uncovering latent needs is characterized by understanding more about users than they know about themselves.
The distribution strategy will always be obvious, but the behavioral insights are important secrets that can only be discovered through rigorous testing. Zynga had one obvious way to acquire users, namely Facebook ads. But the company has a cadre of behavioral insights it uses to craft addictive games. It collects terabytes of information daily to alter game dynamics to boost user engagement. Quora primarily drives users to its site through Google search traffic. But the conjecture about all the reasons why the service is so sticky spills over a long question thread. Instagram posted images to Twitter and Facebook to drive user acquisition, placing its growth strategy in plain sight. However, the founders, one of whom studied psychology as a Symbolic Systems major at Stanford, acquired a deep understanding of what makes users tick and click.
But why can’t behavioral design be copied like a distribution strategy? Because competitors are not able to recognize and act upon these kinds of insights. You can know the competition’s product feels better to use, but you won’t know why. Engaging products gain their advantages by leveraging tiny improvements, which together create huge advantage. From the outside, you can’t tell what’s working and what isn’t.
For example, the iPhone is objectively a better designed, more user-friendly, and ultimately more engaging product than the Android experience. But why? Nearly everyone, when given the choice between an Android interface and an iPhone, chooses the iPhone. There are plenty of good reasons to own an Android, but intuitive interface ain’t one. Google knows this and yet they can’t replicate Apple because they don’t know the answer to “why?” You can’t make decisions between seemingly identical interface choices unless you’ve walked the path of user behavior. Without this knowledge, copying the competition becomes a game of throwing darts at features.
Habit design requires a fundamentally different, though complementary skill set to growth hacking. Designing high-engagement products is an art which is increasingly becoming a science. The craft crosses the disciplines of psychology and design – both fields which are hard to learn in a short period of time. Unfortunately, designing habits often falls in the organisational abyss between the founders’ vision and what is technically feasible.
But those companies able to habituate users quickly enjoy massive advantages. Not only does engagement drive growth for the reasons stated above, but users tend to shut out other, sometimes superior, solutions. In fact, business history is peppered with technically inferior products beating competitors because of the fierce loyalty of habituated users (I’m looking at you Apple addicts). Users only have time and brain cycles for a limited number of services. If a high proportion of users are using your service daily, they aren’t using the competition’s.
But focusing on engagement without growth is also a losing proposition. For one, virality is not something that can be bolted on to a product after it is in the wild. Distribution is not an afterthought and it needs to be built into the core of the experience. Either the company has a viral growth mechanic or it doesn’t. So no matter how engaging your service is, it will remain niche unless there is a way to get it in front of new users en masse.
Creating a company with both high engagement and high growth requires a sound distribution engine fueled by active users. Both engagement and growth are essential to a company’s viability and by adhering to the tao of DAU and MAU, founders have an accurate point of focus to increase their odds of success.
Read the original: Never Take Your Eyes Off This Hacker Metric
Ask a VC is back this month (finally!) after a long hiatus. This week we have a freshly-minted VC, Nabeel Hyatt, of Spark Capital. Hyatt just joined Spark after 15 years of starting and building companies. His most recent company was social gaming outfit Conduit, which became Zynga Boston after it was acquired.
We have a couple of questions from readers — one from Nicky, who asks whether the VC model is “broken”, and another from a reader Alex, who asks about how the landscape for user acquisition is changing.
User acquisition is Hyatt’s home turf. As a general manager at Zynga, which is the most data-driven company on the Facebook platform, he had to know the arc and decline of social games like the back of his hand. This is super-useful in light of the explosive growth that apps like SocialCam and Viddy are seeing on Facebook and iOS.
What’s happening right now is that Facebook is finally becoming a potent force in mobile app distribution, as Hyatt explains in the video. A year ago, if I talked to top free or grossing developers, hardly any of them said Facebook was an important channel for acquiring users. They were too addicted to paid channels like Free App A Day, offer walls or even download bots. Several of these channels are now banned by Apple, which is making more room for apps that are genuinely engaging or are getting users virally.
Secondly, Facebook is realizing that if it wants to stay relevant in a mobile era, it has to step up its ability to push traffic to mobile apps, Hyatt says. A year ago, they were bent on bypassing the native app route and pushing the ecosystem toward building with HTML5 instead. But over the last few months, they’ve stepped away from that rhetoric and are now driving traffic to both native and HTML5-based apps.
With the new mobile platform, Facebook is now able to single-handedly drive apps to the top of the free charts on iOS — something it didn’t do a year ago. That pushed apps like Viddy and Socialcam to the top of the charts over the past month. Now the question is what does this mean? How should these companies be valued? Are investors being sophisticated enough about how to value them?
Hyatt says working at Zynga has given him many insights about what to look for. Namely, downloads and even daily active users aren’t sufficient enough metrics for judging apps.
You have to ask for other numbers. For example, he’ll look at one-day retention: how many users come back the second day? And he’ll look at seven-day retention: how many users come back seven days later? He also looks at DAU/MAU, which is the ratio of daily active users to monthly active users. It’s a measure of stickiness: out of all the users that touch an app every month, how many come back every day?
If you look at Viddy and Socialcam, this DAU/MAU metric is looking troubling on both of them, according to AppData. (Though, keep in mind, there have been some reporting errors over the past week.) He said that very few games can thrive if they fall below 12 percent on this metric. Even though Farmville is about three years old, it’s still hovering at around 19 percent. Viddy has been below 10 percent for more than a week and SocialCam just dipped below that key level. This is a negative sign.
That said, when an app gets a slug of growth like this, the peak number of users is not actually that important. It’s the number of users that it keeps engaged for the following several months.
Here is the original post: Ask A VC: Spark’s Nabeel Hyatt On Hyper Growth And Whether Facebook Is Fueling Future Rivals