Tag Archives: Facebook

The Death of the 3rd Party Link on the Facebook NewsFeed

Viral Traffic from Facebook DeclinesAnyone who has been in online publishing (I just had a stint at Playbuzz.com running their owned and operated sites) has experienced first hand the decline in traffic coming from Facebook. Digiday recently cited SimpleReach showing drops of 32 percent or more since the beginning of the year.

Let’s be clear, third-party links aren’t dead – they are just de-emphasized in the typical user’s feed depending on how frequently a user visits and interacts with friends and brands (publishers). A Facebook user’s feed is customized based on the recency and frequency of their interactions with friends and brands as well as time spent looking at the content those friends and brands produce.  But third-party content – stuff created by publishers and brands that users end up sharing – definitely has been de-emphasized over the last year.  In reaction to what Facebook saw as abuse of their algorithm and a lot of click-bait driven tactics by publishers, they changed the algorithm to favor friend posts over page posts. Including all the other changes in the last year, you can paraphrase Facebook’s newsfeed strategy as follows:

  • Higher emphasis on things posted directly by users (status updates, photos shared, check ins, events, Instagram images)
  • Higher emphasis on things posted directly ON to the network by publishers (video and instant articles)
  • De-emphasize the weight of things posted 2nd-hand (the traditional links to articles on publisher web sites)

Remind anyone of the FarmVille days?

Even though I knew all that conceptually, it wasn’t until I stepped away from my newsfeed for over 24 hours that I really began to understand it.

So here’s the experiment: I (some would say insanely) did not look at my Facebook feed for over 24 hours. I then went through the first 59 posts in my feed on my mobile device. Then I switched to my desktop and reviewed the first 60 posts. There were no duplicates – Facebook knew I had seen the 59 items on my phone and did not include them when I switched over to my desktop. So nearly 120 posts later, there were some really interesting insights:

  • The majority of my first 60 posts were friends status updates, photos, videos (I classify Facebook-owned Instagram reposts as not a 3rd-party link) — only 5 in my first 59 posts were 3rd party link/video posts.
  • In fact, the first 3rd party link I came across was the 35th post in the queue – relatively low down
  • I saw four of the in-line ad blocks before coming across a single third-party link
  • Ads show up neatly every 10 posts, so we can actually look at the percentage of 3rd party links based on the set of 10 posts – and that’s pretty illuminating:

EVCinNYC posts linking to third party content are relegated to the bottom of the Facebook Newsfeed Queue

  • Clearly third-party links only begin to show up as the number of user-created posts begins to thin out. Publishers that are relying on Facebook automatically get pushed to the back of the line and then they fight amongst themselves based on the popularity of their links

As a publisher, it’s clear that the days of viral within the feed have changed dramatically as 3rd party links are de-emphasized over user-created content. You can see that ads are the only way to break through all the user-created posts at the top of the queue – a boon for Facebook, but a real jolt to publishers who don’t have a ton of direct traffic and are looking to build their audience.

For the geeks – the raw data

Listed below on each line is the type of post I saw in my feed from the top to the bottom.  Items in bold are items that I deemed as items shared as third-party link, video or photo.

ON THE MOBILE DEVICE
status
ad
status
status
photo
status
friend posted video to a group I’m in
status
status
Instagram post
ad
status
status/photo
status
photo
photos
status/photo
checkin
status
status
ad
friend posted photo
friend posted photo
status/photo
check in
“listening to” a song post
status
status/photo
friend tagged in a photo
photo
ad
photo
status
friend tagged in a photo
gothamist article shared
Cards Against Humanity link shared
friend posted photos to a group I’m in
status/photo
status
friend shared memory
ad
status/photo
image shared by xkcd.com
image
photos
photo
status
status
shared image from a FB page
ad
soundcloud link shared
friend tagged in photo
photo
photo
2 friends share a post from emilycheath.com on starbucks red cups
photo
friend tagged in photos
friend updated cover photo
ad

THEN TO THE DESKTOP

starbucks groupon shared
ad
wtvr.com article share
nerdist.com article share
huffpo.com article share
brooklynvegan.com article link
notrightinthehead.com – image share
friend changed profile picture
businessinsider.com article share
ad
friend tagged in photo
photos
friend interested in an event
instagram share
housebeautiful.com article share
status/youtube video share
oculus.com link share
shared mashable.com link with status
ad
friend changed profile picture
someone commenting on my post
memory
dailymail.co.uk article share
video posted in group I’m in
friend posted Nova 106.9’s video
friend going to an event
video shared by two friends
theatlantic.com article shared
share of someone’s twitter handle
ad
Missy Mwac video shared
Michael Franti & Spearhead video shared
link shared
link shared
Isha Foundation photo shared
Reddit.com article shared
Collegehumor.com video shared
ad
Paolo Tuci video shared
newyorker.com article shared
event shared
video shared
video shared
npr.org article shared
image shared
awesomethings video shared
daily graphic video shared
ad
photo
xkcd.com article shared
montagne-en-scene.com article shared
video shared
friend’s birthday
updated cover photo
mlb.com article shared
video shared (3rd party)
photo
photo
ad
youtube shared

From Draw Something to $40 million TV Ad Campaigns

OMGPOP's first Facebook game was Cupcake Corner
OMGPOP’s first Facebook game was Cupcake Corner

I’ve spent the last four years on quite the ride, so pardon the long absence. I jumped into OMGPOP in 2010 as the product lead for a Facebook game; we delved into mobile apps in 2011 as the market shifted and then I was fortunate enough to be part of the crazy explosion of our game Draw Something. After getting acquired by Zynga, I jumped crosstown to FreshPlanet to grow and monetize SongPop as well as try to develop more social causal mobile hits. And here in 2015 I’m watching the market shift yet again as mobile game developers are now turning to traditional brand marketing by spending tens of millions on TV ads.

So here’s a couple things I’ve observed along the way:

#1 Overnight success takes years to perfect

Two of the biggest hits I’ve been associated with (Draw Something and SongPop) came after years of iteration.

Draw Something started off as a real-time flash game Draw My Thing on the now defunct OMGPOP.com site (but you can still play the embedded game on sites out there if you are looking for it). Drawing with your mouse sort of sucked but it was fun in real time – you had to beat the timer and draw well enough that others in the group could guess it.

Before Draw Something on mobile, there was Draw My Thing on OMGPOP - a synchronous game where you had a timer to draw and other users tried to be the first to guess the word you were drawing
Before Draw Something on mobile, there was Draw My Thing on OMGPOP – a synchronous game where you had a timer to draw and other users tried to be the first to guess the word you were drawing

We ported the game to Facebook which greatly expanded the audience, peaking at about 2.1 million Monthly Active Users (MAU) – but that was a mere blip compared to what we saw when we jumped to mobile. The iterations on mobile were huge: In 2012 doing real-time mobile play wasn’t an option so we shifted into asynchronous game play; the clock and “winning” or “losing was eliminated; and typing out your guess on a mobile device was a pain, so we shifted to a sort of scramble-like listing of letters from which to make your guess.

Before SongPop, an initial iteration was The Crazy Cow Music Quiz
Before SongPop, an initial iteration was The Crazy Cow Music Quiz

SongPop too had an earlier incarnation – can you believe The Crazy Cow Music Quiz? Obviously we ditched the cow, and a bunch of pre-game power ups, opting for a simpler and more direct game play.

Neither of these game ideas were a success overnight. And that’s about the same for every indie darling that makes it big, the latest being today’s #1 hit Trivia Crack by Etermax. Trivia Crack was built on the success of its Spanish-language version predecessor Preguntados which was built on the success of it’s Scrabble-like game Adworded which was built on the company’s past experience as a third-party developer.

#2 Sometimes there CAN be too much of a good thing

People binged on Draw Something similar to the same extent people now binge on NetFlix - games now have to figure out how to not let the player get sick of the game and burn out
People binged on Draw Something similar to the same extent people now binge on NetFlix – games now have to figure out how to not let the player get sick of the game and burn out

When Draw Something first came out, it was like crack. People couldn’t get enough of it. They played non-stop, during class, over night – it was this incredible social binge event. But unlike binging on Breaking Bad episodes on NetFlix, when you were “done” on Draw Something, there were a ton of opponents waiting for you to draw back. Unlike Scopely’s Dice with Buddies where a round is literally a couple seconds, Drawing took quite an investment of time and thought. That’s cool with three or five of your close friends, but having to draw for 50 people gets a bit overwhelming.

When Zynga bought us, Draw Something was barely a month old and no one had a clue what the eventual retention curve would look like. I won’t second guess anything we did in Draw Something to become such a cultural phenomenon, but today social mobile games put a lot of effort into gating the content a bit, to get users to stay a while instead of binging and leaving.

SongPop limits the number of opponents you can have at any one time. Many games like Two Dots or Candy Crush have “life” systems – you lose lives when you fail a level and either have to ask friends, wait or pay to regenerate those lives and keep playing. Trivia Crack sort of combines the two – limiting opponents you can play by a “life” system. The balancing of these gates and when they appear are key both to monetization, but also regulating the consumption of your content.

#3 Deja vu all over again

When I was younger, I remember a movie would stay in the theaters for weeks. Today, most movies are gone in a flash — maybe two weekends at the cineplex and then gone. But hits last longer – just not as long as they used to. In 1977 Star Wars was in 40% of it’s max theater release for 29 weeks. For last year’s Guardians of the Galaxy, it was in 40% of it’s max theater release for only 10 weeks. The economics have changed – there are about 4x as many theaters today so more movie-goers can see them in the first couple weeks. Looking ahead it’s becoming more clear that theaters will eventually give way to direct-to-home streaming.

Hit games like Luxor would remain at the top of the charts for six months before the downloadable PC game market got saturated
Hit games like Luxor would remain at the top of the charts for six months before the downloadable PC game market got saturated

When I was marketing downloadable games in 2005, a hit game like Luxor was at the top of the sales charts for six months – a game we all called a “AAA” game back then. Within two years the top selling game was at the top of the charts for just two to four weeks. The economics changed – there were 3x as many games being made. And then this little platform called Facebook started making it easier for users to get free games instead of paying $9.99 a pop for the downloadable game.

Are we seeing a similar trend in mobile? Saturation makes it hard for a new game to get heard. The majority that do break through — with either a burst campaign or the lottery ticket of an Apple feature — don’t end up lasting long on the charts.

Developers in the download space tried to pivot to Facebook games, but it was the early adopters on the platform like Zynga and Playfish that were able to take advantage of looser viral channels. Few other developers were as successful than those early entrants on the Facebook platform.

Are we seeing a similar trend in mobile? Over half of the top grossing games (Clash of Clans, Candy Crush, Big Fish Casino, Hay Day, Soltomania) were released in 2012 or earlier. In 2012 there were 500,000 apps to compete with in Apple’s App Store – today there are 3x as many. The companies that were able to establish hits in or before 2012 have been able to maintain those franchises as early adopters on the platforms.

King.com has saturated select cities with branding campaign for Candy Crush Soda Saga - including the tops of yellow cabs in New York City
King.com has saturated select cities with branding campaign for Candy Crush Soda Saga – including the tops of yellow cabs in New York City

So while we all know it’s been getting more and more expensive in the last two years to create a new app (more depth and polish required) and acquire new users (CPIs easily can climb over $3), the money accumulated by these early adopter developers have allowed them to grow and move to an entirely different level: full-fledged brand marketing (cue Kate Upton). With city take overs by King.com and apps buying commercial time during NFL football playoff games, these are the marketing tactics that a small to mid-size developer can’t even begin to compete with.

So what’s an indie developer to do?

Well the great thing is that based on #1 above, there’s always room for iteration and innovation. When you can provide a unique experience, tell a unique story, change the way you interact with a device, then you might be able to find that big hit. Just realize you might have to fail 20 times to get there.

Second, be on the look out and try different platforms. Facebook disrupted the downloadable/PC games market and created new developer power houses. Mobile disrupted the Facebook games market hierarchy and has created new developer power houses. Eventually new platforms will come, we just don’t know where. So innovation and iteration is key for the indie developer to find a new market where they can be successful. Can someone crack Instagram and Twitter to create a new mash up of game play with social? I’m betting they can.

But this is my big question going into 2015: Can the medium-size developer shop survive? Or will we continue to see consolidation with a bunch of power house developers and a lot of small 1-3 person teams trying to create something new and unique?

Zynga Making 4.4 Cents per DAU

Zynga ARPU of 4.4 centsZynga filed their S1 last Friday, and while some have focused on how rich CEO Mark Pincus and his officers could become, I’ve been looking at the financials and trying to tie it back to their daily active users (DAU) to look at revenue per DAU or ARPU (Average Revenue per User) so we could have some industry benchmarks.

Revenues

The S1 shows gross revenue by quarter as follows:

Q3 2009 – $31,311,000
Q4 2009 – $55,721,000
Q1 2010 – $100,927,000
Q2 2010 – $130,099,000
Q3 2010 – $170,674,000
Q4 2010 – $195,759,000
Q1 2011 – $235,421,000

Some interesting notes on this revenue figure:

  • This revenue includes everything, including advertising and virtual goods, but advertising has never made more than 5% of total revenues
  • The revenue figures are impacted by some accounting rules as to when the revenues can be recognized. Items that are consumables (like fuel or energy or components that are purchased and used) can be recognized within a month, while durable items (like tractors or décor that are used over time) take a longer time to be recognized (it isn’t specified in the S1, but it typically is amortized over the life of a user – for more discussion to revenue recognition, see this blog post from Bill Gurley). Zynga was only able to discern consumables versus durables starting in October of 2009, and only in one game. By January 2010, they were able to reflect this split in all games, meaning they could escalate revenue recognition. Thus it is likely that this new ability helped improve the revenue in Q4 of 2009 and Q1 of 2010. It also provides an incentive to focus games more on consumables than durables moving forward.
  • Facebook Credits were implemented in their games, starting in July 2010 and completed for all games as the sole purchase currency in April of 2011. Facebook Credits take 30% of gross, and since the full implementation didn’t occur until the 2nd Quarter of 2011, we won’t see the true dampening impact on Zynga revenues in the S1 (which only covers through Q1 201).

For our purposes, I will divide the quarterly revenue by the number of days in the quarter to come up with an average revenue per day.

Daily Active Users

DAU data by game is available back to June 2009 via appdata.com, so we can at least come up with an average DAU for a quarter from Q3 2009 forward. I am not including numbers from RewardVille and the Zynga Games Bar as they don’t appear to drive revenue directly (there may be search revenue) and these are peripheral to the games (one would not go to either if they did not play at least one game).

DAU per quarter:

2009 Q3 -24,309,425
2009 Q4 -56,757,183
2010 Q1 – 66,082,862
2010 Q2 – 58,012,018
2010 Q3 – 48,975,539
2010 Q4 – 47,631,850
2011 Q1 – 59,082,981
2011 Q2 – 52,447,745

These numbers are about 2% below what Zynga reports in their S1 so it may be that Zynga is adding some DAU from platforms outside of Facebook, but that figure is relatively small (3 million at most).

The caveat of course is that the DAU figures of several games doesn’t provide you DISTINCT DAU (e.g. Zynga showed 59 million total DAU across their games in Q1 2011, including 11.7 million in FarmVille and 19.4 million in CityVille. But many users could be playing both games and thus over-stating the distinct DAU across both games). Zynga did not report distinct DAU in their S1.

So what our DAU is showing is not distinct DAU, but number of game players. Whether the user plays one or five of Zynga’s games, the Revenue divided by DAU number here shows the average revenue they see per game player.

Since Q4 2009, Zynga’s aggregate DAU is relatively flat, while Facebook’s US audience has nearly doubled in the same time period:

Facebook doubles US users since Q4 2009

Of course that FarmVille peak happened in the era where developers could blatantly leverage the notifications API on Facebook. You can pretty clearly see the results now of Facebook turning off those Notifications on March 1, 2010 when you look at it by game:

Zynga DAU By Game

From the chart above, you can see FarmVille took the biggest hit and successful games like FrontierVille couldn’t grow enough to make up the gap as users continued to decline.

CityVille gave a nice bounce in the Q1 2010, helping boost Zynga’s aggregate DAU out of the upper 40 million DAU range for the previous two quarters (the game was touted to be the first launched simultaneously in French, German, Italian and Spanish and helped it to reach 100 million MAU very quickly, but the DAU never reached that of FarmVille at it’s peak).

These numbers appear to show that there is an upper bound on the number of users playing games – that getting consistently 50-60 million aggregated game players (again, could be 20 or 30 million distinct users playing multiple games) is difficult. Even with successive releases and fast-growing games like CityVille, Zynga has not been able to get an aggregated DAU that matched the peak of FarmVille in Q1 of 2010. The impressive gains of CityVille users seems to come at the expense of their other games, like FarmVille and FrontierVille.

Faced with the prospect that there is a ceiling to the number of aggregate users, you can possibly see a benefit in the recent pattern emerging of game releases every six months (FrontierVille June 2010, CityVille December 2010, Empires and Allies June 2011): Extract every dollar from the six-month life-cycle of a game before releasing the next game. Much like game developers learn that on the Facebook platform you need to keep providing your hard core players new content and features to retain them, Zynga is playing it’s own meta version, releasing new games on a steady pace tied to when they see their users beginning to drop off.

Aggregated Revenue per DAU

Thus the revenue per DAU by quarter shows that for the average player of a game, they’ve hit 4.4 cents in Q4 2010 and Q1 2011:

2009 Q3 $0.0145
2009 Q4 $0.0108
2010 Q1 $0.0172
2010 Q2 $0.0245
2010 Q3 $0.0378
2010 Q4 $0.0447
2011 Q1 $0.0443

Graphically you can see the trend here:

Zynga Estimated Revenue per Game User

So reviewing the factors that have impacted this growth number:

  • Revenue recognition of components that would have improved revenue starting Q4 2009 and fully implemented Q1 2010. Any games coming out moving forward that focus on consumables would like drive more revenue shortly after release.
  • Reduced virality in games due to changes in the platform (such as removing Notifications on March 1, 2010 and a severely dampening wall posts in September 2010). This actually may have left games with only their more hard core users and thus better monetization, but also resulting in a shorter lifetime of games as it became more difficult to back-fill those users that churn off the game.
  • The impact of Facebook Credits integration – a dampening impact from Q3 2010 to Q1 2011 but whose full impact we wont see until Q2 2011, but that could explain why the successive quarter-to-quarter growth in ARPU declined slightly in Q1 2011 after four successive quarters of growth
  • The success of game releases in driving better revenue per user and the percentage of aggregate DAU that each game makes. It’s well-known that Zynga optimizes everything and there is learning from game to game. In addition, if a larger percentage of your game players were playing Café World or YoVille (arguably poorer performers because they predominately focus on decorations which are durables and the difficulty of getting users to redecorate), then that impacts the aggregate number. Migrate Café World and YoVille daily users to newer games like FrontierVille and CityVille (which have more consumables and mechanics that drive a user to pay), and that can possibly increase the aggregate revenue per game player figure as well.

Is 4.4 cents per game player per day the holy grail for a mass market Facebook game (as opposed to RPGs which have smaller audiences but higher revenue per player)? Getting users to play every day and across multiple titles is key to driving customer lifetime value and insight into those numbers are lacking from the S1. Their most visible efforts at cross-game promotion and retention (RewardVille getting 1.25 million users per day) suggests that about 2.2% of users are engaged each day).

With Zynga DAUs declining in Q2, will they be able to match the quarter over quarter revenue growth? Will their latest game Empires and Allies, with a slightly more RPG style, provide enough additional revenues to make up the DAU gap? And if it doesn’t, will Zynga push ahead another release before December?

We will have to wait for more quarterly reporting to get a feel, but for now the 4.4 cents per day per user is a handy benchmark.