Image: Flickr - Greenchameleon

I’ve been thinking about this while chewing hard on an interesting nugget from the recent ‘Anatomy of Facebook’ study published by the Facebook Data Team and collaborators.   (You can find a summary of the study here and a link to the full article download here.)     The team selected five different age bands in the Facebook population, and for each of these five age cohorts, they looked at what age all that person’s Facebook friends were.

It’s the oldest Facebookers in the sample which have the widest variation in age range within their circle of FB-friends, and the youngest who have the narrowest.

Figure source: The Anatomy of Facebook - summary (Facebook Data Team)

Does this mean that if you want to achieve the widest potential reach for a message,  your best bet is to target 60 year olds to spread it?   That would be delightfully unexpected, if  true.  But I think it isn’t.

Let’s take a stroll together through a sanity check.  For every 60 year old who’s FB-friends with a 20 year old, there is a corresponding 20 year old in the data who’s FB-friends with that very same 60 year old.   For 60 years, the 20 year olds are noticeable, as a part of their friendship circle.   But,  when you look at the age distribution of FB-friendships for 20 year olds,  60 year olds form a vanishingly small part of their FB-friendship circle.

How can both these things be true?  My guess about what’s going on is that the 20 year olds tend to have many, many more friends than the 60 year olds.  I think that’s the main way in which 20 year olds could be important in 60 year olds’ FB-friends’ age distributions, but 60 year olds are not important (numerically speaking) to 20 year olds. This is just a guess but it’s my best guess about how these two facts could fit together.  If you have other ideas let me know what they are!

What’s the consequence of this pattern of connectivity?   We all know that raw connectivity and influence are not the same thing.  If you’re feeling digressive and a bit geeky here’s a fun paper on the topic of how network structural characteristics affect viral distribution patterns by Kitsak et al.   (Have fun but come back soon.)   But without connectivity, there is no path for influence to propagate.   So connectivity is  interesting.   It’s just not the only thing that’s interesting.

Considering the universe from a path connectivity point of view, what can we say about the relative potential power of 20 year olds vs 60 year olds?   From the point of view of outbound communication,  we can guess that as a 60 year old FB-friend it’s probably pretty hard to get the attention of the 20 year olds you’re connected to.  Speaking purely in connectivity terms, you have to fight for attention against all those inbound comms channels from all those 20 year old age-mates.    Your input is one of many.

But look at the information flow from the opposite perspective, and a tantalising possibility emerges.   The 60 year olds have the most broadly balanced feed, in terms of the age range of their information sources.   They will be better listeners, as their mixing deck is better adjusted to a wider range of signals from reality.   They will know a greater diversity of things.  They will be wiser, for structural reasons.   If you think knowledge is power then you should bet on the 60 year olds.   But I bet you knew that already.

Computing the network metrics characterising this population provided employment for 2,250 servers, deployed in a Hadoop cluster.  Awesome™.  Some of it was done on a  a 64 GB machine via a stream-based algorithm, and some on a 24-core 72 GB machine via a novel algorithm.   (Still Awesome™.)

Why did they look at all their data, when doing so was pretty difficult?   Personally, I think they fed their whole active user base into their crunchers partly because they could.  Facebook is not an ordinary engineering company, and it is not just an engineering company, but it is definitely an engineering company.

But the reason given in the text is different:

Network completeness is especially important in the study of online social networks because unlike traditionals social science research , the members of online social networks are not controlled random samples, and instead should be considered biased samples.(p.2)

Now, that is really interesting.    It is interesting not because it’s flawed, but because of the way it’s flawed.   Making the Facebook ‘sample’ as big as humanly possible doesn’t help, if what you want to talk about is the structural characteristics of human social networks.  The whole population of Facebook is a biased sample if you view it as a sample of human social relationships.  The logic of the authors’ assertion that you need to look at the whole population because using just a sample is biased is wrong.   Even if you use the whole population, the entire Facebook universe is still biased as a sample if what you want to talk about is ‘human social networks’.

However, if what you want to talk about is the nature of Facebook-friendships, being able to look at the social graph patterns  from the whole population of active Facebook users is just amazing.   (And, of course, Awesome™.)  But there’s simply no need to confuse Facebook with reality, and say the reason you are looking at your entire data set is to avoid sample bias.  Looking at the whole shooting match doesn’t help.

The authors go on to say:

…the most accurate representation of our social relationships will include as many people as possible.  We are not there yet, but in this paper we characterise the entire social network of active members of Facebook in May 2011… (p. 2)

Again,  a pivotal bit of confusion revealed by choice of language.  The study reveals the structure of Facebook-friendships, and Facebook-social-networks.    It does not study “the entire social network of active members of Facebook”.

The question of how people form and maintain relationships using Facebook is a  truly fascinating one.  So, too,  is the question of how the design of Facebook facilitates and influences these fundamental human processes.   The answers are important for a variety of  reasons, theoretical and applied, personal and commercial.  They are important to a huge range of stakesholders in the system in addition to Facebook itself.  So I’m really glad Facebook thrashed thousands of servers to do this study, and I’m gladder still that they published it.   I am looking forward to cherry-picking some of the tastier findings in future posts.

But I will always have social connections who aren’t connected to me on Facebook.

Before I went, I made some excited noises about the Sorrell talk because my guess was that he was going to be talking about “social game TV”.    Just think.   If social TV is hot and social games are hot then just think how hot “social TV games” could be.  Ooh.  (And natch it could also be utter crap, if you do it badly.  But the upside is real.)

Only one hitch: I was wrong, wrong, wrong.    Not that a social game layer on top of TV wouldn’t be awesome™.  It could be.

But social TV games just weren’t the tack Sorrell took in his talk.   He did a pretty  straight rant-flame about what games can learn from TV.   His take, as I understood it, is that TV is way more successful than games in terms of the value of the industry, and the number of eyeball-hours, probably always will be unless games get their act together and listen up to the following words of wisdom:

• TV understands how to use familiar music to cue emotions – games are pretty rubbish at this by comparison
• TV uses story loops well, if games combined story loops with rat loops [i.e. compulsion loops] nobody would ever leave the house again
• the games industry will never succeed until it hires more women, 15% isn’t enough
• TV has cracked the recommendation engine problem, it’s called channels, and 80% of shows are still watched this way, real-time at broadcast time
• there’s a special experience to do with live events, although ‘live-ification’ is a  word only used by particularly horrible TV execs
• no game justs gives you a task with no choice, but turn on the TV and you’ve done everything you need to do, and passive is good.

All very interesting, in an intentionally controversial but still content-ful way.  Just not what I was expecting.

So, in the question period I asked Sorrell what he thought of social TV.    He rolled his eyes a bit,  in an “oh dear oh dear not that question again” kind of way, but said that if anyone was going to do it,  it would be  Zeebox, one reason being that they allow unofficial apps.

Les jeux sont faits….

Image: Plindberg - Flickr

Mark Sorrell from Screenpop says games can learn from TV shows like the X-Factor. Tough talk considering the shift we’re seeing away from TV and towards Facebook.

According to recent research, people in the UK spend more hours on Facebook than they do watching telly.   The  writing is on the wall.   Facebook is stealing time from TV.    And Facebook is, in its own special way, a games platform.   So what can TV teach games?  At first glance, it looks like the score is: Facebook+Games 1, TV nulle points.

But I think Sorrell is worth listening to.   For one thing, Fremantle Media, Screenpop’s parent company,  is a leader at producing and licensing insanely popular TV shows and formats -  think X Factor, the Idol series, Britain’s Got Talent.  And more.  (I’d go on but we don’t have all day.)  Not only that, but Freemantle have built and bought a place in the TV-games tie-ins segment, via their innovation centre Screenpop, and also via recent acquisition  Ludia, who make cross-platform games for a starry roster of mass market American TV shows which again I don’t have time to list and you don’t have time read about.   Montreal-based Ludia is set to hire 100 new staff over the course of 2012.

Sorrell will be speaking at the Evolve games conference in London on December 1st next week.    And I’m looking forward to what he has to say.    (If you’re thinking of booking, don’t forget that you can get a 20% discount by quoting discount code ELELTE.  Here’s why.)

Why?  Sorrell is well placed to know what he’s talking about, and I think what he’s talking about is interesting.  The ‘TV meets games’ market is set for some really rapid evolution in the short term.  There are two things going on in the market which are responsible for this:

•  the general shift in viewing habits towards simultaneous multi-media consumption
•  the sudden appearance, stage right, of an 800 lb gorilla (not King Kong – Facebook)

The story that’s unfolding about Facebook and TV is more complicated than can be summarised via a leaderboard.   While Facebook has clearly stolen time from TV, what’s even more interesting in terms of market trending is the fact that simultaneous media use is becoming the norm.   Last year’s US Nielsen survey, sponsored by Yahoo!, showed that about half of viewers regularly use the TV and an Internet connection at the same time.  Interestingly, as of yet, this Internet use and TV use is mostly unrelated.   But it needn’t be.    There’s a clear opportunity here to increase engagement with programming through online participation including game-play.    This opportunity has a very particular shape.   Lots of games simply won’t fit the bill.   But some will.

The other plot line that’s unfolding has to do with the gorilla.  The 800 lb gorilla.   With its recent set of announcements at the F8 developer conference, announcing partnerships with Netflix and Hulu,  Facebook is extending itself into a media consumption platform.   I think it’s well placed to make a go of it.   For now,  the impetus is on social-graph-powered content discovery.    But game tie-ins will not be far behind.   Watch this space.

What’s next for social games?    The only certainty is change.   (Plus ça change!)

Yazino’s CEO‘s Big Idea is that the future of social games is synchronous – that’s what he’s going to be talking about at Evolve.   In the background, I hear a chorus of angels singing a song about the next big thing.   Do you?  It goes like this:  Oh dear oh dear oh dear.  Not another big thing.   But – here’s the thing.  He’s partly right.

‘Classic’ social games have some prototypical features: they are played via social networks; use players’ friend networks as distribution channels; and hook in to deep-seated pre-existing social patterns of display and exchange.    They aren’t synchronous.  Player to player interactions in a game are, in general,  non-blocking. Multiple social exchanges run concurrently with the main plot line.    While inter-player interactions facilitate game play, they are designed so they can occur in parallel with the main player flow, which is about interaction between the game engine and the player.

One shift we have already seen in this ‘classic’ model for social games is the emergence of turn-based adversarial player-vs-player games.   This isn’t a new model.  Quite the opposite.  To date, most of the successful titles in this budding genre have been reworkings of board and card game classics, such as Poker and Scrabble.  Plus ça change indeed.   Turns act as a strong synchronisation mechanism for play, but there is no realtime synchronous dynamic to the game play.

Tolerance for lag between turns

I recall being told quite flatly that turn-based games were not ever going to take off in casual social games, as people would not want to wait for their turn.   Wrong.   But – more interestingly – why wrong?  What I think is going on here is this:  there is a relationship between the time and effort required to make a move, and toleration for lag between moves.   Where your own moves are easy to make, and your adversary’s moves have little impact on the game,  having a big lag between turns just doesn’t work, as shown in the Figure.  The other move isn’t worth waiting for, and your move doesn’t need a refractory period.   This is in the ‘not ok’ zone.   But there is an OK zone.

You could make a case for parallel asynchronous interaction in browser based social games having evolved as a creative response to the technical limitations which are fundamental to browser architectures.  But these limitations don’t bite so hard as they used to, as other application architectures have creepingly colonised the browser model.  And the commercial incentives are motivating people to put effort into jumping as high as they can, to try to clear those hurdles that do exist.

So I believe, alongside Yazino’s CEO, that we can expect to see more synchronous social play.    What will the impact be?  Synchronous play, as opposed to turn-based play or parallel play, opens up new design spaces.  It can be more:

• realistic
• immersive
• challenging

The way I see it, it is more suited to ‘traditional’ realtime competitive engagement, whether team-based or pairwise p2p.   It is less suited to symbolic operations.    Will it take over the world?  My own view is that this is a ‘not only but also situation’.    There are games for which having the concept of turns, and the ability to take time between turns is beneficial.   For these, truly synchronous play is not a requirement. I’d use the word synchronised, rather than synchronous.

For a 20% discount to the Evolve in London conference, use the code ELELTE when booking.

In case you haven’t read my other post – I’m acting as an Ambassador for the Evolve Conference in London Dec 1st, where Yazino’s CEO is speaking.    As a part of this role, I’m doing some commentary on what I think the most interesting bits of the agenda are.   I don’t agree 100% with the view that synchronous will dominate social.   But I’m looking forward to learning more about it. Let me know if you’re coming!

Use the code ELELTE when booking – this gives you 20% off!

I am an Ambassador for the conference: my remit is to make noise (any noise I want to), and in return, I am being given a complimentary ticket.  This is a win for me.  I am very keen on the programme, I like saving money, and I like making noise.

My own personal  top picks are:

• Synchronous and Social: Why Games That Connect People in Real-time are the Real Future of Social Games, Hussein Chahine Founder and CEO, Yazino

• What The X Factor Can Teach Games,  Mark Sorrell, Head of Games, Screenpop

More later on WHY they’re my top picks.

• Facebook’s move to socialise media consumption is a huge direct threat to Apple’s content business, and an indirect threat to the hardware business that Apple’s content business enhances
•  Facebook’s separation of status updates into Meaningful and Ephemeral, combined with other changes,  make it more like Twitter.

There is more than one story to be told about Facebook’s relationship with Apple.  There are ways in which Apple and Facebook are symbiotic.    But make no mistake, they also compete for my attention and my spend.    Yours too.

I think Facebook holds a better hand, because, together with its partners,  it is so strongly placed to do utterly amazing things with content discovery by making use of social graph inference.   (And integrated billing can’t be a million miles away either.)   But Apple is clever, commercial and creative.    This is one story that will run and run.  I don’t know how it will end but it will be worth watching.  As Sheryl Crow says:  “sit back, enjoy the show”.

Twitter is another matter.  Facebook has been moving, amboeba-like, to embrace Twitter’s functionality, while still being Facebook  and extending into new areas (e.g. TimeLine’s LifeBook).     Facebook’s introduction of uni-directional subscription based communications is an important element of this creep, as is the UI change which make it easier to control publication privacy in a granular way, per post.    Both of these are pre-F8.    The F8 “icing on the stake” (the one pointing at Twitter’s heart) is the separation of status updates into small more trivial and ephemeral updates, which go into your ticker display, and more important stuff, which makes it into your timeline and your friends newsfeeds.   (I think.  I don’t have the new interface yet.)

There are important other reasons why Facebook has gone this route besides wanting to become more like Twitter.   One reason is to do with app and content discovery.   Another is to do with a perfectly self-interested need to try to hang on to or even increase the interest value of the newsfeed – Facebook is hoping people will self-triage their posts and the fluffy ephemeral flow will go into ticker.

Facebook isn’t just Twitter.  But it increasingly has a Twitter-like side.   The encroachment on Twitter is real.   Put it this way.  With Facebook, you get the free knife set!  Why would you buy the knife set on its own?   There are reasons, of course, why you would.  I’m not sure they will be enough of them.

What I really want to know, in order to try to call this one,  is what the current overlap is between Facebook and Twitter user bases.   Anyone got some guesses they want to share?

But there’s a big slam dunk, too.  Facebook’s bid to become the arbiter of taste and faciliting platform for media consumption  is  likely to succeed massively – at least for music.   Which is emotional and social, or at least potentially so.   Huge possibilities here, particularly in co-presence and shared experience.

Mind you, those founder and CEO chaps they had up from NetFlix and Spotify didn’t look too happy about things.   Really they didn’t.  Not sure who they’ve invited to dinner,  and exactly what’s on the menu, I reckon.   Not to mention whose family silver the dinner is being eaten off of. But if they get it right the pie will grow, and everyone will have some.

I’m less sure about the NewsCorp comment that in future the way you will discover news is through your friends.   Before you can say “filter bubble”  five times, I’ll be outside getting a chill wind in my hair.

We’ve all got our histories. Yours. Mine. Facebook’s. I can just hear someone saying: ‘awesomely un-rockstar Facebook can’t colonise the past’. And then some other voice saying: ‘awesomely wrong, dude! we can!”.

The TimeLine is pretty.  The TimeLine is a huge improvement over what Facebook had before. Well done to the designers.

But it seems a strange reversion to an old-fashioned publishing model.  (Except, perhaps, for the economics.)

Why?  For me, personally, Facebook is about communication.  For me communication is fundamentally two way.   At least.    So this new emphasis on unilateral self-display is a turn towards the past in more ways than one.  But no one is guarding the door, and anyone may enter.

“The past is a foreign country….” has already been used as the opening of a novel (by L.P. Hartley).  The foreign country is one that is about to be occupied.

I’m not sure that I will invest my personal resources there.  But where gold is deposited miners will follow.

Ever-so-briefly:  funnels enable you to define event sequences that occur in the world you are trying to understand,  and measure what proportion of users make it through the successive stages of the sequence.       Funnels are often used to look at drop-out rates for important ‘straight line’ processes such as check-out, where a user’s failure to complete all the steps of the process is most probably  your fault (and is probably at least partially fixable). Nice.  But you can use funnels in other ways, too.

I’ve gotten into using funnels to compare paths in a user journey.

You take the high road and I’ll take the low, and I’ll be in Scotland afore ye.

Which is the most successful route to Scotland?  To assess this you set up more than one funnel, and you compare them.  I call this ACB testing.

What is ACB testing?

ACB testing uses funnel comparisons to look at how the  route a user takes through an app affects the likelihood of reaching a particular goal.

I’m sure you can think of lots of cases when this is an interesting question, but here are some examples to get you started:

• Does levelling up affect a player’s likelihood of returning to a game?
• Does interacting with other users via chat affect the likelihood of buying virtual goods?
• Does looking at a review for a product affect the likelihood of purchase?

ACB testing can help you get a grip on these questions.   What you do, is compare performance on a simple AB funnel with performance on a funnel that includes an intermediate step, let’s call it C.   This can give you insights into whether engaging in activity C promotes reaching B, or inhibits it.

But (watch closely now) there’s a trick to it.  You can’t directly compare results from the AB funnel to results from the ACB funnel, and sit back with a nice cup of coffee.   This is because, as shown in the illustration below,  the AB funnel includes all the user data from the ACB funnel, and if you did the obvious thing, you’d end up double counting people who go to step C en route to B.

Let’s say that C is engaging in online chat, during a visit to a web site.   You want to know if chatting influences the likelihood of buying, B.  What you want is to compare how likely chatters and non-chatters are to buy.  You want to compare ACB with A~CB (the ~ means ‘not’, so ~C means ‘not C’).

But you can’t directly measure something which doesn’t happen, using most funnel implementations.   They can’t tell you when the dog didn’t bark.  So how do you do it?  It’s easy.   You just need to compute the “~C” results by looking at the difference between the AB funnel, and the ACB funnel, as shown in the table below.   (Before you get excited and start implementing chat functions left right and centre, I do need to point out that this is not real data – I’m just using it to explain how ACB testing works.)

In our made-up example, when you compare the traffic on the path taken by people who chatted (which you get from the ACB funnel), with the traffic on the path taken by people who didn’t chat (which you get by looking at the difference between the AB funnel and the ACB funnel), you see that people who chat are way more likely to buy.

How exactly did I compute the A~CB path?

The different funnels, AB and ACB,  are just different ways of looking at the same information.   So you know that if 100 people start the process with Step A, and of these 100, 20 people chatted, you can confidently infer that 80 people didn‘t chat.   Similarly,  if you know that of the 50 people who performed Step A and completed Step B, 20 of them chatted en route, you can confidently infer that 30 of them didn‘t chat en route.

This enables to to assess whether the activity of chatting can predict buying.  In our made-up example, it can.  And you can use funnels to try to trap this effect so you can look at it.  Which is pretty cool.

How is this different from AB testing?

With AB testing, you put people into two different treatment groups and see what happens. You observe the consequence of these different experiences for your downstream measure of interest. If you see a significant difference between groups, and you have assigned people to groups randomly, you can be pretty sure that it’s because of what that group’s experience was.

ACB testing is different.  With ACB testing, people put themselves into groups by their behaviour. If you see a difference it outcome, it could be due to some other factor which is responsible for people’s self-segregating into behavioural groups.   It’s different – but it’s still interesting.

Are funnels the best way to do ACB testing?

Funnels are not actually the most direct way of assessing the differences in paths taken.   The best way to do that is to parse the event stream yourself, and tally it up.  However, since funnels are commonly used in Web Analytics, and are easily available on a lot of fab (and free) analytics platforms, I thought that others might be interested in this way of using them that’s a bit different from their original purpose. It’s a new way of using them to funnel off value, if you like.   Try it in your sandbox and see what you find out.