March, 2018

Rail v Road During Holidays: Compare and Contrast

The above press release dropped into my inbox the other day.* It’s from Highways England, and is of the sort that is regularly published at holiday times of the year. So as we enter the Easter weekend, roadworks all over the country have been temporarily lifted to enable the flow of traffic.

They even have a quote from the MD of National Express: “It’s great that Highways England have lifted roadworks on key routes, including those serving airports, helping us make sure we can get passengers where they need to be for their Easter plans.”

Compare and contrast with this message from the National Rail website.

While the roads are cleared to ensure easy travel, lots of rail works are scheduled for precisely this period, notably including Bristol, the West Coast main line, Manchester Victoria, London Euston and more.

The reason given is that the period is, “A time when less people travel on the railway and when traditionally a considerable amount of improvement and engineering work needs to be undertaken on Britain’s rail network.”

Let’s parse that a little shall we? The first part of that sentence, grammatically should probably say, “fewer people.” But we’ll come back to that first part of the reasoning in a while.

The second part of the sentence is basically saying, “We’re doing works now because we always do works now.”

I’m not really sure that’s an excuse.  I completely understand the need for maintenance and improvements – these are essential. What I’m not clear about is why these have to be scheduled at a time when large numbers of people are travelling often long distances to be with family and friends.

Network Rail actually has a page on their website explaining, “Why we carry out work at weekends and bank holidays”

Here’s the key text from that page:

We plan works for certain times so they cause the least disruption to passengers, such as on bank holidays, Sundays and overnight, when the network is less busy.

An independent review in 2016 looking at how the rail industry plans and schedules major improvement work concluded that Christmas, Easter and bank holidays are the best times for upgrades that need major lines to be closed. Even though it might seem strange to carry out work at Christmas – when people are travelling to see friends and family – on average, around half the usual five million people travel by train each day during the Christmas period.

I’m not going to dispute the claim that fewer people travel during the Christmas period (although that doesn’t mention Easter), but there a couple of things that I would bring to bear on this.

First, overall rail travel is vastly driven by travel in London and south east, and in particular commuter traffic. That largely stops over the holiday period, and might easily account for most of the overall reduction. According to the Department of Transport’s most recent Rail Factsheet, 69% of all passenger rail journeys are accounted for by London and the south east alone. Much of that is commuter travel.

What would be much more useful would be to understand how much the traffic flow changes for different types of journeys. For example, does long distance or inter-city traffic decrease, stay the same, or even increase?

The second thought I have is that because rail travel at holiday times is so unpredictable, more people take to cars. But this disadvantages those who don’t drive or don’t own a car – notably many of those in inner-London boroughs, or those who are poorer.

The statement above talks about a 2016 independent review, and I confess that I had trouble tracking that down. I did find a 2015 report commissioned by the Rail Delivery Group: Planning and Timing of Engineering Works on the GB Rail Network. This followed the failure to complete works on time over Christmas 2014 when there were overruns and serious problems with people travelling around the country.

Interestingly, it seems that getting accurate and full data for the report was something of a problem:


“Whilst rail travel is popular around Christmas passenger volumes are lower than the rest of the year. We looked at passenger numbers and type of passenger (leisure or commuter) during the year, which were difficult to obtain in any detail. Although we expected the passenger mix to vary with the time of year we did not find significantly lower passenger flows during the summer holiday periods or around the bank holidays on the major London routes. Obtaining more detailed insights into passenger flows during the year as a base for planning is essential and is one of our recommendations.”

Another point of note was this:

The passenger mix at Christmas is different than at other times of the year with a higher proportion of leisure passengers who are unfamiliar with the railway and less capable at coping with modal transfer during disruption.

But they also noted that fewer elderly travel at Christmas – perhaps because people travel to the elderly rather than expecting them to do the travelling.

I found this to be an interesting paragraph:

At present the major blocks at Christmas and Easter contain a range of work. Some of this can be done only at these times: other work can be undertaken at weekends but often is not done because the amount of weekend access is limited and there is pressure to add work to major possessions to improve the overall productivity of work. If the industry were able to make greater use of extended midweek night access (having full due regard for revenues generated by traffic that operates at night, especially freight) it would be possible to move some work undertaken at weekends into midweek nights. This would, in turn, free up weekends to do work that is currently being squeezed into the margins of long blockades. However, this will need to be balanced with the potential revenue benefit from reducing weekend access, which has been a focus of APSCM work.

In other words, if some freight traffic were disadvantaged, then weekend work could move to mid-week overnights, and holiday work move to weekends. That at least would leave the big holiday periods more free of disruption.

There are lots of other issues including adjacent line working (work being carried out alongside a working line), bi-directional signals (lines being capable of running trains in both directions – largely not the case in the UK), and other factors. Not least is the various recharging and pricing elements in terms of the timing of works. It seems Christmas overtime costs might be negated by other pricing determined by government.

It’s interesting to note that in other European countries they do things differently.

  • In the Netherlands: “There is no project work undertaken during the Christmas holidays and the burden of engineering projects is better divided over the year including long blockades during summer holidays.”
  • In France: “While passenger flows into Paris are similar or higher then these into London, enhancements and renewals are being done throughout the year, but not at Christmas. Long possessions are taken during the August summer holiday period, even at the RER for which busses and alternative routes are being offered as alternative; It should be noted however that there is hardly commuter traffic during that month.”

What I didn’t see in the report is any comparison of rail travel over different route types. In other words – shorter distance largely commuter travel v longer distance inter-city travel.

It’s evident that this work needs to be done, and I’d never want to underplay the complexity of track access, and the various calls there are on our rail network. But as I’ve argued before, it feels as though those who need to travel during holiday periods are actively disadvantaged. Furthermore, carrying out works in short bursts is less efficient than closing a line down for a longer period – a blockade in the industry parlance – and getting more work done. It’s notable that in cases where some major work is required, shutting down a station for a longer period, despite inconveniencing commuters, leads to fewer overall disruptions.

I would just like to see more innovative thinking from the Department of Transport and Network Rail.

 

* No. I’m not entirely clear why I was sent it either. But I was.

Facebook Pixel Tracking

This morning Nieman Lab had a really good piece asking whether if there was a certain amount of hypocrisy coming from certain news organisations castigating Facebook for leaking data, when at the same time they’re helping Facebook collect more data on you.

Recall yesterday, when I said that some of Facebook’s data was missing from the download, and I highlighted Facebook Pixel? Quora is a good place to go and have a look (Note: Quora itself has a Facebook cookie):

The pixel is a transparent, 1×1, unique image file that can be embedded on pages outside of Facebook (unique = 1 per advertiser account). That image file, however, sits on Facebook servers. So, each time it is loaded, it increments counters on Facebook’s side.

And each time the pixel file is being seen by a user… Facebook servers can see which browser is used, which machine and which IP address. In other words, they are able to reconstruct that signature – they know which Facebook user has seen the pixel.

In essence, the Facebook Pixel lets you then target people who visit your site when they’re back on Facebook. And of course, Facebook now knows that you’ve visited a particular site, deepening the picture they hold on you.

And Facebook also has the Facebook Audience Network, which basically extends Facebook’s advertising business beyond the bounds of the Facebook website. In particular, they’re targeting mobile sites and apps.

Using the EFF’s excellent Privacy Badger browser plugin, I looked at some of the UK and US’s biggest news websites to see which ones allow Facebook to track you. This obviously isn’t a comprehensive list, but it gives you an idea.

Sites with Facebook Cookies

  • The New York Times
  • The Washington Post
  • Forbes
  • The Daily Mail
  • The Sun
  • Metro
  • The Times
  • The Economist

NB. These are at time of writing.

It must be said that I’ve not really gone into detail about Facebook’s business model here, but it gives you an idea.

And there’s a wealth of data being collected by many companies beyond Facebook – and a multiplicity of ad tracking cookies going around. Upwards of 20 cookies on a website is not unusual. Sometimes they’re just there for analytics purposes. All the advertising networks use them, with Google and Facebook being by far the biggest networks globally.

And there can be good reasons to use tracking cookies. This very site uses Google Analytics to count the number of people who visit, for example. I’ve embedded Vimeo videos and Flickr images on this site, and they have tracking codes built into the code I copy to this site. If you comment, there are various ways you can log in, and they have tracking codes too. I’d prefer there not to be, but if I want to properly use those sites’ services then I have to play ball with them.

While everyone kind of knows that the pair of shoes they looked at over lunch, but didn’t buy, is now following them around the internet, and that must be using some kind of tracking information, I’m not sure that many of us really understand how widespread this is, and how much data is being captured about us.

[Update: In related news, Mozilla has announced a Firefox plugin that stops Facebook tracking you around the web. Useful if you’re not already using something like Privacy Badger or Ghostery.]

Is Netflix Quite As Smart As Everyone Says It Is?

That’s possibly a provocative title, but I’ve come to the conclusions that while Netflix is very good at some things, I’m not certain that its recommendation engine is entirely as linked up as you’d think it’d be.

A couple of recent cases in point.

I was really looking forward to the new Alex Garland film, Annihilation. While I was slightly disappointed it wasn’t getting a cinema release, I was very pleased that Netflix was investing in it (well, buying the rights), and making it available to its subscribers. I dutifully searched for it ahead of its 12 March release, and added it to “My List,” Netflix’s somewhat clunky system for saving things you want to watch.*

Although I believe the film was made available at midnight UK time, but I waited until Monday evening to open the Netflix app on my Nvidia Shield and settle back to watch. I thought that they’d probably have the film front and centre when I opened the app. After all, it was a big coup them getting it. Plus I’d explicitly added it to my list.

There was no sign of it. It wasn’t in trending (too early I guess), or in any of the top lists of things I might want to watch. I ended up using Search to find it. It was – but hidden.

Then over this past weekend, while I was out and about, I got some Instagram advertising for a film called Paradox with Darryl Hannah and Willie Nelson. I’d not heard of it, but clicked through and saw video for some kind of western themed film. “I might watch that,” I thought – vaguely intrigued. Netflix are obviously promoting it, I’d catch up with it at some point.

Later, with that thought having drifted out of my head, I did open Netflix again in search of something to watch. Had I spotted Paradox, I’d have at least given it a second look.

But it wasn’t there. Or more to the point, it wasn’t obviously visible. In any case, because a film I’d seen promoted precisely once, was no longer in my view, I didn’t search for it. I only remembered this at all because I saw a second Instagram ad for it earlier today.

But again, it feels like Netflix is being a bit slow and doesn’t have all its ducks lined up. It’s not that I don’t think they can do some clever stuff, but they’re not as good as they make out.

Have you heard of a Danish comedy drama called Rita? Maybe if you’re Danish, but otherwise, you might not have. Netflix never recommended it to me. It was someone on Twitter who noticed it. It’s very amusing.

I started watching a Spanish series called La Casa de Papel. It’s a series about a gang of thieves who try to rob the Spanish Mint. It starts well, but like another Spanish series I saw on BBC Four last year, the strong hook doesn’t last the course, and we end up with an interminable number of episodes where not a lot happens, and the villain is really villainous. More plot and fewer episodes please Spain. I mention this because after I’d watched a few episodes on Netflix, the series promptly changed its name. It’s now called “Money Heist,” although it wouldn’t be obvious to those like me who’d started watching it under another name entirely. I had no idea what Netflix had done!

I’m always suspicious of over-claims about how briliant someone’s algorithms for discovery are – mainly because I’ve yet to experience anything that’s really that good. Amazon is pretty bad at recommending me books I didn’t tell it about, and music recommendation engines are pretty poor in my experience – especially if you move beyond the obvious.

Maybe they work for some, but I’m underwhelmed.

* I say it’s clunky, because it’s incredibly binary, and doesn’t allow you to make lists for different things. Furthermore, when you watch something that was on the list, it doesn’t then remove that item from your list. I’m also not aware that Netflix alerts you when something that’s on your list is shortly to be removed. Another useful feature.

Examining My Facebook Downloads

One very good consequence of the Facebook/Cambridge Analytica story is that a lot of people are discovering the surprisingly large amount of data that Facebook holds on them. The BBC’s Rory Cellan-Jones was “somewhat shocked” to see what it had on him. And The Verge has a good piece on the subject with particular reference to Android phones.

In essence, Facebook always asks for quite a lot of data when you install its apps, and people seem to be too quick to offer that data when it comes to installing those apps. Only now are they discovering what they’re sharing.

“Yes, yes. Just install and let me get onto Facebook,” seems to be the default thought process.

Now I’m not going to pretend that I’ve always slavishly careful about those permissions myself, but I certainly wanted to see what Facebook holds on me. So I went to the Facebook Settings page and clicked on the Download A Copy link at the bottom.

Facebook first has to prepare the data, crunching it into a Zip file for you. You need to re-enter your password to begin the process, and Facebook promises to email you when the link is ready.

Based on others, I thought it may take a while to compile, but in face it took just 16 minutes. Fast considering the volume of data and the number of users who are perhaps also doing this right now. You have to re-enter your password a second time, and then the file downloads.

I’ve been on Facebook since 2007, and I thought that this could be a big file. In the end it was just over 1.1GB. I’ve uploaded a lot of photos to the service in the past, but particularly in the early years of Facebook, they heavily down-sampled those pictures. (Another reminder that you shouldn’t use Facebook as your only photo backup.)

Anyway, the file extracts easily enough and Facebook has built a fairly intuitive html interface for you to examine your data offline.

My profile data is an interesting place to start. Facebook seems to have detected a single family relationship. While relatively few of my family are on Facebook, some of those who are, were not picked up here as family members. If they don’t have the same surname it might not be obvious to an algorithm.

The interests section is very odd, and not very accurate. When Facebook first started, you just had empty text boxes to fill out. I wrote a general stream of consciousness about music, TV, movies and so on. At various points Facebook has tried to clean that up a little, isolating artists and titles, and linking them to official accounts or lists that it has.

But despite prompts to help them (and help me!), I never really played ball. So there is one novel listed in books, which I think I was probably reading at the time. There is one TV series – one that I absolutely do not recommend. Movies are a little more populated, but with films I may have referenced directly on the service rather than anything else. And music is very limited. Facebook really doesn’t know much about my media consumption.

In general, Facebook would learn a lot more about my media choices if they scanned through this blog!

Otherwise, most of the rest is either groups or people I’ve taken an interest in. I would say that they’ve used Instagram heavily for the latter.

Probably the most contentious area is the list of contacts. And for me, that’s a moment in time, when I did at one point let Facebook into my phone or Gmail account. The list of contacts is old, and while many of those email addresses and phone numbers still work, they’re cast in aspic. Over the years I’ve had any number of phones, and if and when I install a Facebook app, I never give permission for it to see my contacts.

My Timeline is as you would expect – everything I’ve written on Facebook. I link my Twitter account to Facebook, because I’m far more active there. All those Tweets are also captured here. But nothing I wouldn’t expect Facebook to have.

As I mentioned above, I’ve uploaded a number of photos to Facebook over the years. They tend to be more social photos than anything, and Facebook was an easy way to share with friends and work colleagues. Latterly, anything that I’ve cross-posted from Instagram shows up. [Update: A friend – on Facebook – noted that captions for photos are not included]

There are only a limited number of videos, again social, and no surprises.

Messages lists all my Facebook message and Messenger interactions. I loathe Messenger and don’t ever have it permanently installed (On occasion I’ve installed it for a short, but necessary period of time. I uninstall it immediately thereafter). Nontheless, again there were no surprises.

The data supplied by Facebook on “Pokes” (Remember them?) was incomplete. I only had one poke listed!

Security lists a variety of things including devices, and even IP addresses from which I’ve accessed Facebook.

The final two key pieces of note were Applications and Ads. I recently cleared out the list of applications that I allow Facebook links to. It’s always worth doing this on a regular basis. I know precisely which apps are currently linked, and there is a good reason for each of them. There are only five.

Ads are broken into three parts. There’s the list of topics that Facebook thinks you’re interested in. This is a curious mix of very broad things (“Music”) and very narrow things (“Dan Martin (cyclist)”). It’s reasonably fair, although I don’t really have a particular interest in Citroen, nor Motor Sports or Auto racing. And I’ve no idea why “BBC Radio Solent” is one of a handful of radio stations listed as being of interest to me [Update: I worked out that a former work colleague of mine works there now, and I’ve liked some of their activities]. They do at least list my current employer! My previous employer is not listed. It’s possibly that this list is dynamically updated and pruned accordingly.

Ads History claims to list all the ads I’ve clicked on. They only have two listed – both this year – and one without a named advertiser. This is clearly missing data. While I do recall clicking the one named advertiser, and although I rarely click advertisements, I have clicked others in the past. Incredibly, I once actually bought something on the basis of a Facebook ad! Extraordinary, I know.

Finally, perhaps most worrying for me, is a list of “Advertisers with your contact info.” Most of the list is made up of KLM subsidiaries. I once entered a KLM competition on Facebook, and must have agreed they could use my data. I rarely participate in competitions that require much data access for this very reason. Uber, Airbnb, Deliveroo and eBay Canada seem to have my details. But there are a hole bunch of seemingly related “Crowdfunding” companies who have my data. I’ve no idea how they got it, and more importantly, I’ve no idea how to remove it from them. In general it’s quite a contained list.

Notably, Facebook does not have a list of my outgoing or incoming calls, and it’s not had access to any SMS messages I’ve sent. I’ve never given permission, and never wanted to use one of its products as my default SMS app.

The most sensitive data is my list of contacts. But that data is old and is not being updated since the Facebook app on my current phone does not have permission.

As I’ve said repeatedly on this blog, I’ve never found Facebook the most trustworthy company. But on the other hand, there aren’t any surprises to me from what Facebook has in my data.

I think that there are some incomplete aspects of it. I’ve clearly clicked on more ads that Facebook is admitting – but perhaps they delete that data after a period? Less importantly, the list of Pokes was incomplete. I mention that only as it suggests that this might not be a truly complete picture of my Facebook activity.

But I also know that if I carried out the same process for Google, it would be a lot larger. Google has all my email. It has all my contacts. It stores documents, photos and videos for me. I use its browsers multiple times per day. It knows what YouTube videos I watch. It knows what music I listen to. I’ve had phones running its software for years. They know where I go.

In all of that respect, it’s potentially a much scarier proposition.

And yet, I do have more trust in Google than I do in Facebook. Perhaps that’s misplaced? Perhaps not. But in general terms, I think people are clearer in their knowledge of how their Google data is used.

Auditing who knows what about you is important, and we should all be doing this on a regular basis. It’ll be a much bigger job, but it looking at my Google Data might be worth doing too…

[UPDATE]

It’s probably worth highlighting a few things that you don’t get from this data.

  • Likes – Given that a key part of the Cambridge Analytica story is about trying to determine OCEAN psychographic measures from Facebook likes, a record of comments and pages I’ve “liked” is data that’s relevant but not here.
  • Facebook Pixel dataFacebook Pixel is the technology that Facebook uses to determine where users also go. While that could be websites that simply allow you to comment via your Facebook login, it might as well be websites that you never realised had installed the pixel. In effect, when you visit such a site, Facebook knows about it. It gives them some of the data that Google collates about you via its ad networks.
  • Geographic data – Facebook loves to know where you are. I mostly have this turned off, but couldn’t definitively say that this has always been the case. While Google has its Timeline History that tells you where you’ve been, there doesn’t seem to be an equivalent for Facebook’s location data. Incidentally, if you’ve never explored that Google data, I’d urge you to. You’ll be delighted, scared and possibly both. (Note to crime drama and fiction writers: Nobody ever uses this, although I understand it potentially increases the difficulty in plotting your story as mobile phones in general have.)
  • Whatsapp or Instagram data – I’ve noted that some of my Instagram information does seem to have fed through to parent company Facebook’s data. But that doesn’t seem to be the case for WhatsApp. Within the EU, Facebook has been limited quite significantly about how much data it shares. The UK’s Information Commissioner made that very point again recently. But it’s worth noting nonetheless.

Facebook and Cambridge Analytica

I’ve been following the stories surrounding Cambridge Analytica and Facebook for some months now, and in recent days, following stories from The Observer, The New York Times and Channel 4, the story has really blown up.

However, I do think that the story, while completely valid, and asking some really critical questions, perhaps over emphasises some aspects at the expense of others. It’s true – I tend to see cock-up rather than conspiracy in most things. But here are some of my thoughts on the two key players in this instance.

On Cambridge Analytica

My underlying belief is that Cambridge Analytica is a company that claims it’s more powerful than it actually is, and uses the “digital voodoo” to win business in the murky world of political consulting.

If nothing else, tonight’s Channel 4 News exclusive shows that aside from stories about how powerfully the company is able to target voters using sophisticated social media targeting, more than anything they’re an ethically dubious political consulting group who will do just about anything to see that their paymasters win.

On Newsnight, their CEO Alexander Nix somewhat disingenuously claimed that he was being targeted by British media. That’s certainly not what came across in the Channel 4 video.

As they themselves will admit, it’s clear that Cambridge Analytica did indeed use Facebook targeting to try to micro-target individual voters during the last US presidential election, with advertising that pushed the right buttons with those voters. So if they thought that you might be swayed with promises about jobs, then that’s the hook that they would use to get you to vote Trump. But I also think it’s true that previous campaigns have done the same – notably Obama’s.

Of course we’ve not really had many elections in the time of social media to truly measure the impact, but while the data does allow a political party or its supporters to make potentially hundreds or even thousands of different pieces of copy that might tick the right boxes amongst voters, I think that it’s incredibly naive to believe that so many are willing to change their voting intentions on the basis of a few Facebook ads. It takes more than that.

That being said, this is clearly the narrative that Cambridge Analytica have painted for themselves. They worked for a campaign that wasn’t seen as likely to win the US election and yet it did. That means that they can go out and win lots more business all over the world with the suggestion that it was them “wot won it,” – as the Sun once claimed after the Conservatives won the 1992 UK General Election.

It’s a great gimmick. Nix has been able to go out and proclaim that his company has discovered the secret to winning elections. And campaigns of every hue have been queuing up at his doorstep.*

Beyond that, it does sound as though the company has been, if not bending the rules further than it should, taking and using data it really isn’t allowed to. After months of questionning, as the big stories came out over the weekend, Facebook finally pulled Cambridge Analytica’s access to use the platform and says that it will be conducting a forensic audit on how the data was being used. Meanwhile the UK’s Information Commissioner is seeking a warrant to look at the company’s servers and databases.

I wouldn’t in any way excuse the company. If the claims made by both whistleblowers and others are proved to be true, then the book needs throwing at the company.

But I still need a lot of convincing that the company is as powerful as it would like to portray itself. Instead, I see a company who has decided to lead on it being an expert at data science in the world of political consulting to give it an edge over its rivals. And given the scale of social media networks and the way that users and the networks themselves use this data, we are more willing to believe that they are more powerful than they truly are.

On Facebook

I think Facebook has some real problems that are largely of their own making. This boils down to a few key areas, all of which they really only have themselves to blame for.

Privacy Settings

Facebook’s privacy settings have always been a nightmare for users. With a network as complicated as the one it has built, relatively few really understand what they’re sharing and who it is with. The settings change relatively frequently meaning it’s hard even for the most assiduous privacy minded user to keep up with who has access to what. What stories you like; who your photos are shared with; who your friends can further share things with and so on. It’s truly complicated.

Some people manage to lock their accounts down quite a lot, but they’re the exception rather than the rule. Who can friend you? Who can share things with you? Who sees things that others have shared with you? Can you be found with your email or phone number and who can find you that way? Are you searchable via Google? When you like a company’s page, how much data is that company able to capture on you? What about when you enter a competition?

Lack of Control

Such is the breadth of the data that Facebook is able to capture, that one of the most troubling aspects of this story is the general lack of control the company seems to have. One part of this story concerns the use of a personality prediction “app” on Facebook called thisisyourdigitallife. The questionnaire was completed by 270,000 users but the data captured wasn’t just for those users, but for many of their friends as well. In total that gave data for over 50 million users. Facebook refers to these as “friends who had their privacy settings set to allow it.” But just how many of those people who didn’t complete the survey remember, truly realised that this kind of data could be captured? While it may only be how that data was subsequently used that broke Facebook’s terms and conditions, it’s worrying that someone is able to capture that level of data regardless of whether they’re using in accordance with those rules.

And I suspect that a very small minority of users realised just what kind of catnip they’re giving to companies like Cambridge Analytica when they complete “fun” surveys such as this. For most users they’re just communicating in a fun environment with their friends.

Advertising

Facebook has built one of the most powerful digital platforms in the world; a platform that reaches billions and one which generates nearly all its revenue from advertising. That advertising is targeted using a massive database that it has generated based on user data that the company would say users have volunteered themselves. That in itself might be troubling considering that users largely don’t really consider how that data about themselves is used. But even worse is the idea that third parties can come in and use that data to target individuals without it even being revealed who is paying for those ads. We’ve seen that with ads paid for by Russian sources during the Clinton/Trump election campaign, and we’ve seen it in other campaigns. Channel 4 News highlighted last year’s Kenyan election for example.

It’s surely critical for any platform that runs advertising to be up front and honest about who is actually paying for that advertising.

Data

Facebook captures enormous amounts of data and it’s not simply when you use the Facebook website. You might be sharing your location with Facebook via its mobile app. Third party websites incorporate what is known as the Facebook Pixel. As with other advertising networks including Google, it means that Facebook knows where else you are outside of its own network. All those helpful sites that let you log in with your Facebook account? That’s more data you’re sharing with Facebook on top of the thousands of fields of information it is already keeping about you.

This can all make it seem to users that Facebook is actually doing nefarious things like using the microphone on your phone to listen in to your conversations. It’s almost certainly not, but as I said in a piece last year, if the company is so uncannily accurate, then the perception trumps the reality.

Summary

This is serious story – scandal even. And as much as anything, I’d like to think that this is a wake up call that gives people a greater understanding about how their data is being captured used, and potentially misused.

But I suspect that Cambridge Analytica is really just another political consulting company who’s USP is that they target voters with social media. They hit the big time by working for the Trump campaign, and being allied with Steve Bannon. Yet as a result, their marketing claims have become so hyperbolic that it has led to a widespread disdain for what they do, and since this story has begun to unravel, they’ve been rowing back how impactful this aspect of what they do actually is.

Instead, based on evidence from Channel 4 News, the company is perhaps more about the grubby world of sending prostitutes to the homes of political opponents and capturing it on video, or giving the appearance of having developers pay backhanders for property deals. This is all as low rent as you like.

Facebook’s problem is that it has too much data to the point that nobody seems to be able to keep on top of things. They have so much that some users suspect them of actually listening in to them via their phones. But they truly do follow you around the internet. The danger for Facebook, beyond what are likely to be short term falls in their stock value, is that users do start to rebel and close down their accounts.

I’ve always had a problem trusting Facebook. I don’t think they’re evil as such. But they have played too fast and too loose, and have ended up in a powerful position. As a by-product – alongside Google – they have just about completely cornered the digital advertising market which brings with it its own problems for society in general.

I’m not going to underestimate the problems with “fake news” and the ability of propaganda to spread like wildfire on social media platforms. There are some serious questions to be asked about how these platforms can and should be regulated, particularly in regard to elections where we have seen continuing problems.

On the other hand, just because I can be micro-targeted using all this data, it doesn’t necessarily win elections.

People who work in advertising always love to tell you about how their methods work – explaining that they built the biggest brand in the sector using their methods. That’s the same in politics too. Saatchi & Saatchi were widely credited with helping Margaret Thatcher win the 1979 General Election with their famous “Labour Isn’t Working” outdoor poster. The poster did indeed sum up very nicely the prevailing political landscape. But did it actually win the election, or was the country moving away from Labour anyway? It was a sizeable win for Tories. I’m pretty sure that it wasn’t a poster “wot won it” that time, and I’m pretty sure that a company led by an old Etonian with some Facebook data wasn’t actually responsible for Trump (or Brexit) either.

Further Reading:

A really decent piece from the New Statesman that seems to accurately summarise the full Facebook data part of the story.

It’s been said in some more breathless quarters of the internet that this is the “data breach” that could have “caused Brexit”. Given it was a US-focused bit of harvesting, that would be the most astonishing piece of political advertising success in history – especially as among the big players in the political and broader online advertising world, Cambridge Analytica are not well regarded: some of the people who are best at this regard them as little more than “snake oil salesmen”.

A Verge piece that really gets into psychographics and microtargetting and what it can and cannot do.

Taken altogether, it seems like Facebook was taken in by a shady firm that misused data and lied about it. When Facebook found out, it did nothing. And making matters worse, we can’t even point at Cambridge Analytica’s deception as the reason Trump was elected: a closer look at its methods suggests they might not even work.

Cory Doctorow at Boing Boing says that he doesn’t think Cambridge Analytica is actually able to do what they say they can.

So, as I’ve written before, we should take Cambridge Analytica’s claims to Svengali-like mind-control with a boulder of salt, because until Sunday, they made these claims to drum up business (now they’re busily declaring that they are no more persuasive than any other ad agency, of course, because they’ve gotten in trouble for it).

Antonio García Martínez at Wired magazine on the noisy fallacies of psychographic targeting.

For the impatient, my fundamental thesis is this: Cambridge Analytica’s data theft and targeting efforts probably didn’t even work, but Facebook should be embarrassed anyhow.

* In a strange radio related twist, Cambridge Analytica’s offices at 55 New Oxford Street are in the same building as many of the UK radio industry’s offices including Radiocentre, RAJAR and Digital Radio UK! But it’s a largish office, and other companies share the building too.

The Business Models of MoviePass and cPass

Over the weekend, a new company raised its heads above the parapets. cPass is a new subscription cinema going scheme that allows members to see one film a day at the cinema for a monthly fee of £9.95.

Cinema membership schemes aren’t unheard of, but they tend to be more expensive. Cineworld has its Unlimited scheme that costs £17.90 a month, rising to £20.40 if you want to include central London cinemas. My local suburban Cineworld charges £12.10 for a regular ticket in peak times. So I need to see 2 or more films a month to make a Cineworld membership worthwhile for me. But under the cPass scheme I’d only need to see one film to start making savings.

Odeon has it’s Limitless scheme and it’s very similar to the Cineworld offering. It costs £17.99 a month, or £19.99 if you want central London cinemas included. Note that the flagship Odeon Leicester Square is currently closed for renovations, although a single visit to that cinema can easily exceed £19.99 alone. However my nearest suburban Odeon charges £12.50 for a regular ticket in peak time. So as with Cineworld, I’d need to see two films a month for it to be worthwhile.

So how is cPass offering a seemingly better deal than either of two of the UK’s largest chains offer on their own, with the advantage that I can see films in any cinema (subject to terms)?

Well, they’re copying nearly precisely the US MoviePass model. That is, they’re not working directly with the cinema chains at all. What they do is send their members a debit card – I assume either Visa or Mastercard – and when someone books a ticket on their app, it puts some money in the debit card’s account for a limited time and lets you buy a ticket as you would with any regular debit card. The cinema gets paid, and you get a ticket.

But if I’m paying £12.50 for a ticket, yet have only paid £9.95 for an entire month, how does that make any business sense?

The Recode Media podcast recently interviewed the CEO of MoviePass to try to understand the model, and a few things emerge, whilst others remain unsaid.

The Gym Membership

Part of the model is the assumption that we won’t all be trying to see 31 films in a calendar month. The average person probably sees 2-4 films a year. i.e. not that many. Everyone knows that gyms are packed in January and then settle back down to a more manageable level shortly thereafter. Gyms have more memberships than members they can cater for. They hope that while some are going 5 days a week, more are going much less frequently and are too lazy to cancel their memberships. We’ve all heard the tales of onerous cancellation procedures – there was a whole Friends episode about this.

MoviePass are quite honest about the fact that when members first get their cards, they rush to see lots. But fairly quickly they drop back to a more modest level of film going.

We’ve all got subscriptions to things we don’t use as much as we should. Part of MoviePass’s model relies on that.

Different Areas – Different Prices

I live in Greater London and the price of a visit to my local cinemas is just more than £12. Other parts of the country may charge closer to £6. So the maths can be different, and even if they make a loss in London, it could be offset elsewhere.

That all said, cPass says it’s launching in the capital.

Growing the Market

On this week’s Kermode and Mayo podcast, the thorny subject of film piracy was raised again. It’s clear that lots of people are using dodgy streams to serve pirated films – much more so that downloading torrents of a few years ago. Kodi boxes with the right (i.e. “wrong”) plugins have made it simple to watch recent releases on your TV at home, perhaps having to suffer some dodgy pre-roll adverts.

Mark Kermode’s solution to this is the day and date multi-format release. That way, you could choose to buy the DVD or Blu Ray, stream for a fee on your preferred service, or go to the cinema.

I suspect that wouldn’t work very well and would swiftly see the end of cinemas altogether. How many families would honestly spend £25-30 or more for a trip to the cinema with all the add-ons when they could get the DVD, Blu Ray or legal stream for closer to £10? Visits to the cinema would drop away massively, and they would start closing. I don’t deny that the right film seen with a large audience is great fun, but I’m uncertain that this is enough to prevent a serious dent in cinema attendance.

While I’m not certain that the music model is quite right at the moment, it is true that the likes of Spotify have removed the reasons for downloading music illegally. You get high quality music either for free with ads, or for a relatively modest monthly sum, ad free.

Part of the cPass/MoviePass model is that more people will go and see more films. They grow the overall market and encourage those who see relatively few films to see more. In turn that generates more tertiary revenues for cinemas – i.e. popcorn sales.

The downside is that the likeliest people to take up something like cPass are those who already go to the cinema a lot. Indeed, subscribers to current unlimited schemes would surely swiftly cancel their current memberships and move to the cheaper model.

Scale and the New Normal

Scale is what it’s really all about. These companies want to become dominant in the market place and have their members become a significant part of the overall audience. That gives them an awful lot of power with the chains (see next section).

You need deep pockets to play this game, and the backers of these services are clearly spending to get to a certain level whereby they can start to use this scale to their own advantage. This is the familiar Silicon Valley model of spending heavily to get an audience or user base, and then turn it around to monetise it.

At the same time $9.95/£9.95 becomes the new normal for pricing schemes. As alluded to at the top, this is close to half the price of existing schemes that are generally limited to a single chain. It instantly becomes harder for Cineworld to market Unlimited when there’s a cheaper, better option out there.

As it stands the chains know who you are, how often you’re going and what you’re watching. That’s valuable data. They instantly lose that as patrons move to the cheaper non-affiliated deal.

Deals with Chains

This is the big unspoken bit. MoviePass in the US is already negotiating with smaller chains and indies to get both discounted tickets and even kickbacks from sales of food and drink.

If a large proportion of the population is using something like cPass or MoviePass, then a certain amount of power is wielded by those companies. They might try to “incentivise” members to use one chain over another by temporarily or permanently removing certain cinemas.

If this result in sizeable declines in box office takings at the removed cinemas, these companies could throw their weight around and “force” the chains to provide them with deep discounts.

This has happened in the US with MoviePass, who have excluded some AMC locations, seemingly to pressure AMC into giving it discounts.

It’s worth noting that deep discounts do already exist in the marketplace. Many corporate “perks” websites and other third-party membership deals offer significant discounts to cinema tickets. 50% discounts aren’t unheard of. You can safely expect that these pass companies will be pushing to get discounts at least as deep as these, because they can redirect audiences away from anyone who doesn’t play ball with them.

It’s a fine line of course, since if none of the chains play the game, then the pass companies could be left out in the cold, haemorrhaging money. But it’s going to be interesting to watch.

Would the chains consider themselves as being “held to ransom,” and being forced to co-operate when they don’t want to?

Marketing Opportunities

These pass companies will also be chasing revenues from film distribution companies as well. They can do deals to heavily promote certain titles and push their audiences towards them.

The data they collect on their subscribers viewing habits could potentially be used to point consumers to relevant films – or at least do as good a job as people like Netflix can.

Summary

This isn’t by any means a proven business model, and if the big chains hold firm on their pricing, then it’s unclear how it can ever be. If lots of people don’t go and see many films yet continue to pay $9.95/£9.95 a month, then the sums work anyway, but I suspect it’ll take more than that. Particularly if these cards become popular among the younger demographics for whom money is tight, but demand is high.

The strange thing is that cinema owning isn’t as profitable as it might be. Deals with distributors mean that for big releases a large proportion of the ticket revenue, particularly in early weeks of a release. It was reported that in North America Disney wanted 65% of box office revenues for The Last Jedi, and also required that the film played for at least four weeks in the largest auditorium. Failure to do this meant that the share would rise to 70%!

Lots of cinema chains are desperately seeking reasons to get people to upgrade their seats and spend more. We have premium seats, 3D films, IMAX and more recently 4DX – all trying to get you to spend more. These pass schemes mostly push for the cheapest tickets. Does this work against the needs of the chains?

And what does this mean for overall box offices? If tickets are being sold at cheaper price points, then even if there are more people on seats, the overall value might fall if a high proportion of movie goers are paying what are effectively discounted prices.

The pass companies would counter that they’re about growing the market. And that might be the case. But as mentioned above,the first group of people who are likely to jump on these schemes are regular cinema goers who can quickly save lots of money. Just as the early adopters of Spotify were those who in the past had spent a lot of money each week on music, now only had to spend a tenner a month for the same quantity!

A really interesting scenario would be if some of the UK chains formed their own cross-chain pass. In other words, merge Unlimited and Limitless with any other schemes out there. It’d potentially be cheaper to operate since you wouldn’t need to get any banks involved and all the costs of processing debit cards.

Making those unlimited/limitless cards cheaper would certainly grow the market, but would distributors be happy? Do the sums add up to fuller cinemas and a net increase in box office?

I find the whole model curious, and I’m really uncertain that it’ll work. But I’ll be paying close attention!

Netflix: $8 Billion and 700 NEW Shows?

How much programming is Netflix actually making?

The answer is a lot, but I think that the widely reported numbers are a little misleading.

Heavily retweeted earlier today was this:

I’m not trying to pick on one person; these are figures that have been reported elsewhere.

Most pieces reference a Variety story: Netflix Eyeing Total of About 700 Original Series in 2018. But you’ll note that the Variety headline includes the word “total” in it.

The key section of Variety’s report is this:

The “700-range” figure [Netflix CFO, David Wells] cited includes 80 non-English-language original productions from outside the U.S., such as psychological thriller “Dark” from Germany and “Club de Cuervos” from Mexico. The total encompasses both new and existing original series (such as “Orange Is the New Black” and “Narcos”). [My emphasis]

In other words, this is a cumulative figure and represents the total number of original series on the platform.

It does not mean an additional 700 originals!

The Variety report is based on an investment call that Netflix had, and as is the way with these things, the transcript of the call is available online.

Here’s the relevant section:

Unidentified Analyst
Right. So moving from maybe the big-picture stuff to more into here now. What are your priorities for 2018? Where are you focused and where is the team focused in making sure the company executes this year?

David B. Wells – Netflix, Inc. – CFO & Principal Accounting Officer
Well, I think — a lot of what you hear many of us say is internal execution, right? So we think we have a large market. We just talked about there’s so many more nonmembers than there are members, and so our focus is really to continue to improve the product that we have. We’ll be adding increasingly more and more of our originals in our global content. This year, we’ll have 80 originals in the global category, meaning these are non-English language original produced content things, like Club de Cuervos, Dark — O Mecanismo is a new one coming from Brazil. And so the — our muscle in that area is increasingly being built and exercised, and I’m excited about lots of great stories coming from different parts of the world. And again, people seem to love high production quality and a good story. It doesn’t really matter where it comes from. So I think our focus is building out our production muscle, building out our global production muscle, increasing our product in various parts of the world. We’re the newest in Asia. So I’d say it’s continuing to sort of localize pieces in Asia, continue to improve the product there. But we also have an eye towards not losing our leadership position in other parts of the world as well. So it’s not like we’re not also improving the Americas.

Unidentified Analyst
You mentioned 80 global originals. That’s TV series, so that’s distinct from your film strategy?

David B. Wells – Netflix, Inc. – CFO & Principal Accounting Officer
Yes. That’s distinct to film, and it’s even distinct from television series that you might describe as sort of global, like Orange Is the New Black or Narcos. These are things that are produced in a non-English language market. So I just want to make that distinction. So there’s even more than 80 that are sort of for the global market. If you think about the total number, it might be somewhere in the 700 range.

That makes clear that there are 80 original “global” originals – non-English language originals. And there are 700 in total. They obviously measure movies differently, and categorise them separately, but then they are still both commissioning original movies and also buying them outright after festivals such as Sundance, beyond the regular licencing of movies from studios. Ted Sarandos, Netflix’s Chief Content Officer has previously said that they will release 80 original movies in 2018.

But how do you even determine what is a Netflix original? It’s not that simple.

Stranger Things or Narcos are relatively simple. They’re 100% Netflix. But for others it’s less clear. For example, in the US, the science fiction series The Expanse appears on SyFy, but it counts as a Netflix original in much of the rest of the world. Star Trek: Discovery appears in the US on the CBS All Access streaming platform. Everywhere else it’s a Netflix Original. Troy: Fall of a City is currently airing on BBC One and was co-commissioned by both the BBC and Netflix where it’ll appear globally.

Even seemingly homegrown series like Orange is the New Black and House of Cards, aren’t strictly Netflix exclusive. Orange is the New Black is currently airing on the Sony Crime channel in the UK, having done a deal with Lionsgate the producers. In France House of Cards originally aired on Canal+ since there was no Netflix in France and the producer, MRC, was able to sell it to them. On more recent 100% Netflix commissions, it has reportedly tightened contracts to prevent that programming appearing elsewhere – unless they choose to allow it.

In any event, a Mashable report makes clear that this 700 number includes some of these co-commissioned series:

A Netflix representative told Mashable that this content budget includes properties we already know and love like Stranger Things, as well as licensing content from partners like AMC’s The Walking Dead.

Note that The Walking Dead is not available on, for example, UK Netflix, because Fox International has the rights and they distribute it on Sky’s platform in boxsets.

It should also be pointed out that “originals” can include one-offs as well as series or seasons of shows. Think about all the stand-up comedy specials that Netflix is commissioning.

So to summarise, there will be 700 originals in total at the end of 2018, which includes new commissions, previous commissions and co-commissions.

Netflix is definitely spending a lot, although it’s in the ballpark of what other large media companies also spend each year. But it’s not launching new series at the rate of two a day!

They’re also losing money – negative free cash flow in the parlance. I’m not arguing that there isn’t an underlying business model that makes sense, but it’s worth noting all the same. The theory is that as they build up their library of originals, they don’t have to licence as much third party material (See also the recent news that Disney won’t renew their Netflix deal and will shift their output to their own new streaming platform).

Netflix faces the issue of needing to have relevant programming in multiple local territories, and while there’s value in older series, viewers will continue to seek new programming. Netflix will have complex calculations about how much it needs to spend on new programming versus catalogue versus subscriber growth versus how much it licences. It’s a complex grid.

Snow and Mist

Snow and Mist from Adam Bowie on Vimeo.

The country has been covered with snow for the last week or so, but it’s not straightforward to get some spectacular drone shots because of the weather. Consumer drones aren’t capable of flying while it’s snowing. And you also have to consider wind speed, and there’s been quite a lot of that.

So my only practicable solution was to get up very early in the morning. Although a fresh fall of snow had been dropped the previous afternoon, and overnight the temperatures had remained sub-zero, but this morning the melt was very much on.

I shot this video and these pictures during a misty dawn. There was still plenty of snow on the ground, although it would disappear fairly rapidly as the day went on. The key thing to always remember with snow photography is that you need to increase the exposure beyond where the camera thinks it should be.

On Sky Q, Netflix and BARB

Note: This is a subject likely to be of even more niche interest than many of my other blogs here. You have been warned!

Yesterday Sky announced that it had reached agreement with Netflix to add Netflix to their Sky Q platform. What this means is that Netflix programmes will appear within the wider Sky Q ecosystem. It means that individual programmes will be promoted without necessarily having to dive into a Netflix “app” as you would on many TV platforms.

On Twitter this raised a few questions: Sky’s on BARB, so might we see ratings for Netflix in due course? Does this mean that Sky will know what their customers are watching on Netflix?

The first thing to say is that this Sky deal isn’t unique. On Amazon’s Fire TV platform, for example, individual Netflix programmes are recommended alongside Amazon’s own shows. You can dive straight into an episode of Stranger Things if you want to (and have obviously previously logged into your Netflix account). And that probably means that Amazon has some idea of what programmes their customers are watching.

I’ve no doubt that the carriage agreement between Netflix and Amazon stipulates that data can’t be shared, since as we all know, Netflix doesn’t publish ratings figures. But Amazon probably has some view on that. Of course, it’s only a partial view since there are a multiplicity of ways to get Netflix on your TV, from built in Smart TV apps, to games consoles, Chromecast, Apple TV and many others. But you know, if you have a big enough sample… I’ll return to samples later.

It’s worth noting that a multiplicity of companies potentially have some kind of awareness of how Netflix is performing. Netflix has a complex Content Delivery Network (CDN) globally to ensure that you get what you want when you want it and where you want it. More importantly, some larger Internet Service Providers (ISPs) are part of Netflix’s Open Connect network. In other words, the ISPs cache Netflix’s programming locally on their own data servers, speeding up delivery time to viewers, but also saving the ISPs money in terms of connectivity with the wider internet. Both the CDNs and ISPs have the potential to understand in great detail what is being consumed (albeit there may be encryption and obfuscation to limit or prevent this).

One thing that Netflix being on Sky doesn’t mean is that there will be BARB data. Recall that Netflix is already on other TV platforms such as Virgin Media. It’s also worth noting that even if all parties wanted to, BARB isn’t really set-up to provide meaningful data at this granularity, and that’s to do with sample sizes.

There are about 26m TV households in the UK, each of which has one or more people. BARB is the UK TV ratings organisation and it recruits a panel of households that provide detailed viewer data which is used to determine overall ratings. There are roughly 5,100 BARB households, with each BARB household representing around 5,000 other households.

That’s a decent sample; it’s large enough to provide robust data on programmes across the larger TV networks. (Things do get a bit more complex with smaller channels, and I’m always wary of anyone telling you how few people watched a show on a minority interest channel, since the sample doesn’t really cater for it). BARB has to work across each TV region, since advertising is sold on a regional basis.

But back to samples, and specifically Sky Q. How many Sky Q households are there in the UK?

In late January, Sky’s results for the six months until 31 December 2017 revealed that there are now 2m Sky Q homes in the UK and Ireland. BARB only measures UK TV viewing, but we’ll use 2m as a working number, alongside 26m TV homes. BARB needs to make sure that it’s panel is reflective of how people actually watch TV – the right number of Freeview, Sky Q, Sky+, Freesat, Virgin Media homes and so on.

Sky’s figures suggest that 7.7% of UK homes, roughly, have Sky Q. If that’s replicated in BARB’s 5,100 panel, then we’re talking about 392 BARB households with Sky Q.

In a paper published in January 2018, BARB says that 7.5m homes have Netflix subscriptions in the UK, or 22% of homes (This is based on a large survey that BARB conducts called the BARB Establishment Survey. It’s used to ensure that BARB knows properly how the UK public is watching TV and video).

As things stand then, assuming that Netflix subscribers are spread evenly across all TV platforms (a massive “if”), then we need to further reduce our sample to find Sky Q homes who currently have Netflix. It seems likely that current subscribers would be able to link their accounts to the Sky Q platform, and of course Sky and Netflix would like to see subscriber growth coming from those homes.

Let’s be generous and assume that instead of 22% of Sky Q homes having Netflix, it’s likelier to be closer to 50%, based on the greater disposable incomes of those households and them likelier to be earlier adopters. That means our sample is down to 196 homes. And that’s a very small sample to start trying to model what those households are watching on the Netflix platform.

This is all moot anyway, since unless Netflix changes tack radically, they don’t need to join ratings bodies like BARB. They’re not an advertising led business (the reason commercial services need BARB), and they know precisely how much their shows were streamed.

These digital providers are rolling big dice. Netflix says that it will spend close to $8bn on programming this year. Amazon is expected to spend $5bn. Facebook and Apple have promised to spend big too.

Recode has an interesting chart that puts 2017 spends in context.

But we probably need to contextualise that a little because Netflix and Amazon are spending globally, while many of the other players on this chart are predominantly US companies with relatively little ex-US programme spending.

I’m not sure where any of this gets us except that these digital streaming platforms are going to be everywhere even more, and we’ll probably never get a clear picture of how popular much of their programming really is.