X Close

Digital Education team blog

Home

Ideas and reflections from UCL's Digital Education team

Menu

Facebook assassins?

By Mira Vogel, on 27 October 2014

Four protesters about Facebooks real name policy

There is a lot of social media use in higher education – see Rey Junco’s research, for example. One reason for this is summarised by David Wiley:

Q: What would happen if Facebook worked like Blackboard?
A: Every 15 weeks Facebook would delete all your photos and status updates and unfriend all your friends.

So here’s bit of a comparison of four social networking services Facebook, Google+ and two upstarts Ello and Tsu – far from the only alternatives, but four of a kind. As a comparison it’s pretty one-dimensional because rather than what the services allow members to do I’m focussing on the preliminary but fundamental matter of how they turn a profit, pieced together from a quick scan of articles.

All are venture capital funded and consequently they either rely on member subscriptions or they commoditise their members by putting pricetags on their biographical data, shares and relationships. In the latter case they depend on members to befriend and be befriended, follow and be followed, feed news and be fed news. Google+ is Google’s answer to Facebook. Facebook probably needs no introduction, and rumours of its death may or may not be exaggerated. At any rate there is currently a disaffection with advert-saturated intrusive business models where members generate profitable content and everybody but the members seems to profit. Ello and Tsu are two such responses. So how do they compare?

Business model

Ello is set to become one of a small number of public benefit corporations in the US – this broadly relates to who owns the assets, how much profit is reinvested in the company, and constraints on the future direction of the company. Ello managed to raise more venture capital investment than usual for this type of business. Its freemium service offers members a basic service for free and charges for advanced features – Ello founders point out that this is basically the successful iPhone app business model, except there’s no need to buy anything at all. Ello don’t mine members’ data.

Tsu calls itself a “combined social network and payment platform”, and members can share from it to Facebook and Twitter. Tsu revenue comes from page views (so favours members who post lots, find lots of new members, and have lots of friends sharing their posts), but members keep a stake in their data and “therefore they own the royalties generated from advertising, sponsorship and partnership dollars wrapped around their content”. Data here includes a member’s network – a portion of the revenue goes to the person who invites a member to Tsu – this ‘parent’ is in line for a kind of finder’s fee every time the member gets a payout, as is their parent, their parent’s parent, in diminishing fractional payouts. There are questions about motivation to join in and share, and the revenue-sharing processes are said to be rather complicated for the average punter – but they are the main innovation.

Facebook uses member data to personalise and target advertising to them. It also has its members advertise to each other by default. It has a real name policy – one account per member, using their own name. For some time there have been privacy concerns about Facebook, such as Instant Personalization which shared members’ demographic data with selected commercial partners.  At the same time, for those whose networks are already established in Facebook there tends to be profound inertia about migrating elsewhere. In 2013  a Facebook user was reportedly (methods subject to debate) worth $98, and a Facebook Like, $174.

Google is an advertising company which seeks to collect user data over time in order to refine its contextual advertising algorithm. To do this, Google wants you to log into one or other of its services. Some have commented that Google+ is less a social network and more a demographic research tool – for some time you only had to be logged into, for example, Android, or gmail, or Chrome, or YouTube) to be passively using Google+, with resulting privacy concerns.  Google+ recently eased up on pressurising users into the service, abandoned its real name policy, and stopped requiring a Google+ account to use Google Hangouts. It’s currently hard to work out its business model, and this has led some to question its future.

On screen advertising

Facebook depends on its members’ profile and user data to serve demographically-targeted adverts to its members which appear throughout the site. Its terms require members to use one account in their own name and be truthful in their profile. Because members use Facebook to socialise rather than view content, it has a relatively low rate of clicks on the adverts it serves but its sheer numbers of members compensates for this. Contextual advertising is also Google’s line – although Google+ doesn’t itself currently have on-screen ads, they aren’t ruled out and in any case there are ads on Google Search and other Google services. Ello‘s charter [PDF] commits it to be free of “paid advertising on behalf of a third party” and it doesn’t mind what you put in your profile. Tsu has advertising.

What’s at stake?

There has been a recent spate of class actions against intrusive, exploitative practices by social media companies with advertising-driven business models – not least this humungous one – and these are meeting with some success. Concerns about algorithmic manipulation were popularised by Eli Pariser in his 2012 book The Filter Bubble, and are a major theme in technology commentary – see for example Zeynep Tufeksi on Twitter’s roll-out of algorithmic curation. The problem (which is particularly antithetical to academia) is that algorithms opaquely serve members things they are calculated to like, and people tend to prefer agreeable or affirming stuff over challenging or troublesome stuff (you won’t be surprised to learn there are algorithmic answers to that; there are also alternatives such as Pariser’s Upworthy which piggy backs on other algorithms, although that’s a balancing act). The question is the same one people are asking about learning analytics: is the algorithm secret and commercially sensitive or is it available for scrutiny? It’s very early days for Ello – it is currently nowhere near as lubricated as the leading social networks in terms of sharing between networks, for example, but Ello’s is the only business model where members will explicitly not be subjected to data manipulation. Some commentators urge a more dispersed form of social networking – for example the POSSE, or ‘Publish (on your) Own Site, Syndicate Elsewhere’ approach allows members to share posts from one place (here, their own) to many other places.

More

Are any of these things a factor in your choice of social network? Perhaps not. Feel free to poke holes. On Facebook privacy see this snappy guidance from Sally Burr at the University of Sussex. Finally, how to join. With Tsu you need to be invited – it’s very easy to get invited but given the money involved I’m not going to recommend any one or other Tsu parent, so search for somebody you’re happy to generate revenue for. At this stage with Ello you also have to be invited, but you can request an invitation from its homepage. You can join the million-plus waiting list, or look out for somebody to invite you. Google+ and Facebook are open to anyone over 13.

*Image source: Facebook forces drag queens and trans people to use “real names” – marked by Google Image Search as ‘Labeled for reuse with modification’.

HT Natasa Perovic and Niv Setru for some of the links

Moogle Analytics

By Matt Jenner, on 20 April 2012

We have had UCL Moodle linked to Google Analytics for the past three or so years, the data this allows us to see about Moodle usage and browsing habits is getting increasingly interesting. The concept of looking into data for measuring learning is obviously bonkers, no amount of pure data of browsing behaviour of Moodle will give indication about any actual learning taking place. Observe a classroom or people on the bus, you can see how many people are there and what they’re doing, but you won’t know what they’re thinking, what connections are being made etc. But, do not despair, there is a reason for this post – albeit it’s a bit dry and technical…

What Google Analytics does is give us an anonymous view into what happens when someone comes onto Moodle. For example, you do this:

  1. Browse to the Moodle homepage
  2. Login
  3. View a course
  4. Send a forum message
  5. Log out

Data Trail

You’ve just left a data trail behind you about your computer (not you). This trail isn’t used for snooping on you personally, it’s anonymous, it also isn’t used for snooping at all. What it does, and I hope you believe me, is leave data tracks telling us something useful. For example, let’s look at the above and break it down.

You and your browser

Google analytics can find this information out:

  • Web browser (Internet Explorer 9)
  • Operating system (Windows 7)
  • Device-ish (only really useful for mobile devices like smartphones)
  • Rough location (London, UK)
  • Internet Provider (University College London)
  • IP Address (your computers’ address to the internet)
  • Web software capabilities (Flash, Java, Javascript)
  • Some other stuff (passport, NI number, date of birth, name – all not included!!!)

Don’t be scared about this, none of the data can be used to identify you. As a side note: if you’re worried about UCL Moodle collecting this data, remember that for other sites (such as the big ones outside of UCL which you may use daily) your data is their product, and you give them nearly everything about you, this data is only about your computer, but let’s not go down that route, this is about learning…

Other data collected

Google’s Analytics also collects data about your pages you view; so for example each page you load is marked as a ‘view’ on that page. Pages with higher numbers of views are clearly more popular, for whatever reason. Then there’s unique views, one person viewing a page ten times means one thing, ten people viewing the page once each means something else. Adding to this there are other metrics, such as time on page, page exits (if they leave the page, can we see where they go) and eventually building pathways through a site.

So, what’s my point?

What Google reports is really hard to read into but there’s a few changes to Moodle code (really small changes) which makes this data far more useful. Over the coming year I hope we can make these changes, and next year, report back on what the data says about Moodle usage. We want to know what’s popular (read: working) and what isn’t (read: not?).

Informative and technical links:

Perhaps mostly useful for the technically inclined, but it’s what we work with to make Moodle more useful for you.

More soon!