Linked data, attention data, recommendation, ownership, privacy and rights

Introductions

Matthew Wood

Head of Architecture

Michael Smethurst

Information Architect

BBC Audio and Music

...and apologies

for our design skills (hellomatty.com/wod/s5.html)
if it's all a bit BBC
that we're really not experts...

Some things we're not going to be talking about:

Resources (information or otherwise)
Representations
Content negotiation
Ontologies
Triples
RDF
Semantic web

One thing we are going to be talking about:

What makes Tesco so successful?

Tesco: people and groceries

People pay money for groceries

Tesco: the more products going out...

Increasing the number of products going out...

Tesco: the more money coming in

Increases the money coming in

Tesco: people, groceries and attention data

Loyalty cards capture a person's attention data to products

Tesco: diversification

Loyalty card accounts can be linked to postcodes and from there demographics, house prices etc

Tesco: colonisation

Loyalty card accounts can be linked to postcodes and from there demographics, house prices etc

Tesco: people, groceries and connective tissue

Loyalty card accounts can be linked to postcodes and from there demographics, house prices etc

Tesco are linking data

not in an open way
not in a webby way
not in a way that brings value to anyone other than Tesco...

Well almost

Tesco buy back attention data at the rate of one penny in the pound...
...which is interesting because...
...it reverses the traditional producer / consumer roles
...it reverses the traditional flow of capital
...it's the one place where the consumer sets the price :-)

Loyality (cards) boiled down - convenience vs privacy

We sell our privacy for money back, special offers, self scan, home delivery etc

Tesco recognises:

that we live in an attention economy blah, blah, blah
that there's value in well maintained data sets
that there's even more value in the links between those data sets
that by bridging data siloes the aggregate business becomes greater than the sum of its parts

Many other businesses also recognise this:

Google
Amazon
last.fm
Waitrose / John Lewis
Apple / iTunes
Thomson Reuters
Sky

Back to the BBC...

BBC: people and programmes

People pay money for programmes

BBC: pushing programmes to people

Push as many programmes to people as possible

BBC: pushing programmes to people - rights permitting

The biggest problem is rights

BBC: pushing relevant programmes to people

The other problem is identifying relevant programmes to push

More intelligent promotion gives...

...less promotion of content we hope people will like
...more promotion of content we know people will like
...more promotion of content we think people will like (data based recommendation)
Less money spent - happier audience - more public value

BBC: people, programmes and attention data

People express attention to programmes - by watching, listening, commenting, rating etc

BBC: people, programmes and connective tissue

Social graph enables social recommendation

Linking data - Sky+++

Courtesy of:

Sky, NDS, Experian, Dunnhumby - linking Sky accounts with Tesco etc

Direct marketing the data way

Clive Humby

Credit-card data tells you how they live generally, supermarket data tells you their motivations, the media tells you how to talk to them. If you have those three things you're in marketing nirvana.

Convenience and privacy redux

Again trading privacy for convenience

This is not supposed to be a Daily Mail web horror story...

Daily Mail Horror stories

But...

when you get to the level of an individual open your data is clearly naive
there are issues of ownership
issues of privacy
issues of rights

Data ownership

The user should always own their profile, social graph and attention data

Privacy

If users own their data then...
...they should be able to control the privacy of that data
But privacy on the document web ≠ privacy on the data web
Document web privacy = who can see
Data web privacy = who can see + who can use + how they can use it
...but the document web and the data web are merging (REST, semantic HTML, ufs, RDFa)

Rights and licencing

If users want to open their data and control how it's used...
...they need to control how it's licenced
but the law hasn't caught up - You can't copyright facts
so you can't claim ownership of the fact you watched Eastenders last night...
and you can't choose to give away that ownership either

Back to authentication

So if you can't control how your data is used...
...then it's still about controlling who can see it
OAuth etc

Privacy - anonymity via personas

If I had 3 accounts: one on Twitter, one on Wordpress, one on YouPorn...

Privacy - 3rd party claims

..and Matthew came along...

Privacy and linked data - 3rd party triangulation of persona

...he could

owl:sameAs, XFN...

Every time we publish data we need to question...

...what could this data be combined / linked with that could make good things happen
...what could this data be combined / linked with that could make bad things happen
think Holorith Machine???

UX challenges

how do you inform users without scaring them?
how do you ensure informed consent
children!

Informed consent - beyond checkboxes

Twitter, last.fm and Yahoo! checkbox terms and conditions

Finally

What should the BBC's role be?
bbc.co.uk/users/joebloggs
...or...
joe.blogs.org/bbc
encouraging / educating / enabling people to own their own website / data space could be the BBC micro of the 21st Century?!?

Questions...?

...and hopefully answers...
...but maybe not from us
:-)