Data is the oil of the 21st Century, but data democratization is a term that is hard to get a handle on. Let's talk about democratization of "Data Proceeds"
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Cognitivo PoV - Market power in the data-driven economy
1. Balancing market power in
the data sharing economy
Nov 2019
contact@cognitivo.com.au
www.cognitivo.com.au
2. Our lives are increasingly mirrored online
2
Online (facilitated)
Interaction
Wearables / IoT Digital Twin
+
• Over 2.5 quintillion bytes of data are created every single day and it’s only going to grow from there.
• 90% of the world’s has been data created in the past 2 years.
• By 2020, there will be more than 50b devices connected to the internet, that’s over 8 devices per person.
• Humanity is undergoing a digital metamorphosis.
3. What you are
interested in
There’s a race to “monetise” data
3
What you
buy
Who you are
Where you buy
When you buy
Where you go
Those that are creating a data honey pot (with
hopes of being bought out)
Those using data to create a genuinely great
experience to compete for customers
4. Consumers’ online attention is concentrated on
certain channels
4
News
outlets
Rev: $136b USD
Market Cap: 725b
Rev: $55.8b USD
M. Cap: 519b USD
R $110.36b USD
M. Cap: 779.7b
R: $265.5b
USD
M. Cap: 700b
Other
(e.g. gmail)
Search
Youtube
20.5%
Facebook
Messenger
Instagram
WhatsApp
18.6%
Office365 /
Outlook.com
3.4%
2.3% 2.2% 2.3%
Revenue $9b USD
M. Cap ~8b USD
R: $1b USD
M. Cap:
~20b
Source: Nielsen Digital Panel Feb 2019,
Cognitivo Research Revenue and Market Value as at Dec 31 2018
Note: Time spent on other websites: 50.8%
• US based search, content
aggregators and social
platforms dominate our
online-attention
• ~40% of all Australia’s time
spent online relates to
Google of Facebook
• Telco’s are restricted to
compete in this space
• It is clear that traditional
news / media / advertising
firms cannot compete this
is reflected in market
values
ShareofTimeSpentOnline
5. This concentration of market power is not
unprecedented
5
• Standard Oil worth $1T in 1900 (inflation adjusted),
controlled ~90% of the refined oil in the US. Broken up
into 34 companies in 1911
• Microsoft found to have a monopoly, ordered to be
broken into 2 (OS vs Software), later appealed and
ordered to share API’s with 3rd parties.
Source: GS.statcounter.com
• Today, Google has >90% of the global search market
and Facebook has ~70% market share in social
media.
6. There are growing calls by politicians to invoke
anti-trust legislation against digital giants
6
However, the world is a very different place today and the US isn’t the only competitor for leadership in the digital
economy
1890
8. Democratisation of Data “proceeds” is coming
8
Andrew Yang – US 2020 Presidential candidate
• “It’s not like any of us wants to use the fourth-
best navigation app” That would be like cruel
and unusual punishment. There is a reason why
no one is using Bing today.”
• “We say this [data] is our property and we
share in the gains, that’s the best way we can
balance the scales against the big tech
companies.”
Democratisation is about how every citizen can
benefit from a unique resource, without eroding the
competitiveness of that very resource.
9. 3 key aspects of Data Market Power need to be in
balance for healthy competition
9
Competition
Data Protection
& Privacy
Consumer
Protection
Compatibility and
portability of data
increases competition
Offering consumers
choice and lower prices
Consent options and disclosure that is easy for consumers
to understand and are adhered to by businesses
Competitive, data-driven
markets competing for
well-informed
consumers on all
dimensions of price and
quality, including level of
privacy protection
Addressed by: Open
Data Regulation
Addressed by: GDPR / Australian Privacy Principles++
Source: Adapted from European Data Protection Supervisor (2014) and ACCC (2019)
Addressed by: Anti-trust
and Competition Laws
10. Deep mistrust in institutions, both governments
and companies, will stifle innovation
10
Almost 10 per cent of eligible Australians
have opted out of the My Health Record
- Australian Digital Health Agency
The future of the digital economy
relies on trust, by both
consumers and business users.
The Productivity Commission has
noted:
People need to “feel they have
some control over how their own
data is used and by whom, and
an inalienable ability to choose
to experience some of the
benefits of these uses
themselves” - ACCC
47% of Australia distrust both Governments and the Media
11. Technology enables new business models and can
be used to build consumer trust
Differential Privacy
Fine-grain Data
Management & ABAC
Homomorphic
EncryptionConfidential Computing
(SMPC)
Distributed
Analytics
Safe Harbour
Tokenisation
R4 Tool
Input parties Computation nodes Result nodes Result party
Re-identification
Risk
Assurance
Federated Learning
Consensus &
Micro-currencies
OrganisationalModel Bus. ClassificationScheme
Property Local Laws
Stat
Planning
Health &
Disability
Parks &
Infra
Asset Mgt.
Leisure &
Sport
Community
Dev
Sustainable
Living
Building
Support
Strategic
Planning
Projects &
Strategy
Common Functions
Finance IT Legal Risk
ConceptualDataModel
Asset
Customer
Arrangement
Event
1.CRM
2.Finance
Org
Sales HR Operations
Corporate
Services
Event Asset
Risk
H
M
Data ObligationsLibrary
Policylandscape:
Privacy
Records Retention &
Disposal
Confidentiality & PCIDSS
https://www.socialmediatoday.com/news/how-much-data-is-generated-every-minute-infographic-1/525692/
Over 2.5 quintillion bytes of data are created every single day, and it’s only going to grow from there. By 2020, it’s estimated that 1.7MB of data will be created every second for every person on earth.
Statistic about 90% world’s data created in past 2 yrs
https://www.sec.gov/Archives/edgar/data/1652044/000165204419000004/goog10-kq42018.htm
What do we does a domestic economy in aus to comepte with these guys? Digital products are global, australia doesn't play in this space
What do we do with the telcos? They know everything but their value has been eroded by whatsapp – we could open a domestic capability to compete with this, we could set a regulatory framework to enhance local competitiveness against intl competitors
A lot of the oil companies still survive today
On June 7, 2000, the court ordered a breakup of Microsoft as its remedy.[19] According to that judgment, Microsoft would have to be broken into two separate units, one to produce the operating system, and one to produce other software components.[18][20]
The D.C. Circuit Court of Appeals overturned Judge Jackson's rulings against Microsoft
The proposed settlement required Microsoft to share its application programming interfaces with third-party companies
Sherman act – was this the basis for breaking up the oil companies -
We chat becoming a leader in NLP
Dji is drones
Ubitech robotics
Competitiors in other markets have access to talent, data, a loose regulatory environment and they will develop capabilities and companies to support their economy for the next 200 years
- democratisation of proceeds because data itself can be copied
- it doesn't make sense to break up companies that are doing really well but we need competition, everyone should share in the gains
- every citizen can benefit from a unique resource without eroding the competitiveness of that resource (e.g. the oil dividends – we want the digital dividend/data dividend)
Ensuring market power is balanced
Getting the balance between the three forces in the circles is vital – this is what the ACCC has to look at – considering perception of digital giants havin too much market power
In the past, we have had initiatives e.g. me and my bank account online – MAMBO – about bank account portability, check if there Is something similar in europe, open banking, open energy and open telco
Link to tech in next slide
The future of the digital economy relies on trust, by both consumers and business users. As the
Productivity Commission has noted
Businesses, as much as governments, rely on the willingness of the public – the source of so much of
the data – to continue to trust data handling and use. Against the background of an ocean of personal
data that is already public, there is now, and will be in the future, a need for continued community
acceptance and trust in the handling of personal data by both governments and business.
Social licence will develop if people:
have a sound basis for believing in the integrity and accountability of entities (public and private)
handling data
People need to “feel they have some control over how their own data is used and by whom, and an inalienable
ability to choose to experience some of the benefits of these uses themselves”
turn this into the competition slide
Source: http://www.roymorgan.com/findings/7640-social-media-distrusted-june-2018-201806260211,
https://www.theguardian.com/commentisfree/2018/jan/22/distrust-social-media-traditional-journalism-fake-news
Fine-grain data mgmt & ABAC – build a conceptual data model and define your policy landscape to ensure effective data management
R4 tool – quantify re-identification risk tolerance and evaluate a dataset for compliance with this tolerance level BEFORE data is released
Distributed analytics – business models based on big data analytics are emerging – but considering the existing distrust on behalf of the consumers who provide the data, new techniques such as differential privacy and homomorphic encryption are likely to be important to reassure consumers that their data is being used safely
Differential privacy – noise is added to real data in a way which ensures that the data distribution is maintained - so models can be reliably trained on it – but the data itself is no longer real data so a person’s real information is not leaked, even if a data breach occurs
Homomorphic encryption – enables computation on ciphertexts, rather than real text, without affecting the results
OpenMined – open source community working on these privacy preserving techniques, Cognitivo’s Melb director is the Australian country lead for this
Homomorphic – cypher based encryption that enables an organisation to safely encrypt it’s data while a second party can perform some analytics over the top and will never know the details. SMPC – it’s an encrypted communications protocol.
Federated learning – bring the model to the data – google keyboard