The document discusses how operators should view and utilize big data. It provides perspectives from industry experts on how big data analytics can provide value if aligned with business strategy, how different tools may be useful for different business functions, and how big data can integrate with existing analytics environments. The experts discuss starting with pilots and focusing on clear goals, selecting tools based on needs, and emphasizing interpretation of results over just collecting more data.

1. 6 6
Should operators be aware and conscious of Big
Data – as opposed to seeing it as just another
term used by the IT department?
Ales Gornjec: Operators should consider deployment
and use of big data analytics very seriously. The term is
perhaps new, but the concepts have existed in the gam-
ing industry for a long time. In the past only the biggest
operators would invest into infrastructure that allowed
them to perform good analytics, but nowadays this
technology is available to wider audience as long as
they have a strategy and clear vision which data to col-
lect and what to do with it.
Jerry Bowskill: Big Data is more than hype; it’s a useful
label to help focus organisations on the potential benefit
of extracting greater value from data. However, because
this is an umbrella term with no clear definition, the
specific implementation and strategy is not prescriptive
and up for interpretation. Rather than focus on Big Data,
the true value here is the Big Opportunity that is emerg-
ing in today’s data rich business environments.
Jeremy Thompson Hill:Yes, operators should
absolutely be aware of the opportunities that Big Data
presents to their business – and the threat a lack of
understanding and adoption will inevitably bring. For
an operator, Big Data represents the chance to under-
stand their customer and act to ensure that their offer-
ing is truly relevant and engaging. No matter the trade,
the customer should always be placed at the centre of
any business and its strategy .
Big Data is an umbrella term for a whole group of
analytics tools and technologies – how do you
breakdown the tools that are actually needed for
specific purposes? Many gaming operations are
heavily focused on customer-oriented analytical
applications involving personalisation, ad as well
as email targeting, and search engine optimisa-
tion, in addition to quantitative techniques involv-
ing data mining, marketing, and experimental
design. How do you decide upon the Big Data tools
needed for specific businesses both now and in
the future?
Jerry Bowskill: A Big Data strategy, first and foremost,
needs to be completely aligned with the commercial
strategy of the organisation. Being clear from outset,
how Big Data can deliver commercial benefit is a criti-
cal pre-requisite to any technical development or capital
expenditure. Opportunities will vary across different
businesses and therefore the tools, data and skills
required will vary accordingly. For Scientific Games, our
approach has been to focus initially on creating a solid
data architecture on which our gaming platforms are
built, feeding a data warehouse that can support the vol-
umes of transactions generated by our ‘omni-present’
01 Data Lake - is a massive, easily
accessible data repository for
storing "big data". Unlike traditional
data warehouses, which are optimised
for data analysis by storing only some
attributes and dropping data below the
level aggregation, a data lake is
designed to retain all attributes,
especially when you do not yet know
what the scope of data or its use.
Currently, Hadoop is the most
common technology to create a data
lake. It is important to distinguish the
difference between Hadoop and a
data lake. A data lake is a concept, and
Hadoop is a technology to implement
the concept.
TECHNOLOGY - BIG DATA
Interactive
Data - the bigger
the better?
Everyone collects data - the trick is in
using it to the best advantage. We
interview the Big Data players in the
gaming industry, speaking to
Comtrade, Playtech, Openbet,
Scientific Games Interactive and
TapCentive about the importance of
this technology for the future of the
land-based and interactive industry.
Jeremy Thompson Hill,
CEO, Openbet
Ales Gornjec,
GeneralManager, Comtrade
Jerry Bowskill,
Chief Technology Officer,
Scientific Games Interactive

2. 6 7
customers across their landbased and interactive sites.
With all data formatted and available, we could then
deploy any toolset, giving us the flexibility to evolve the
types of tools as our needs change over time. It is impor-
tant, as a technology company, not to fall for the trap of
thinking it is cost effective to develop in-house tools.
Our advice is to select from the vibrant community of
tool suppliers. At Scientific Games, our initial approach
has been to use a small number of analysis toolsets
offered by companies such as IBM and Google, as the
needs differ between our Social and Real Money Gaming
products.
Jeremy Thompson Hill: There is not one right answer
here as we are all on a journey towards creating a more
customer-centric approach to business and one gaming
operator is always going to have a different product set
up and set of priorities to the next.
That said, in the main, operators are trying to retro fit
new approaches, tools and technology into existing
infrastructure and operating systems, rather than start-
ing with a blank canvas, so the decision as to which
tools to go for tends to be driven by multiple teams
across the business. For example, a marketing or prod-
uct team might decide to utilise a tool like Maximiser to
drive multiple content variants to specific user segments
whereas a data insight team could implement a specific
business intelligence tool that can be used to extract key
insights into the business.
At OpenBet we know that the velocity of technological
change in the market means we need to be flexible
enough to cater for the requirements of our customers,
regardless of their chosen Big Data set up. As a result,
we are moving towards a truly 'open access' data
approach where our platform becomes an enabler for
growing a Big Data solution through APIs rather than a
blocker trying to force operators down a specific route.
Ales Gornjec: There are many different areas within
gaming operations that can benefit from a Big Data
approach, and we would recommend different tools for
them. To start with IT, there are now various tools avail-
able that collect all kind of system logs in one place and
help IT to find potential problems proactively, and to
troubleshoot and solve those that have already hap-
pened more efficiently. Splunk is the leading commer-
cial option, often used in gaming, but we have also
found an open source combination of Elastic with
Logstash, and Kibana very useful.
On the marketing side, there are many tools that can
increase targeting precision, improve relevance, and
increase campaign efficiency. Google Analytics has
been used for many years now to improve web site
design and to understand user acquisition channels. IBM
Unica was quite popular within the gaming industry to
drive campaigns. Moreover, quite a lot of operators are
now linking and improving efficiency with results of Big
Data analytics. These links are often still indirect and
not automated because of an additional integration
investment that is necessary. There are many tools
available today that help in various areas of marketing
activities. They are usually not integrated into one cohe-
sive solution, and substantial manual work is needed to
gather all the data together, process it, and start the nec-
essary actions. Usually this work is split between the IT
and marketing departments, where good business ana-
lysts and data scientists can have important roles in
bridging the gap between business and technology
mind-sets.
Big companies with large investments in their data
warehouses have neither the resources nor the will
to simply replace an environment that works well
doing what it was designed to do – it’s also not
practical to acquire all the Big Data-enabling tech-
nologies needed in one fell swoop. Assuming the
business can establish acquisition tiers for key Big
Data solutions, what are the corresponding budget
tiers?
Jeremy Thompson Hill: OpenBet provides multiple
levels of data solutions to meet the needs of our cus-
tomers and commercially the business is moving more
towards a shared risk or license fee model. Our data
solutions are costed based on volume of data and also
tangible results. As we make data feeds available to cus-
It is important, as a technology
company, not to fall for the trap
of thinking it is cost effective to
develop in-house tools.
TECHNOLOGY - BIG DATA
Interactive
Hadoop cluster
A Hadoop cluster is a special type of computational cluster designed specifically for storing and analysing huge
amounts of unstructured data in a distributed computing environment. Such clusters run Hadoop's open source dis-
tributed processing software on low-cost commodity computers. Typically one machine in the cluster is designated
as the NameNode and another machine the as JobTracker; these are the masters. The rest of the machines in the
cluster act as both DataNode and TaskTracker; these are the slaves. Hadoop clusters are often referred to as
"shared nothing" systems because the only thing that is shared between nodes is the network that connects them.
Hadoop clusters are known for boosting the speed of data analysis applications. They also are highly scalable: If a
cluster's processing power is overwhelmed by growing volumes of data, additional cluster nodes can be added to
increase throughput. Hadoop clusters also are highly resistant to failure because each piece of data is copied onto
other cluster nodes, which ensures that the data is not lost if one node fails.
As of early 2013, Facebook was recognised as having the largest Hadoop cluster in the world. Other prominent users
include Google, Yahoo and IBM.

3. 6 8
tomers the commercial model for the business is pivot-
ing towards a more results-oriented approach where
both parties are motivated to drive success.
Ales Gornjec: Big data solutions are available in vari-
ous business models – usually cloud-based solutions
have some volume- or usage-based pricing structures,
where the initial investment is kept low and ongoing
costs depend on the volume of data, the number of play-
ers, or some similar parameter. Many operators want to
protect their data and are interested in on-premise big
data installations – in such a case the initial setup cost is
higher, since it also includes hardware and licenses, but
has a lower ongoing component that is usually struc-
tured as a license support.
Jerry Bowskill: A Big Data strategy does not require an
organisation to mothball any of their existing data
warehousing or business intelligence infrastructure.
Just because an organisation is running a Hadoop clus-
ter does not automatically mean that it has an effective
Big Data strategy. The first tier should be to build on
what you have for a very well defined goal. With mod-
est changes it is possible to get select amounts of dis-
parate information connected using simple database
technologies and analysed by opensource tools.
Running a pilot doesn’t need to break the bank; start by
trying to answer a simple question that adds value to
the business, and get a feel for the tools and resourcing
before moving forward. At Scientific Games we get an
enormous amount of benefit on our Social and free-
play games from Google Analytics and this has influ-
enced where we have invested in our Real Money
Gaming analytics. Ultimately the power of business
intelligence, over more traditional reporting, is that it
allows scenarios to be run and data to be segmented in
real-time so you can get results or spot trends in some
cases before you really know how to articulate the
question.
Firms such as Google, eBay, and Facebook were
built around Big Data from the outset. They didn’t
have to reconcile or integrate Big Data with more
traditional sources of data. The same isn’t true for
the majority of gaming industry businesses, which
need Big Data to co-exist with analytics on other
types of data. Can Big Data fit into the overall data
and analytics environment of the gaming indus-
try?
Ales Gornjec: This depends on the data that is collect-
ed. IT infrastructure logs are new, but sometimes have
similar information as server-based gaming systems or
central monitoring solutions that collect from gaming
machines. These solutions collect structured informa-
tion that is defined by Gaming Standards Association’s
G2S protocol, while big data logs contain unstructured
information. Both can be used to search for potential
fraud, but in a different way.
Player data on the other side can be extended with
activity information that add game play and payment
transactions that are already stored by existing systems.
This could include browsing activity on operator portals
and social sites like Facebook and Twitter. Operators
already monitor user activity on the portals with Google
Analytics.
Jerry Bowskill: I would take a different perspective
here and say that Google, ebay, Amazon and Facebook
are all great examples of organisations that have fully
embraced a data-driven, evidence-based culture
throughout every part of their organisation. They are
obsessed with data and using this to make the most
robust, accurate and timely decisions possible. In this
respect, Big Data is more of a philosophy rather than a
specific technology solution. For SG Interactive, having
a big data philosophy is essential to the success of our
social gaming products. With millions of players per day
and tough competition, we have to AB test our games
and marketing campaigns; failing fast and building on
success in near real-time is an operational necessity.
Our Real Money games have a big data mindset, devel-
oped from our experiences with Social gaming, which
has helped us better understand player and product
behaviour. This ultimately allows us to create better
games for our customers to get a greater ROI from mar-
keting them because they are not tied to a set number
of reports. Rather, the games have the tools to run
whatever analysis they like.
Jeremy Thompson Hill: Absolutely. Operators who
have been using more progressive algorithms to under-
stand customer behaviour in greater depth and utilise
this information to define a user-centric front end expe-
rience for a customer are winning the battle in the mar-
ket. There are operators in the market that have already
made this shift.
For businesses that are still on this journey, the majority
in the gaming market, there is always going to be an ele-
01 Hadoop is talked about as if it's one
monolithic thing, but it's actually a
family of open-source products and
technologies overseen by the Apache
Software Foundation (ASF).
Theoretically, HDFS (Hadoop
Distributed File System) can manage
the storage and access of any data
type as long as you can put the data in
a file and copy that file into HDFS. As
outrageously simplistic as that
sounds, it's largely true, and it's exactly
what brings many users to Apache
HDFS and related Hadoop products.
After all, many types of big data that
require analysis are inherently file
based, such as Web logs, XML files,
and personal productivity documents.
TECHNOLOGY - BIG DATA
Interactive

4. 6 9
ment of stitching different data sets together to build a
more granular picture of the customer. Transactional
data, front end analytics, payments, demographic data,
game paytables and features - the list goes on and while
it is not straightforward to connect these data sets
together the benefits are significant for Operators who
do.
OpenBet has partnered with a business to help solve
these complex problems and to allow the business to
move forward aggressively into predictive analytics and
machine learning. It is a new partnership for OpenBet
but we have high hopes for the output. The reality is,
complex backend tools that need a PHD-level of under-
standing to use efficiently are going to be obsolete in
years to come. Moving forward, the battleground centres
around automation and machine learning.
Does the collection and storage of increasing vol-
umes of data mean we are ultimately better
informed? And how do you go about turning cus-
tomer data into customer loyalty?
Jerry Bowskill: Potentially yes - The collection and
storage of more data does create the opportunity to
make better decisions, however it also introduced a risk
of more spurious correlations. Advances in medical
imaging and analysis provide a clear example of one of
the potential problems. It actually makes it harder to
recognize what’s ‘normal,’ as for the first time having
more data has illustrated how greatly individual param-
eters vary between patients. Differentiating between
correlation and causation is a potential negative unin-
tended consequence of Big Data. In this respect, the
technology may allow us to more effectively store and
analyse data. However, this places a greater emphasis
on the interpretation and business implementation of
the results. Perhaps the ‘AB testing’ of players provides a
non-ambiguous example of where analysis can reveal
the real underlying preferences for changes between
products variants.
The reality is, complex backend
tools that need a PHD-level of
understanding to use efficiently
are going to be obsolete in years
to come.
TECHNOLOGY - BIG DATA
Interactive
Even bigger terminology
Terminology like megabytes, gigabytes or perhaps even terabytes already become out-dated. Today we are talking
about Exabytes, Petabytes, Zettabytes, Yottabytes or even Brontobytes. It is estimated that the current digital uni-
verse is 1 Yottabyte in size. This is the equivalent of 250 trillion DVD’s and our digital universe is expanding at an
exponential rate.
Jeremy Thompson Hill: This is a key question - data
does not automatically lead to value for a business, it is
the interpretation of the data, the extraction of action-
able insights and the automation of processes that gen-
erate value. A business can drown in insight data but
remain lost. Give me one actionable insight over a
mountain of data any day of the week.
The question of how you turn insights into greater cus-
tomer loyalty is answered by an efficient product devel-
opment methodology. If we listen to our customers they
will show us. In the words of Silicon Valley entrepreneur
Eric Ries: ”We must learn what customers really want,
not what they say they want or what we think they
should want.” The key to the process is starting from
customer insights and putting in place a process that
allows the business to surface key insights and enhance
a product iteratively to ensure that it is both relevant
and engaging. Once this is achieved then keeping cus-
tomers loyal is a much easier process. Price will, of
course, always be a factor for someone who likes a bet
but I firmly believe that good use of data can drive cus-
tomer engagement and loyalty.
Ales Gornjec: We are definitely better informed, but
just collecting data is not enough. Big Data analytics
might not give us direct answers to all operational ques-
tions, but we can complement them with traditional
systems to improve our decision-making process and to
improve player experience. If IT departments can
improve site performance and stability with Big Data
solutions, this will translate into increased player loyal-
ty over time.
In the past, data storage was an issue, but now,
with decreasing storage costs, other issues have
emerged, including how to determine relevance
within large data volumes and how to use analyt-
ics to create value from relevant data. At what
juncture is the gaming industry in relation to these
issues and where should it be heading?
Jeremy Thompson Hill: We have always been facing
these issues, regardless of the operational efficiencies
that have emerged from cheaper and more effective
storage. The key challenge is surfacing actionable
insights, making changes based on these insights and
then automating processes to ensure that data, or the
customer, drives the business. The gaming industry has
always been very data rich, I think we are all heading
towards more efficient use of this data, an increased
level of data throughout the customer lifecycle and
automation to ensure that people are focused on new
ways to improve the customer experience rather than
the manual facilitation of campaigns and operational
procedures.
Ales Gornjec: Collecting data is just the first step. This
one requires some investment into infrastructure, but is
nowadays is quite simple from an implementation per-
spective. Creating value out of collected data is not triv-
ial and requires usage of good analytical and visualiza-
tion tools combined with people who understand the
business. They also must be able to use the data to make
proper conclusions that translate into improvement
actions on the marketing or customer service sides.
Jerry Bowskill: Whilst relevance remains an ongoing

5. 7 0
challenge, the ability of the entire organisation to adopt
a data driven culture remains a challenge in many
organisations.
Dealing with torrents of data in real-time is a chal-
lenge when the desire is to react quickly enough to
deal with data velocity in real-time, and make
reactionary changes to the customer experience.
How do you make this possible?
Ales Gornjec: There are appliances and on-demand
solutions on the market like HP Vertica that are capable
of near real-time analytics. Many Facebook applica-
tions providers already use them, and the data gathering
and analytical part would be quite similar also for gam-
bling operators. Real-time changes to the customer
experience are, however, much more difficult for gam-
bling operators because of the potential risk exposure
and regulatory frameworks that demand additional cer-
tification steps before new products can be deployed.
Real-time gamification features and rewards during
game play are typical tools that operators are now try-
ing to deploy to prolong game play and improve player
experience. They both depend on real-time information,
but don’t require changes in the product. At the moment,
they are triggered by analytics derived from structured
transaction data but could in the future be extended by
additional big-data analytics.
Jerry Bowskill: Whilst optimal real-time interaction is
one potential goal of a Big Data strategy, this needs to
be founded upon a clear understanding of knowing
exactly when, how and more crucially why an interac-
tion is required. A deep understanding of what has hap-
pened historically is important in defining a real-time
strategy. In this respect, predictive models are only as
good as the rules embedded in the underlying algorithm
or machine learning methodology. For many organisa-
tions, there is huge value in understanding player
behaviour and what makes a successful or promotion
strategy without needing to develop a costly real-time
solution.
Jeremy Thompson Hill: This is a key priority for
OpenBet. We recently released a feature which opens
up a data stream of real-time transactional information
for our customers. OpenBet and our customers are in
the process of building some interesting new features
on top of this data feed that we believe will create a
much more engaging customer experience.
The gambling industry at large can learn from the casual
gaming sector here when it comes to creating an engag-
ing real-time customer experience. The use of intrinsic
motivators in games like Clash of the Clans which
encourage users to return consistently is a really inter-
esting example of how real time data can be used. The
gaming industry has spent too many years focusing on
building out the optimum customer lifecyle communica-
tions process that is based around a set of industry stan-
dard offers. This approach has clearly worked but there
are multiple dimensions to customer retention and
building innovative and interesting features that stimu-
late customers to return and bet more frequently is a key
battle ground. For OpenBet, opening up a real-time data
feed to enable innovation is a big step forward. Platforms
01 One of the more profound
developments in the world of big data
is the adoption of so-called data
visualisation. Unlike the specialised
business intelligence technologies and
unwieldy spreadsheets of yesterday,
data visualization tools allow the
average business person to view
information in an intuitive, graphical
way.
TECHNOLOGY - BIG DATA
Interactive

6. A closed-loop management system
A closed-loop management system is a management system that promotes a controlled base of both preferred
outcomes and feedback from the system. Implementing this management system would mean that business activi-
ties such as inventory levels, production schedules, and supply chain functions are guided not just by the sales team
of continued forecasts and orders, but rather includes feedback from ongoing operations from across many business
units. This common style is different from the open loop management systems, which support zero feedback and
have all inputs driven off of set calculations designed around anticipated outcomes only.
Companies who choose to adopt and implement the closed-loop management system facilitate the capacity to avoid
many shortfalls.The closed loop consists of five stages:
● Discovery: This involves identifying internal tools, procedures, and ideas such as the mission and vision state-
ments, SWOT analysis, competition analysis, and core capabilities to articulate a strategy statement.
● Modeling: The platform above is in turn modeled into objectives and initiatives, utilizing additional tools and pro-
cedures which should include flow charts and key performance indicators.
● Deployment: The deployment stage will link the strategy to the physical operation and a third set of tools and
procedures such as quality management, process improvement, engineering, forecasting, planning and budgeting.
● Monitoring: As deployment pushes forward, necessary monitoring continues to evaluate and understand data
and the business environment.
● Optimisation: In the last stage, teams begin to evaluate strategies and optimize a key component which in turn
begins another loop.
7 1
that do not offer this freedom to innovate are restricting
their customers’ ability to challenge and grow.
How do you draw together numeric data in tradi-
tional databases, i.e. information created from
line-of-business applications, unstructured text
documents, email, video, audio, financial transac-
tions, etc.? How do you manage, merge, and gov-
ern these different varieties of data and determine
which are the most important statistics at your
fingertips?
Jerry Bowskill: Organisations who are committed to
developing a data-driven culture often find that a twin-
track approach is most effective. An enterprise-wide BI
platform allows a large number of users to access and
receive relevant and timely information. Equally, a
data-mining and predictive analytics capability can act
as the low-cost and very effective ‘path-finder’ tool to
quickly discover the most relevant data types to deliver
insight to the business. Once identified, these new data
sources can be built into the enterprise-wide BI toolset.
For our B2B Real Money Gaming and Play4Fun social
products, the amount of data we can capture is limited.
All player data is anonymous to meet the needs of the
regulator, so this constrains both the amount of data we
store and has probably helped us focus on unlocking
value by starting with solving simple questions around
game performance across multiple customers and chan-
nels.
Ales Gornjec: Combining all this data into one common
format is a very difficult to even impossible task.
Structured data in existing databases will offer better
reporting and analytical possibilities than unstructured
data usually dumped into big data systems. We recom-
mend analysing each system with the most appropriate
approach and to combine results into an analytical
superset that can then be used to drive the business for-
ward.
Jeremy Thompson Hill: The variety of information
created and available is one of the key challenges in Big
Data analysis and one there is rarely a silver bullet for
tackling. We don't underestimate both the effort and
huge downstream benefits of good data cleansing. With
the plethora of data sources at the disposal of our opera-
tors, quick exploratory data analysis in traditional tools
such as R (and even SQL) can give a good initial indica-
tion of the potential value of a data source and allow
prioritisation of which seams to mine to first.
Data flows can be highly inconsistent with periodic
peaks and daily, seasonal, and event-triggered
overloads – how do you create an infrastructure to
meet these structured and unstructured chal-
lenges?
Jeremy Thompson Hill: The gaming industry has
extreme peaks of both user activity – such as just
before the football on a Saturday; those final ten min-
utes of in-play bets; the Grand National – and offline
transactional activity including settlement of bets on a
major event. There are also frequent permanent shifts in
data volumes from emerging markets and industry
innovation. Event processing infrastructure needs to be
both horizontally scalable to handle the message
through-put and also needs to support a variety of con-
sumers. Some consumers will be tuned and scaled to
handle the peaks in real-time to interact with customers
whilst they're on the site. Whilst other consumers may
be able to lag behind during the peaks and catch up dur-
ing the quieter times. To support the volatility and vari-
ety of use cases we designed our messaging platform on
top of a distributed transaction log. The added benefit of
this approach is that it enables us to replay previous
events, tweak algorithms and see if that change better
identifies patterns in the data.
Jerry Bowskill: Simple – always build a hardware and
software infrastructure that is flexible and scalable. The
virtualization of servers and being able to use a mix of
public and private clouds, based on the types of prod-
ucts and jurisdictions being served, allows solutions to
be scaled cost-effectively. The only certainty in any Big
Data strategy is that your data volume will grow (and
probably at a faster rate than you originally envisaged).
Ales Gornjec: Every system has to be scaled for peaks.
If these are extreme this becomes challenging and there
are several strategies possible to address these issues. In
case of Big Data analytics, that is not a mission critical
activity that has direct impact on revenues, the best
approach would be to filter our data sources based on
their priority in case of potential overloads.
How do you link, match, cleanse, and transform
data across systems, before making the connec-
tions, correlating the relationships, hierarchies,
and multiple data linkages needed to keep the
data under control?
Every system has to be scaled
for peaks. If these are extreme
this becomes challenging and
there are several strategies
possible to address these issues.
TECHNOLOGY - BIG DATA
Interactive

7. 7 2
How will Big Data, which is having such a profound
effect on the scale and productivity of major busi-
nesses, shape and contribute to the future suc-
cess of the gaming sector, both land-based and
online?
Many businesses across different industries are
beginning to understand that data is a product
and, if managed properly, can significantly grow a
business both in terms of efficiencies and rev-
enue.
The gaming sector is no different and generates
huge volumes of data. This is increasing at a rapid
rate due to the increasing number of channels,
products, customers and mobile devices. In recent
years, only high level data was evaluated and
acted on, while there was very little technology
available to store and analyse data in an effective
and timely manner and real-time was a mere pipe
dream. Big Data technologies, however have
changed, and are increasingly changing, the way
many suppliers and operators are able to create
more intelligent processes that can react in real-
time in order to shape and navigate the player
experience in the best way possible.
In preparation for this shift in technology and one
of the most challenging, competitive and com-
pelling periods in our industry Playtech has
developed the latest, cutting-edge Business
Intelligence Technology (BIT). By combining the
power of Playtech’s existing IMS platform and our
fully automated BIT software we are able to lower
player churn, extend lifetime value and overall
revenue as well as personalising and significantly
enhancing a customer’s unique gaming experi-
ence.
BIT offers the following elements:
The BI Platform –
Complete operational overview
Eliminates the need to build an expensive,
resource-consuming system; serves day-to-day
operations and high-level management decisions;
and allows licensees to compare key metrics
against competitors.
Data Driven Marketing Tools –
The power of personalisation
Automated data driven marketing tools that inte-
grate with, and feedback to Playtech’s core plat-
form IMS. These tools enhance player experience,
lifetime value and licensee revenues by offering
users a unique personalised offering including
instant game recommendations, promotions and
bonuses, and are based on statistical player
behaviour analysis and segmentation.
Playtech Analytics –
Real-time decision making
Real-time tracking and reporting to maximise
player value and brand profitability; instantly
identify best-performing tools, products, markets
and promotions; rapidly optimise gaming client
In preparation for this shift in
technology and one of the most
challenging, competitive and
compelling periods in our
industry Playtech has developed
the latest, cutting-edge Business
Intelligence Technology (BIT).
and web pages and stay on top of trends and
changes in player behaviour.
Playtech Optimiser -
Every click has a value
Automated, real-time, personalisation and opti-
misation engine; focus on player engagement and
retention; eliminate guess work and improve
campaign performance – instant marketing suc-
cess; measure player engagement with each page
and in-page component.
Smart Installer –
Superior acquisition and retention tool
Personalised pre-registration marketing; real-
time player re-activation; tailor campaigns to spe-
cific user groups.
What actions should operators take to adopt a Big
Data approach to storing, managing, correlating
and governing their customer data?
Data management can be a highly complex mat-
ter, while building big data platforms that are used
to maximise player experience is even harder.
Today, rapid data flow and the sheer volume of
data make it difficult to organize, map and under-
stand.
Operators considering their big data options
should consult experts from other industries and
follow best practice. Cloud services should also be
considered particularly because of the flexibility
they offer during the early stages of deployment.
But no matter how much data you gather, in order
to make the best use of it you need a team of data
scientists who know to mine, build and extract
information.
TECHNOLOGY - BIG DATA
Interactive
Playtech has announced the hiring of
600 additional staff as part of its
Omni-channel technology rollout.
Playtech’s use of Big Data to
manage the increase in customer
data as part of this solution is central
to its success.
Optimising the
future of Big Data

8. 7 3
Ales Gornjec: The first, and very important, step would
be to implement a true 360 degree view of each player.
This means deploying a single wallet solution or linking
their accounts and all activity data behind them within a
data warehouse or Big Data system. Once this is done,
all players activity and all operators’ communication
with them (email, chat, phone calls) should be stored
there. Initially, only part of the data would be struc-
tured, and some of it only available to No SQL type
search queries.
Jerry Bowskill: A path-finding tool can act as a great
way to undertake this type of data discovery at relative-
ly low cost, compared to building this into an enterprise
data warehouse. These tools also allow for exploration
of different tactics to cleanse, transform and enrich
source data.
Jeremy Thompson Hill: Most of the data sources will
contain a unique identifier or digital fingerprint that can
uniquely identify the resource. We also use systems
that make predictive matching based on scoring a num-
ber of customer credentials.
Who is currently using Big Data in the gaming
industry right now? How are they using it and how
successful have these attempts been to really fully
harvest the benefits of Big Data?
Jerry Bowskill: All major players in the gaming indus-
try have a mature data strategy. Many of our Real
Money Gaming customers across landbased and
Interactive have data warehousing, with internet opera-
tors taking the first steps over 5 years ago. However
there are large variations in levels of expenditure and
the effectiveness of these strategies. What is clear is that
there isn’t always a direct correlation between levels of
expenditure and effectiveness!
Ales Gornjec: We see more and more operators and
vendors using Splunk to help them collect all IT-related
logs in a central BigData repository. Some of them are
publicly listed under
https://www.splunk.com/en_us/customers.html.
Also Tableau, that is in our opinion one of the best data
visualization tools has already a lot of clients in our
industry http://www.tableau.com/learn/stories.
Big Data has been credited in being able to affect
the following: a 360 degree customer view; under-
standing markets; finding new markets; person-
alised website experiences; improved services; co-
creation and innovation; reduced risk/fraud; bet-
ter organised companies; and a better under-
standing of the competition. Do you agree that Big
Data can achieve all these things, and what are the
priorities for gaming businesses?
Jeremy Thompson Hill: Yes, I agree that good use of
data can help to answer these questions and many
more. What data allows you to do is get to know and
understand your customer. The customer expects and
demands a more personalized experience, targeted and
tailored communications. Data allows you to build that
insight and knowledge.
The priority with data will always be the customer,
understanding them and keeping them happy so they
remain a customer.
Ales Gornjec: BigData can help in all above mentioned
areas, but many of them can also be solved with tradi-
tional structured data gathering and data warehouse-
based analytics. A single wallet that processes all player
transactions is a very important element of every mod-
ern gaming platform, and in many aspects behave like a
BigData system.
Jerry Bowskill: These are all areas where a Big Data
strategy has the potential to deliver significant value.
The challenge for any organisation is to have a clear
view on where it believes is the greatest opportunity,
and to focus on effectively delivering the insight to
achieve this. Scientific Games develops industry-lead-
ing game titles and products across a range of markets
and channels. Ensuring that we continue to develop
great games and products which offer the best customer
experience is one area where our investment in Big Data
helps us achieve this.
How do you protect your systems from hacking
and data breaches, and how do you make sure
that the data you gather does not break privacy
laws?
Jerry Bowskill: The infrastructure we use to support
our data-warehousing is completely private and our
operational and security teams, along with trusted
third-party hosting partners, work within the regulatory
and industry best practise security standards across
physical, logical, network and hardware levels.
Similarly, the data we store is managed based on the
jurisdictions and products being served; within Real
Money Gaming, all player data is anonymous and simi-
larly for Social Gaming, the platform holder (Facebook,
Apple, Android etc) have the lionshare of the personal
identifiable information. This greatly limits the sensitiv-
ity of the data we manage.
Jeremy Thompson Hill: There are a number of ele-
ments at play here. Firstly, systems employ the principal
of least privilege. We also employ industry-best prac-
tices around system encryption, hashing and anonymi-
A single wallet that processes all
player transactions is a very
important element of every
modern gaming platform, and in
many aspects behave like a Big
Data system.
TECHNOLOGY - BIG DATA
Interactive

9. 7 4
sation, making sure that these systems are regularly
audited. Furthermore, we work closely with each of our
operators and the governing bodies in the jurisdictions
in which they operate to ensure privacy laws are not
breached.
Ales Gornjec: A player database is one of the most
valuable assets of every gaming operator and has been
always very jealously protected. Traditionally, this
information is stored in their on-premise installations,
and is protected with strict security measures.
With adoption of white label solutions for online gam-
ing, and with more and more service providers (pay-
ments, affiliates, risk and fraud, KYC ) operating from
cloud-based solutions this information has spread over
different systems that are not under the control of oper-
ators of IT departments. These clouds potentially intro-
duce risks in terms of data protection and operators
must evaluate their maturity in terms of security very
thoroughly.
How important is data visualisation in allowing the
average business person to view information in an
intuitive way, as opposed to expecting everyone to
become Big Data scientists?
Ales Gornjec: Visualization is very important, and also
quite challenging because it has to help answering busi-
ness questions and can’t be generated in some standard-
ised way. Nowadays business is changing very quickly,
and this means that also technology will not be the
same for very long time. We are collecting more and
more data, and that constant change and growth para-
digm has to be the foundation of any Big Data solution.
In order to understand business implications and iden-
tify potential improvement actions, the results of
BigData analytics have to be presented in an intuitive
way. For this part big data scientists or data analytics
experts are needed that understand both the business
side and underlying technology.
Jeremy Thompson Hill: This is key - giving teams
across a business access to relevant data in a way that
they can use it is imperative. Visualisation is one piece of
the jigsaw, as is data mining and interrogation .
Empowering individuals to do their job autonomously
and giving them the access they need to relevant data is a
big part of this.
By partnering with a business like Feature Space,
OpenBet is making a clear statement that operators need
people who are trained data scientists to be able to take
significant steps forward around predictive analytics and
the implementation of algorithms that can drive a busi-
ness forward. We wouldn't expect a marketer, for exam-
ple, to be able to interpret data as efficiently as a data
scientist with a PHD. That said, the marketer will need to
able to visualise and act on the insights surfaced for the
process to bear fruit.
Jerry Bowskill: There is a lot of hype about data visuali-
sation at the moment and whilst the presentation of
information is important, the accuracy and specificity of
the insight is more valuable. Data visualisation is just one
element in being able to craft an effective story and set of
recommendations based on information. Just because
your data presentation looks pretty, doesn’t necessarily
mean that it is insightful. Sometimes a few words and a
simple “what’s it worth” figure is clearer and more effec-
tive than a colourful, animated 3D histogram.
01 According to IBM estimates, more
than 2.5 quintillion bytes of data are
generated every day. The sheer
volume, speed and variety of the
information make it daunting. Data at
one time was in well-defined formats
and fit easily into standard databases,
but the advent of social media meant
an explosion of unstructured data.
Some of the data in the Big Data
bucket has been available for some
time, but companies had been unable
to use it due to a combination of
storage, processing power, analytical
and timeliness challenges. But these
challenges have been solved or
minimised.
TECHNOLOGY - BIG DATA
Interactive

10. Does the collection and storage of increasing
volumes of data mean we are ultimately better
informed? And how do you go about turning
customer data into customer loyalty?
For all businesses today, becoming better
informed about your customers is dependent on
the breadth, depth, and sources of customer data
as much as it is on the volume of that data.
Many “big data” systems today rely exclusively
on collecting and analysing transaction-related
data. For example, transactional feeds from gam-
ing systems provide a rich profile of customer
gameplay. Food and beverage systems do the
same for purchases in bars, restaurants and enter-
tainment venues, and loyalty systems provide yet
another rich source of data. But what about data
related to customer activities in between each of
these transactions?
In the physical casino environment, a focus on
transactional data only results in large gaps in
understanding customer behaviour. It’s as if cus-
tomers appear on the radar to transact, and then
disappear. What were they doing between trans-
actions? Were they on the property? If so, where
were they on the property? Did they visit any
facilities and not spend money and if so, why not?
There’s an underlying assumption that the data
already being collected by casinos will be suffi-
cient to provide the basis for influencing cus-
tomers to take a next transactional step. But
instead of having to guess, wouldn’t it be better if
casinos had real data clearly showing the connec-
tions between non-transactional activities and
transactions?
Mobile data – if done correctly – is an informa-
tion gold mine for land-based casinos because it
can provide a much richer profile of customers on
the property on a daily basis. With mobile
engagement solutions as a part of a complete
land-based casino experience, customers appear
“on the radar” much more frequently, providing
opportunities to influence their behaviour that
can lead to a next transaction, and helping casinos
develop a true 360-degree customer perspective.
The Tapcentive mobile engagement and gamifica-
tion platform is one such system that delivers
customer engagement to influence customer
behaviour and generate a rich set of associated
data for those activities between transactions.
This new data creates an entirely new dimension
of mobile data in the “big data” story that comple-
ments and increases the value of existing transac-
tional data. Rather than make assumptions about
what customers are doing when they’re not trans-
acting, casino operators and marketers can dis-
cover new connections and behaviours that link
transactions together and lead to improvements
to existing marketing and loyalty program invest-
ments.
Big Data has been credited in being able to affect
the following: 360 degree customer view;
Understanding markets; finding new markets;
personalised website experiences; improved serv-
ices; co-creation and innovation; reduced
risk/fraud; better organised companies; better
understanding of the competition. Do you agree
that Big Data can achieve all these things and
what are the priorities for gaming businesses?
In addition to the 360-degree customer view that
Dave Wentker is CEO of Tapcentive, a mobile marketing
platform that is transforming how businesses drive foot
traffic and engage their customers by connecting the
digital and physical worlds. Bolstered by more than 25
years of experience leading and developing innovative,
technology solutions for global brands, Wentker
cofounded Tapcentive in 2013 after recognising the need
for more interactive and exciting mobile application tools
for in-location marketers. Prior to Tapcentive, held senior
leadership titles at Visa, where he led an in-house mobile
product development team as the company introduced
and commercialised NFC payments globally.
Customers who pay in cash for
some or all of their transactions
provide a data trail that is helpful
but limited. And those visitors
who visit a property and don’t
spend any money at all are even
harder to track and understand.
7 5
can be achieved by adding mobile engagement
data to traditional known-customer transactional
data, there’s an opportunity to learn more about
casino customers that are partially or completely
“invisible” to the transactional systems.
Customers who pay in cash for some or all of their
transactions provide a data trail that is helpful but
limited. And those visitors who visit a property
and don’t spend any money at all are even harder
to track and understand.
With mobile engagement data as part of the “big
data” stream, casinos have an opportunity to col-
lect data on both of these customer groups. Mobile
engagement has the potential to “un-mask”
anonymous players and visitors by offering them
an additional set of offers and opportunities that
they may be willing to consider during their on-
property customer journey. Even if they do not act
on offers presented to them, the presentation of
those offers is the first step in data collection and
understanding these customer groups.
Not everyone is going to opt into mobile engage-
ment opportunities, but with mobile phones in
every pocket and purse, the opportunity to drive
mobile engagement and collect mobile customer
data from anyone on-property is very real.
MOBILE/ONLINE/LAND-BASED
Interactive
CONNECTING THE DIGITAL AND
PHYSICAL WORLDS WITH DATA

11. 7 6
How do you employ and task individuals to imple-
ment Big Data strategies? Do you outsource such
tasks or build in-house structures? Do gaming
businesses have the right skills in place to develop
or customise big data solutions to fit their needs?
Jerry Bowskill: People are the key to any successful big
data strategy. A significant investment in hardware,
software and data will not deliver a guaranteed return
on investment if you have the wrong people using and
developing it. It’s a bit like an F1 team; you need a great
car, great pit crew and mechanics however, if you
haven’t got the right driver, you aren’t going to win any
races. Recognising and developing what skills you have
in-house and then complementing these with external
support is an effective way to develop your strategy and
retain the technical and commercial knowledge within
the business.
Ales Gornjec: Initially gaming businesses can acquire
these skills externally, but we would recommend that
with time they also hire some internal specialists that
will hold all acquired knowledge together and help indi-
vidual departments understand the results of Big Data
analytics – and to make good decisions based on them.
When selecting new platforms or any other middleware
these should include Big Data functionality, or be at
Caesars (formerly Harrah’s) Entertainment has
long been a leader in the use of analytics,
particularly in the area of customer loyalty,
marketing, and service. Today, Caesars is
augmenting these traditional analytics capabilities
with some big data technologies and skills.
The primary objective of exploring and implementing
big data tools is to respond in real time for customer
marketing and service. For example, the company has
data about its customers from its Total Rewards loyalty
program, web clickstreams, and from real-time play in
slot machines. It has traditionally used all those data
sources to understand customers, but it has been diffi-
cult to integrate and act on them in real time, while the
customer is still playing at a slot machine or in the resort.
In order to pursue this objective, Caesars has acquired
both Hadoop clusters and open-source and commercial
analytics software. It has also added some data scien-
tists to its analytics group. There are other goals for the
big data capabilities as well. Caesars pays fanatical
attention— typically through human observation—to
ensuring that its most loyal customers don’t wait in lines.
With video analytics on big data tools, it may be able to
employ more automated means for spotting service
issues involving less frequent customers. Caesars is also
beginning to analyse mobile data, and is experimenting
with targeted real-time offers to mobile devices.
Caesars Entertainment’s Total Rewards program has
more than 45 million members. These members are
tracked throughout their entire travel journey. From the
moment they book until the moment they leave the
hotel or casino, everything is tracked and analysed and
used to provide supreme services to the members.
Due to the data-drive strategy, Caesars has been able to
trace 58 per cent of all costs spent to the customers in the
company in 2004 to 85 per cent by 2014. A remarkable
achievement and in addition it has given Caesars unique
insights into the behaviour of their customers. As Joshua
Kanter, vice president of Total Rewards for Caesars
Entertainment, Las Vegas, claims: “Big Data is even more
important than a gaming license.”
With all the data, they are able to give loyal visitors very
targeted benefits, while preventing paying too much
money without results. The goal is to determine the right
profile for the guests that arrive at a casino. Cameras
record everyone’s action in the casinos and these cam-
eras even record the choices the gamblers make. A play-
er who is guessing, is more likely to lose money than a
disciplined player. Caesars combines this data with data
that Caesars collects from customers while booking
stays, travel arrangements, dining, gaming and enjoying
other activities at the company’s properties.
All this information is stored, analysed and used for per-
sonal benefits to the guest. Subsequently, a guest
receives a free dinner or hotel room to keep him or her
satisfied. On the other hand, the tracking software is also
used to prevent any of the 75.000 employees being to
generous in giving away freebies.
WHY ANALYSE SO MUCH INFORMATION?
Companies need to look at customers and lifetime value
rather than counting profits simply by product line or
geography, as mentioned by former Gary Loveman, CEO
at Caesars Entertainment and Harvard Business School
professor with a doctorate in economics from MIT.
But there is more. Caesars Entertainment uses the same
type of analytical program to analyze the insurance
claims of all its employees and their family members.
Managers of Caesars are able to track many different
variables about how medical services are used by their
employees. This aggregated and anonymous data can
help the organisation find differences in medical usage.
One property for example, Harrah’s in Philadelphia,
showed a higher usage the emergency room compared to
the overall organisations. Managers brought this to the
attention of their employees and the rate dropped signif-
icantly.
least Big Data-ready for potential future extensions.
Jeremy Thompson Hill: Operators across the gaming
sector approach this in different ways, some outsource,
some have hired big analytical teams but I think there
is an understanding across the board that this is a key
component of success. OpenBet takes the view that we
need to combine best of breed data scientists at a busi-
ness like Feature Space with our own industry-leading
technologists to deliver the optimum service for our
customers. Data science is a very serious business and
we all need to be mindful of our own strengths and
weaknesses. The creation of a strategic partnership
with experts is the best way to crack this problem.
Should the primary outcome of a Big Data analy-
sis be the automation of the decision-making
process, or to make human interaction with the
data more accessible?
Ales Gornjec: We would recommend automation of
the decision-making process to be based on structured
data within their single wallet data store that is still
more accurate and reliable. Big Data is still a new con-
cept for many operators, and is a very valuable addi-
tional source of information. We would recommend
more emphasis on good visualization to make it easily
understandable to humans, that will then prepare cam-
paigns and improvement actions based on this informa-
tion.
Jeremy Thompson Hill: The primary objective of Big
Data analysis for the gaming industry should always be
to understand the customer better. Getting there requires
a combination of automation and human interaction. It is
neither one nor the other, it has to be both. With the
progress that is being made around machine learning,
who knows how far automation may go in the future. In
the words of the pioneering computer scientist Alan
Turing: "Machines take me by surprise with great fre-
quency".
Jerry Bowskill: Both - some decisions can be automated
once algorithms have been developed, refined and
proven to be superior to an alternative approach. At the
same time, human intelligence is just as important and
more effective depending upon the decision being made.
For example, you might use a predictive algorithm to
choose how a web page is presented or what promotions
and bonuses are marketed to an individual player. For a
large strategic investment decision, you will almost cer-
tainly use data as a critical input factor. However, it will
also require a significant amount of human input to
make the final decision.
TECHNOLOGY - BIG DATA
Interactive
CAESARS: BIG DATA IN PRACTICE