View
303
Download
3
Embed Size (px)
DESCRIPTION
Big data is still relatively new and it is very exciting. The opportunities, if not necessarily endless, are are at least incredibly rich and varied. Aiming to bridge the link between Big Data as a Technology and Big Data as Business Value, we hope our presentation will help frame some of your thinking on how to use and benefit from this topical development.
Citation preview
TRANSLATING TECHNOLOGY INTO BUSINESS
Let’s make money from Big Data!
JUNE, 2014
About “Transla.ng Technology into Business”
June 2014 2
• B Spot helps clients transform technology ideas into business concepts.
• As part of our on going effort to add value, we publish monthly content related to this topic on our website. “Transla?ng Technology into Business” is aimed at organisa?ons and individuals who want to understand some of the changes and impact that technology developments have on industries and business.
• Our short presenta?ons are not deep technical documents; rather, they are business-‐orientated, analy?cal opinion pieces and perspec?ves about the dynamics surrounding technology developments and, most importantly, the opportuni?es that these create.
• B Spot’s presenta?ons are free to download. • If you have any further ques?ons, sugges?ons for new topics, or comments please
contact beatrice@bspotconsul?ng.com
Enjoy! B Spot
Content
June 2014 3
§ Explaining Big Data § Evolu?on § Market segmenta?on § Market size and forecast § Demand analysis § Spot on … what you need to take away
Big data technologies are just tools; the real value comes from what we make out of it Explaining Big Data
June 2014 4
Big Data is data that is too large, complex and dynamic for any conven.onal data tools to capture, store, managed and analyse.
The right use of Big Data allows analysts to spot trends and gives niche insights that help create value and innova.ons much faster than conven.onal methods. Source: Vipro
Volume
Velocity
Variety
Amount of data stored worldwide (in petabytes)
> 3,500 North America
> 50 La?n America
> 2,000 Europe > 250 China
> 50 India > 200 Middle East > 400 Japan
• People to people: Social networks, web logs, virtual communi<es, etc.
• People to machines: medical devices, archives, digital TV, e-‐commerce, smart cards, bank cards, computers, mobiles, etc.
• Machines to machines: Sensors, GPS devices, bar code scanners, surveillance cameras, scien<fic research, etc.
The speed at which new data is being created – and the need for real-‐<me analy<cs to create business value from it -‐-‐ is increasing thanks to digi<sa<on of transac<ons, mobile compu<ng and the sheer number of internet and mobile device users.
Big data is far from new but has only in recent .mes been recognized as an industry Evolu.on
June 2014 5
Source: B
spot ana
lysis
1989 2005 2011 2012 2013
Tim Berners-‐Lee invents the Web and mass digital
data collec.on starts
Steve Jobs became one of the first
people in the world to have his en.re DNA sequenced as well as that of his tumor – first person to use Big Data to try to safe his life.
The open source Big Data framework called Hadoop has been all about innova.ve ways to process, store,
and eventually analyze huge volumes of mul.-‐structured data. From the .me of its incep.on by
Doug CuUng at Yahoo un.l 2011 or so, the majority of enhancements to the plaZorm have been mostly focused on new and be[er ways to accomplish this
core func.on.
The amount of data created both inside
corpora.ons and outside the firewall via the web,
mobile devices, IT infrastructure, and other sources is increasing
exponen.ally each year. From 2005 to 2020, the digital universe will grow by a factor of 300, from 130 exabytes to 40,000 exabytes, or 40 trillion.
Google self-‐drive car based on big data intelligence is being
developed
Further development of visual techniques and technologies used for crea.ng images, diagrams, or anima.ons to communicate, understand, and improve the
results of big data analyses, e.g. tag cloud, clustergram, history
flow, spa.al informa.on flow, etc.
Major IT vendors aggressively entered the big data space despite making li[le revenue from it but recognizing future poten.al and
massive impact on their hardware, sobware and other services impact.
Following an example from retail and stock
exchange markets other industries have started using big data tools for
their internal and external purposes. Mainly for
customer segmenta.on and product development.
2014
Ed Snowden exposes mass surveillance and big data abuse by the US and the
UK authori.es. The issue of privacy and correct usage of big data became an
urgent issue.
Major infrastructure in big data investments
taking place.
The market is s.ll generally very fragmented Market segmenta.on
6
• Storage • Servers • Networking
Vendors include Dell, HP, IBM, Cisco
Hardware Big Data Distribu.ons
Data Management Components
Analy.cs and Visualisa.on Services
• Community Hadoop distribu<ons
• Enterprise Hadoop distribu<ons
• Non-‐Hadoop Big Data framework
Vendors include Cloudera, IBM, MapR, LexisNexis, MicrosoW
• NoSQL databases • Data integra<on • Data quality and
governance
Vendors include Data Stax, IBM, Informa<ca, Syncsort
• Analy<c development pla[orms
• Advanced analy<cs applica<ons
• Data visualisa<on tools
• Business intelligence applica<ons
Vendors include Karmasphere, Tresata, Datameer, SAS Ins<tute, Tableau, Revolu<on Analy<cs
• Consul<ng • Training • SoWware
maintenance • Hardware
maintenance • Hos<ng/cloud
Vendors include Think Big Analy<cs, Amazon Web Services, Accenture, as well as services associated with enterprise distribu<ons (e.g. Cloudera).
Next Genera.on Data Warehouse • MPP, columnar data warehouse appliances
• In-‐memory analy<cs engines
Vendors include EMC Greenplum, HP Ver<ca, Teradata Aster Data, IBM Netezza, SAP, MicrosoW, Kognito
Source: Wikiban
June 2014
Almost 40% of the market is held by 8 companies and they supply mainly hardware Market segmenta.on
7
Big Data revenue split by type compiled by Wikibon.org, 2012
Source: Wikibon, companies data
0
500
1,000
1,500
2,000
2,500
IBM
HP
Teradata
Dell
Oracle
SAP
EMC
Cisco
MicrosoW
Accenture
Fusio
n-‐io
PwC
SAS Ins<tute
Splunk
Palan<
r De
loiee
Amazon
NetAp
p Hitachi
Ope
ra Solu<
ons
Mu Sigm
a TC
S Intel
MarkLogic
Booz Allen Ha
milton
Clou
dera
Ac<a
n SG
I Capgem
ini
1010data
Orginal Device Manufacturers
Others
June 2014
Top 8 players holding 40% market share
but big data revenues are s<ll 1% or less of their overall
annual revenues
• Leading IBM offers the largest product and services por[olio and is one of the biggest promoters of Big Data. • Second revenue generator in 2012, HP, made money from from Big Data-‐related services, followed by sales of hardware to support Big Data deployments. HP by its sheer size is in a posi<on to impact and par<cipate in a number of Big Data deployments.
• Others, combina<on of hundreds of exis<ng and start-‐ups, will be the most dynamic contributors group to the big data companies. • The mix of big data technology developers and big data service providers will be changing. Any company involved in data gathering, and using latest analy<cal tools can call themselves big data company. That will have an impact on exis<ng industry of market research which will be under pressure to either transform or join big data market.
8
There are opportuni.es for different type of players, new and exis.ng, to make inroads into big data Market segmenta.on
Big data produc?on Big data
management Big data
consump?on
Source. CM Research
• Social media • Documents • Databases • Web crawlers • Web robots • Sensors • Voice • Music & video • Email • RFID • Call records • Payment details • GPS
Volume Velocity
Variety
Storage
Big Data quality
Security
Analy.cs
Databases
Data mining
Search
Digital marke.ng
Re-‐selling
June 2014
Big data is the fastest growing market since the discovery of the Internet Market size and forecast
9
0
10
20
30
40
50
60
2011 2012 2013 2014 2015 2016 2017
Source: Wikiban, IDC, IBM; Bspot analysis
Market revenues and forecast for Big Data, 2011-‐2017
USD Billion
7.2 11.4
18.2
28.0
37.9 43.7
47.8
31% growth CAGR
61% annual growth
June 2014
An es<mated total value of big data including revenues coming from the sale of hardware,
soWware and services but also revenues coming from the value big data tools have been
genera<ng.
An es<mated l value of big data including revenues coming from the sale of hardware,
soWware and services.
Growth driven by increasingly more adopters beyond Web star<ng using big data tools not only retailers but also pharma, energy, financial
services.
More investment being poured into big data technology especially by larger companies like Google, Facebook and Amazon driving the
prices dawn and allowing the access to big data tools to wider customer base. The technology of big data is maturing,
especially soWware like Hadoop, NoSQL data stores, in-‐memory analy<c engines and
analy<c databases.
Key growth factors include: matura.on of sobware, growing awareness of benefits, growth in investment Market size and forecast
10 June 2014
• Increased awareness of the benefits of Big Data as
applied to industries beyond the Web, esp. financial services, pharmaceu<cals, and retail.
• Matura<on of Big Data soWware such as Hadoop, NoSQL data stores, in-‐memory analy<c engines, and massively parallel processing analy<c databases
• Industries will start using big data analy<cs more frequently and they will increase the level of decision-‐making process on it following beeer understanding of the services provided by big data vendors.
• Following first wave of big infrastructure investments coming from big companies and organisa<ons there should be a second wave of investment boost coming from non-‐IT companies.
• Smart devices including computers, smart phones
but also smart devices used by industries e.g. smart meters, sensors, etc. will drive faster adop<on of big data usage.
It will help to grow: It will con?nue to be a challenge:
• Data is moving from structured to unstructured format, raising the costs of analysis. This creates a highly lucra<ve market for analy<cal search engines that can interpret this unstructured data.
• Proprietary database standards are giving way to new, open source big data technology pla[orms such as Hadoop. This means that barriers to entry may remain low for some <me.
• Many corpora<ons are op<ng to use cloud services to access big data analy<cal tools instead of building expensive data warehouses themselves. This implies that most of the money in big data will be made from selling hybrid cloud-‐based services rather than selling big databases.
• In future, a growing propor<on of big data will be generated from machine to machine (M2M) using sensors. M2M data, much of which is business-‐cri<cal and <me-‐sensi<ve, could give telecom operators a way to profit from the big data boom.
• Legisla<on issues including privacy concerns, data security and intellectual property rights are s<ll unresolved and it will need to be regulated and cross-‐regional and global standards will have to be introduced.
Source: W
ikiban
, IDC
, IBM
; Bspo
t ana
lysis
Currently hardware suppliers are the biggest revenue generators, but sobware and services are the future winners Market size and forecast
11
34%
22% 16%
8%
8%
5%
3% 2% 2%
Professional services
Compute
Storage
SQL
Applica<ons
XaaS
Networking
NoSQL
Infrastructure soWware
39%
41%
20% Services
Hardware
SoWware Big Data sobware and services revenue split, 2013
Big Data revenue split by type, 2013
Source: Wikiban, IDC, IBM; 2013
June 2014
Hardware sales will con<nue enjoying good market condi<ons in the short to medium term. Once large players will sa<sfied their needs for inves<ng in big data infrastructure, there will be smaller players and companies from other non-‐IT industries needing hardware for building big data internal capabili<es.
At the same <me soWware and services providers will con<nue
to grow and in the long term they will increase in its significance over hardware which will eventually commodi<zed.
According to Wikibon analysis, vendors will con<nue using NoSQL and in-‐memory database soWware, streaming analy<c pla[orms, ver<cally focused analy<cal and transac<onal
applica<ons and applica<on development pla[orms (both on-‐premise and Cloud-‐based) and associated consul<ng and
professional services to address specific, high-‐value business problems and opportuni<es.
Industries focusing on consumer needs like retail, banking, telecoms are the first to use big data tools Demand analysis
12
1
10
5
2018 2012 2015 year
Electronics and computers
Telecommunica.on
Healthcare U.li.es
Media
On-‐line services
Retail
Public services
Professional services
Financial services
Defense and Police
Manufacturing
Transporta.on Automo.ve
Educa.on
Travel
First adopters
Laggards
Source: Bspot analysis
Natural resources
Construc.on Sport
Airline
June 2014
Level of adop.on
In the future, it will be industries driving the big data development, not IT companies (1/3) Demand analysis
13
Financial services
Healthcare
Retail
June 2014
• About 70% of the industry is already using big data and analy<cs. For example big data has been used for a long <me in the trading industry. In fact, using mathema<cal algorithms for lots of data analy<cs is traders specialism but also great trading secret. • Banks and financial services firms are also turning to big data, using insights pulled out of daily transac<ons, market feeds, customer service records, loca<on data, and click streams to carve out new business models and services and transform how they go to market. They also using big data to focus on opera<onal issues – risk, efficiency, compliance, security and making beeer decisions. Some of the ideas financial services firms can use big data for: personalised services, loan decisions support, improve customer loyalty, op<mize return on equity, combat fraud and mi<gate opera<onal risk, iden<fy new revenue streams. • Walmart pioneered the use of big data to improve opera<onal efficiency in the retail industry well before the term big data even existed. The company streamlined its complex supply chain to take advantage of economies of scale, thus limi<ng excess inventory and reducing associated costs. Than, the retailer passed on some of these big data-‐enabled savings to customers in the form of low prices undercut the retailer's compe<<on. • Retailers, service companies and consumer goods producers are the most hungry of big data intelligence on their customers. Big data analysis are used for customers’ segmenta<on, marke<ng to enhance customers reten<on and understanding demand for new products and services. Dynamic price op<miza<on, video-‐enabled store layout and product placement analysis, staffing analysis and decision support, suppliers analysis and op<miza<on of supply <ming, pricing and sourcing, knowledge of customers' buying paeerns and behavior are addi<onal ways how retails can capitalise on big data input.
• The pharmaceu<cal industry began mining and aggrega<ng sales and prescrip<on data because this lever helped companies improve their boeom line by more effec<vely targe<ng sales, managing sales force resources, and selec<ng prime areas for R&D. A number of pharma companies are already using big data, among them, Bristol Myers Squibb. BMS has spent nearly $46 billion on research and development since 1997, indexes hundreds-‐of-‐thousands of clinical documents per year in pursuit of insights that will improve the drug discovery process. BMS is using soWware from HP to analyze research and market data to be used by clinical researchers and scien<sts. • For medical devices manufacturers big data pla[orms can become substan<ally more intelligent by including modules that use image analysis and recogni<on in databases of medical images (X-‐ray, CT, MRI) for pre-‐diagnosis or that automa<cally mine medical literature to create a medical exper<se database capable of sugges<ng treatment op<ons to physicians based on pa<ents’ medical records. In addi<on, clinical decision support systems can enable a larger por<on of work to flow to nurse prac<<oners and physician assistants by automa<ng and facilita<ng the physician advisory role and thereby improving the efficiency of pa<ent care. • Public health can benefit enormously from big data. Wider variety of health care informa<on, making them more informed consumers of the medical system. Pa<ents could be able to compare not only the prices of drugs, treatments, and physicians but also their rela<ve effec<veness, enabling them to choose more effec<ve, beeer-‐targeted medicines, many customized to their personal gene<c and molecular makeup. Pa<ents could also have access to a wider range of informa<on on epidemics and other public health informa<on crucial to their well-‐being.
In the future, it will be industries driving the big data development, not IT companies (2/3) Demand analysis
14
Public sector
U?li?es
Educa?on
Telecos
June 2014
• Intelligent use of smart meter data will allow u<li<es companies to: beeer monitor and forecast energy consump<on paeerns; iden<fy inefficient energy use at both the macro and household levels; accurately predict poten<al power outages and equipment failures before they occur; improve customer segmenta<on and tailor service offerings based on customer behavior. • Smart grids will be the next step of managing energy informa<on but start grids are s<ll not common yet, IT companies need to get started to collaborate with u<li<es now. The level of sophis<ca<on in managing and analysing data from smart grids is even higher. Apart from smart meters data there will also will be grids data, energy distribu<on data, IT databases data and others. • Addi<onally, u<li<es are already able to use data about their customers to offer beeer or new services, reduce customers’ churn, brand monitoring and even support machine performance monitoring and supervision. • EDF Energy, using SAS big data pla[orm, has created a dedicated analy<cs func<on to focus on key areas including customer segmenta<on, churn assessment, probability modeling and product placement modeling.
• Governments have lots of data available and its wise usage can be beneficial for the administra<on as well as ci<zens. Big data used by governments will enable people to make beeer choices about the public services they use and to hold government to account on spending and outcomes. • Big Data is also providing the raw material for innova<ve new business ventures and for public service professionals. • According to the UK free market think thank Policy Exchange, the UK government could save up to £33 billion a year by using public big data more effec<vely. McKinsey has inves<gated that the poten<al annual value to Europe’s public sector thanks to big data is 250 billion Euro.
• Educa<on has always had the capacity to produce a tremendous amount of data, more than maybe any other industry. The benefits range from more effec<ve self-‐paced learning to tools that enable instructors to pinpoint interven<ons, create produc<ve peer groups, and free up class <me for crea<vity and problem solving. Big data could enable customized modules, assignments, feedback and learning trees in the curriculum that will promote beeer and richer learning, customise courses and even big data can be used in admissions, budge<ng and student services to ensure transparency, beeer distribu<on of resources and iden<fica<on of at-‐risk students.
• Telcos already have the customer profile data with demographics informa<on (age, income, gender, profession, etc.), subscriber usage and loca<on. The simple thing is to put together the knowledge of the customer and proac<ve customer service: offer with renewing contract ahead of expira<on, roaming discounts ahead of foreign travel, etc. Basically, the amount of data hold by telcos on their customers is a marke<ng goldmine and apart from helping to increase revenues it will also support to reduce subscribers’ churn, control cost of acquisi<on simula<on tools, reduce opera<ng costs, help with fraud detec<on, help products improvements and tailor upon customers’ needs in real <me, etc.
In the future, it will be industries driving the big data development, not IT companies (3/3) Demand analysis
15
Man
ufacturin
g Av
ia?o
n Au
tomo?
ve
Profession
al
services
June 2014
• Thanks to advanced analy<cs of all customer transac<onal data and external data sources (e.g. social media), automakers will be able to make improvements in customer acquisi<on, customer reten<on and manage beeer return on marke<ng investment. Addi<onally, the automo<ve sector is able to use big data for op<mizing supply chains, predict/an<cipate maintenance; connec<ng data from the vehicles, or the devices they integrate with, to relay informa<on from vehicle to vehicle (V2V), and vehicle to infrastructure (V2I) too; GPS and Satellite Naviga<on systems performing in real <me, etc. • Big data offers significant inroads for making cars safer – mostly through its ability to automate func<onality. On board vehicle systems can now inform each other of their whereabouts and of other hazards in the road so that drivers can avoid collisions. • Google's self-‐drive car is an example of using big data in automo<ve to use external and internal data for this inven<on.
• By analysing data created by jet engines and sensors that collect data on the surrounding environment (temperature, humidity, air pressure, etc.), service providers are able to predict when various parts are likely to fail and take preventa<ve maintenance ac<on. Replacing a soon-‐to-‐fail part before it malfunc<ons is significantly less costly than doing so aWer the part fails during opera<ons. More efficient jet engines consume less fuel and emit fewer environmentally contamina<ng gasses. • Other advantages of using big data tool by avia<on are: preventa<ve maintenance reduces aircraW “down <me” , improved customer sa<sfac<on, <cket pricing predic<ons and others. • New revenue genera<on tools. Bri<sh Airways for its new personalized service and offers program, Know Me. It collects and tracks an usual amount of data on individual passengers, their preferences and travel history. Data on the online behavior and buying habits of 20 million Bri<sh Airways customers, crea<ng hundreds of predic<ve signals that suggest a person’s “behavioral DNA to offer new services.
• Big data can help manufacturers reduce product development <me by 20 to 50 percent and eliminate defects prior to produc<on through simula<on and tes<ng. That a massive saving for the R&D process. • Manufacturers could capture a significant big data opportunity to create more value by ins<tu<ng product lifecycle management. Designers and manufacturing engineers can share data and quickly and cheaply create simula<ons to test different designs. Big data can help with further improvements in product quality, use real-‐<me data from sensors to track parts, monitor machinery, and guide actual opera<ons. • Taking inputs from product development and historical produc<on data (e.g., order data, machine performance), manufacturers can apply advanced computa<onal methods to create a digital model of the en<re manufacturing process.
• First adopters are management consultancy and market research companies to replace manual data mining to speed up analyst work in order to focus more on analy<cs and value to the clients rather than data provider. • Legal firms and accountancy companies are known to be tradi<onal and slow with implemen<ng technologies. On the other hand they collect and store massive amount of data and their services are also based on finding the right data and correctly apply. Introducing big data tools will help them with overall performance, speed and accuracy.
Spot on… what you need to take away
16 June 2014
For vendors: § To mone<se your innova<ons and solu<ons, transform your big data concepts into value proposi<ons
that are based on ac<onable insights that drive revenue and/ or reduce costs for your customers.
§ Integrate big data from structured, mul<-‐structured and unstructured data from various (internal and external) source system together in a common pla[orm.
§ Put safeguards in place to address public concerns about big data, including, but not limited to, privacy, security, intellectual property, and liability.
For companies: § Manage big data as a corporate asset and educate employees on how to iden<fy business requirements
for big data projects and effec<vely communicate insights extracted from big data to the business.
§ Trust big data input and make analy<cs-‐driven decision rather than follow “gut ins<nct”.
§ Protect compe<<vely sensi<ve data or other data that should be kept private or corporate secret.
www.bspotconsulting.com