By embracing big data and predictive analytics to create multidimensional customer profiles, companies can make more informed business decisions that better anticipate customer needs, wants and desires.
Cybersecurity Awareness Training Presentation v2024.03
Turning Customer Knowledge into Business Growth
1. Turning Customer Knowledge
into Business Growth
By embracing big data and predictive analytics to create
multidimensional customer profiles, companies can make
more informed business decisions that better anticipate
consumer needs, wants and desires.
| KEEP CHALLENGING
2. Executive Summary
Customers today can access an unprecedented volume of
information via varied channels before making an informed
purchase. For organizations, this means continuously learning
from customer behavior to stay relevant. But while there is no
dearth of customer data available, organizations often grapple
with the challenge of developing clear, complete and fully
updated profiles of their customers.
In a 2012 study, conducted by Columbia Business School and
New York American Marketing Association,1 39% of corporate
marketers said their company’s customer data was collected
too infrequently and was not up to date. Meanwhile, a January
2013 study by Aberdeen Group2 found that top-performing
companies are more likely than others to use a rich set of data
sources to feed their predictive analytics models, including
internal transaction data and unstructured or real-time data, to
provide actionable guidance for decision-makers (see Figure 1).
Creating Rich Customer Profiles
Data Source
Leaders
Followers
Internal transactional records
93%
74%
Internal customer records
75%
80%
Customer sentiment data
57%
29%
External customer information
56%
36%
Customer interaction data
56%
36%
Clickstream data
40%
18%
Unstructured data
38%
29%
Base: 157
Source: Aberdeen Group report, January 2013
Figure 1
2
KEEP CHALLENGING
December 2013
3. The use of big data and analytics can be extended to customer
relationship management (CRM), as companies need to
combine structured and unstructured data with powerful
analytics tools to create a multidimensional customer profile.
This white paper describes a solution concept and
implementation approach to developing a multidimensional
customer profile and deriving actionable insights with the help
of big data and analytics.
TURNING CUSTOMER KNOWLEDGE INTO BUSINESS GROWTH
3
4. The Data Challenge
Organizations have traditionally used structured customer data stored in their
enterprise systems to develop customer profiles. Additionally, a few have attempted
to incorporate external data purchased from third-party agencies, by converting it
into structured formats that can then be stored in their enterprise systems.
However, this approach results in customer profiles that, at best, are incomplete in
the following ways:
• Data stored in enterprise systems is dated and restricted to past interactions.
Many times, data integrity is questionable; for instance, a promotional mailer
may use a customer address from the CRM system, but if the customer has
relocated, the promotion campaign is rendered ineffective.
• Agency data is based on extrapolated customer surveys, which can never
replace actual data insights on individual customers.
• Customers no longer use only company-operated channels. Consumers have
a much broader footprint through social media to broadcast their experience
with the company’s products or services or even their intent to switch to a competitor’s offerings.
For any sales and marketing team, it is vital to
keep current with the pulse of the customer,
and this cannot be accomplished by relying
solely on internal enterprise data.
Because of these factors — and with the fast uptake of social, mobile, analytics and
cloud technologies (the SMAC Stack™), creating customer profiles without semi- or
unstructured data can render an organization uncompetitive and even irrelevant.
The Customer’s Multiple Dimensions
For any sales and marketing team, it is vital to keep current with the pulse of the
customer, and this cannot be accomplished by relying solely on internal enterprise
data. Information avenues that can provide crucial insights include social media
activity, browsing behavior, mobile app downloads, games played, past purchases,
photos shared, music/video preferences and vacation choices. We call the accumulation of all these activities a Code Halo™, which is essentially the digital footprint
created by enterprises, customers, employees and processes from their online
behavior. Business leaders such as Amazon and Google have quickly risen to the
top of their industries by deriving meaning from the intersections of Code Halos
and building their strategies around these insights. (For more on this topic, read
our white paper, “Code Rules: A Playbook for Managing at the Crossroads.”)
A true view of the customer, then, needs to link the details stored in enterprise
systems with Code Halos, or external customer information. This consolidated or
augmented view presents a near-real-time and complete picture of the customer
(or potential customer) with which the business is interacting. Because the traditional view completely ignores the social aspects of the customer, it can best be
described as a dormant description that is waiting to be brought to life by social
information and the customer’s Code Halo.
4
KEEP CHALLENGING December 2013
5. However, this does not happen automatically; semi- and unstructured data that
supplies information on customer activity in the external world needs to be
analyzed and indexed before it can be melded with structured data from enterprise
systems and delivered in the form of a multi-dimensional customer profile. We call
this process AIM (or analyze, index and meld) & Deliver.
The multidimensional customer profile is like a coin with two sides; the face of the
coin depicts the structured data elements of the customer, and the back depicts the
unstructured data elements. When both of these aspects are melded and delivered
together, the true customer profile can be derived.
The multidimensional customer profile can also be visually represented by a sphere
(see Figure 2). Note that when you slice this sphere, you can look at various aspects
of the customer and company from both structured and unstructured perspectives.
Once the multidimensional customer profile is available, it opens up multiple use
cases that drive real-time actionable insights. The insights can be made available
Creating the Multidimensional Customer Profile
Creating the Multidimensional Customer Profile
Unstructured
Events
Mobile
Life events
Apps Games
Photo
ent
Environm
er
y Weath
Econom
Audio
Searches
Video
Downloads
Comments
Docs
Sharing
Favorites
Company Web site
Web/mobile clickstream
Store
Browsing behavior
Product pages visited
Footfalls
interest
Product
Location
Intelligence
Frequently used Web site
Social
Videos
Podcast
ntiment
se
/brand
Product
nce
l Influe
fessiona
Pro
Social
s
fluencer
in
Product
PAS
Current residence
Frequent visits
Device preferences
Competitor purchase interest
Interaction history
Travel/vacation
Points of interest
Blogs
Boards
Job profile
networks
Social networks Professional networks
Forums
Skill set
Likes
Dislikes
Bookmarks
Product failures
Product comments
Sharing
Circle
Demographics
3
Age
Allied product interest
Payment history Credit history
history
Professional
Customer influencers
Product/brand interest
groups/
Product
y
hierarch
Contact center
Professional Influence
Social
Micro-blogs
Customer surveys
y
Compan
program
Loyalty
Benefits
Tiers
tabase
Offers da
tions
tore loca
S
Gender
Other
Customer
Offer responses
Credit-worthiness
Past offers
Credit terms
Channel
Chat
Direct mail
Grievances
groups/
Product
affinity
E-mail
Search keywords
Accept/ignore
Service history
Cases
Product interest
Purchase history
Preferred mode
Contact preference
orks
er netw
Partn
y
Inventor y
lit
availabi
Data Elements
Attributes
Structured
Figure 2
TURNING CUSTOMER KNOWLEDGE INTO BUSINESS GROWTH
5
6. The CRM Analytics Continuum
Creating the Multidimensional Customer Profile
Customer trigger
Analytics insights
lead to decision maps
for executives.
Analytics-based insights
lead to decision matrix
for field service reps.
Connect with real-time
customer profile.*
Unstructured
Product Affinity
Mobile
Apps Games
Photo
Economy
Audio
Searches
nt
Environme
Weather
Video
Downloads
Comments
Product pages visited
E-mail
Frequently used Web site
Location
Intelligence
PAS
Customer Customer
Product
Cross-Sell
ID
Profile
Purchased
Offer
Web/mobile clickstream
Browsing behavior
Points of interest
Boards
Social networks Professional networks
Likes
Dislikes
Bookmarks
Sharing
ase
Loyalty
Tiers
ions
Store locat
Partner
Age
Allied product interest
program
Benefits
Gender
Offer responses
Credit-worthiness
Past offers
Credit terms
Offers datab
B2
Brand B
Brand C
Y
123
Demographics
Customer surveys
3
Y
Product comments
Circle
Payment history Credit history
Company
Brand C
Skill set
Micro-blogs
history
Brand A
XYZ
Job profile
networks
Product/brand interest
Channel
A1
Professional
Customer influencers
Forums
ABC
Interaction history
Professional Influence
Social
Product failures
groups/
Product
hierarchy
Brand C
CrossSell
Success
Direct mail
Contact center
Travel/vacation
Blogs
Chat
Search keywords
Device preferences
Competitor purchase interest
Current residence
Frequent visits
Grievances
groups/
Product
affinity
Docs
Sharing
Favorites
Company Web site
Store
Footfalls
interest
Product
Social
Videos
ment
Podcast
rand senti
Product/b
nce
al Influe
Profession
Social
influencers
Product
High
A1
Customer Profile Acceptance
Events
Life events
A1
Brand C
Brand A
Y
DEF
C3
Brand D
Brand B
Y
Other
Customer
Accept/ignore
Service history
Cases
Product interest
Brand A
Brand B
High
Low
C3
Brand E
Brand D
Brand B
Purchase history
Low
Preferred mode
Contact preference
networks
Inventory
availability
Structured
*Powered by the AIM &
Deliver process.
Figure 3
5
and customized for different stakeholders in the form of decision matrices/maps
that can be leveraged for real-time data-driven decision-making. The effectiveness of decisions using this approach drives continuous closed-loop feedback (see
Figure 3).
An example of this is real-time cross-sell offers, in which the decision matrices/
maps can vary for different stakeholders (see Figures 4 and 5). Using the multidimensional customer profile derived from big data and analytics, the contact
center agents, sales representatives and any other customer-facing personnel have
access to the exact real-time offers they need to entice customers or prospects.
This kind of decision-making is more operational in nature and targeted to the
timing of the customer trigger.
2
At the same time, the multidimensional customer profile can deliver the muchneeded fuel to power analytics for executive decisions. In order to understand which
offers performed well and the changes needed to improve the offer management
process, executives would need a dashboard providing planning insights such as
purchases made to date, potential pairing across products and categories, and
customer profile acceptance levels to boost success rates.
Cross-Sell Decision Matrix for the Customer Operations Team
Customer
Profile
Product
Purchased
Cross-Sell
Offer
Cross-Sell
Success
ABC
A1
Brand A
Brand C
Y
XYZ
B2
Brand B
Brand C
Y
123
A1
Brand C
Brand A
Y
DEF
C3
Brand D
Brand B
Y
Customer ID
Figure 4
6
KEEP CHALLENGING December 2013
7. Cross-Sell Decision Map for Executives
Product Affinity
High
A1
Customer Profile Acceptance
Brand C
Brand A
Brand B
High
Low
C3
Brand E
Brand D
Brand B
Low
Figure 5
Implementation Approach
To implement the solution, we recommend a four-phased approach (see Figure 6).
Phase 1: AIM & Deliver
To initiate the first phase of the AIM & Deliver process, the underlying data elements
must be identified. This entails merging the customer details available within and
outside the enterprise (see Figure 7, next page).
Disparities across the data sources need to be ironed out to associate customer
data within the enterprise with the right data sources in the external world. This can
be done with advanced analytics. By combining automatic entity extraction with
name matching, users can automatically identify entity mentions in unstructured
data and link them with structured information. This linkage simplifies the process
and combines data about an entity into a complete customer profile.
AIM &
Deliver
Phase 3
Phase 2
Phase 1
Four-Step Implementation Process
Phase 4
5
Analyze
Index
Meld
Deliver
• Entity extraction
• Document
clustering
• Attribute matching
• Customer name
matching
• Link structured data
• Link customer to
enterprise applications
• Real-time customer profile
• Location intelligence
Evaluate
business case and
stakeholders
Define stakeholders and detail the use case.
Assess business relevance, technology and economic hurdles.
Big data
architecture
Design the big data architecture after use case crystallization.
Analytics
engine
Configure analytics engine for actionable insights.
Derive real-time
multidimensional
customer profile.
Deliver augmented
real-time multidimensional
customer profile
Figure 6
TURNING CUSTOMER KNOWLEDGE INTO BUSINESS GROWTH
7
8. Merging Two Worlds of Data
Loyalty
Program
Data
Contact
Preferences
Orders/
Pipeline
/ Analytics
BI
Structured
Data
s
Contracts
En
t
Channel
History
Credits/
Terms
Structured
Surveys
Social
Events
Scan Curation
Documents
Contact
History
eS
pris ystem
er
Payment
History
Campaign
Data
Quote
Purchase
History
Service
History
Agency Data
Quantified
Self
Transactions
Influence
Professional
Network
Activities
E-commerce
Social Bookmarking
Index
Voice Portal/IVR
Agency/
Semi-/Unstructured
Data
Analyze
Photo Sharing
M-commerce
Location Intelligence
Contact
Center Data
Web Clickstream
Direct Mail
Video Sharing
Social Activity
Boards/
Forums/
Activities Boards/
Forums/
Activities Online
Searches
Mobile
Activity
Blogs/
Microblogs
Unstructured
Surveys
POS
Store Transactions
E-mails/
Surveillance
Chat
Figure 7
The key steps involved with combining these two different genres include:
• Analyze
the different types of data, clustering them based on specified
parameters and extracting entities, such as customer name, organization,
product name, location, etc.
• Index the clustered data sets and create structured metadata for each entity,
enabling fast filtering and searching by people, places, company names or other
entities.
8
KEEP CHALLENGING
December 2013
9. • Meld the extracted entities with near-perfect attribute matches (i.e., accurate
customer names with existing customer data in the CRM system).
• Deliver the augmented customer profile, enhanced with location intelligence for
easy consumption by CRM systems, BI/analytics or any other point solutions.
Phase 2: Evaluate the initiative’s business
case and stakeholders.
This crucial step can make or break the overall initiative. We offer a proven approach
to creating and finalizing the business case for big data analytics that is specifically
relevant from a CRM perspective.
The use case-driven approach can help map
the business requirements tightly with the big
data technology design considerations, such as
relational storage and query, distributed storage
and processing, and low latency/in-memory.
This cost-benefit analysis-based approach can help define the stakeholders and
detail the use case while also assessing ROI. It helps answer questions such as:
• How do you approach your first big data implementation?
• Do you have the information necessary to determine the approach?
• How can you ensure you receive the business value of the big data journey?
• What metrics and cost factors affect the success of your big data program?
The output of this step provides the company with a business case and an ROI
calculation to ensure management will fund the initiative. More than a proof of
concept, this process results in a proof of value and helps customers understand
the business relevance, technology challenges and economic hurdles of a typical
big data/analytics engagement.
Phase 3: Design the big data architecture and configure
the analytics engine.
Once the business use cases have been crystalized, the big data architecture and
analytics engine needs to be designed for focused analysis and to derive actionable
insights for different stakeholders. This significantly reduces the time to value and
also brings a sharp focus to the expected business outcomes.
The use case-driven approach can help map the business requirements tightly
with the big data technology design considerations, such as relational storage and
query, distributed storage and processing, and low latency/in-memory. This, in turn,
leads to a sustainable and scalable architecture.
The analytics engine must then be configured for linking datasets around an entity
(e.g., what do I know about this customer?) or around a relationship (e.g., how is this
customer related to others?) Successfully configured, such analytics can produce
qualitatively new insights that result in business value, such as reduced customer
churn rate, next best action and better predictions of risk and failure.
TURNING CUSTOMER KNOWLEDGE INTO BUSINESS GROWTH
9
10. Making Meaning from Digital Fingerprints
Applications
Improved
Upsell/Cross-sell
Real-Time
Offers
Enhanced Marketing
Effectiveness
Proactive
Servicing
Personalized
Campaigns
Service
customers
proactively using
social listening.
Use the multidimensional
profile to personalize
campaigns.
• Positive
• Campaign acceptance
Objectives
Delight customers and
cross-sell/upsell by
making intelligent, realtime recommendations.
Create offers and next
best offers on the fly
based on updates to realtime customer profiles.
Fine-tune offers and channel
effectiveness during
campaign planning and
creation.
Success Metrics
• Increase revenue
generated from
cross-sell/upsell
offers.
Increase share of
wallet.
Increase customer
satisfaction scores.
• Increase number of
• Reduce marketing
•
•
real-time offers sent.
Improve offer
acceptance rate.
Reduce customer offerrelated spending.
Reduce turnaround
time for offer.
•
•
•
campaign costs.
Increase lead conversion.
sentiment
service level.
Customer
loyalty.
rate.
•
•
Figure 8
Phase 4: Create real-time, multidimensional
customer profiles.
Once the multidimensional customer profile is established, the possibilities are
endless. The profile provides access to customer data residing not only in the
enterprise but also from every other area in the external world with which the
customer has interacted. In essence, the profile captures every digital trace that
the customer creates. This invaluable data can now be exploited for driving several
applications (see Figure 8).
Challenges Along the Way
Companies can expect to be faced with several challenges when developing multidimensional customer profiles, including:
• Data explosion: Customers are increasingly interconnected, instrumented and
intelligent. Accordingly, an unprecedented velocity, volume and variety of data
is being created. As the amount of data created about consumers grows, the
percentage of data that businesses can process quickly decreases, because traditional systems cannot store, process and analyze massive amounts of structured
and unstructured data. Business systems are not designed for today’s unstructured data, rapidly changing schema and elastic scaling of storage.
• Privacy and regulatory issues: Another issue is regulatory and privacy issues.
Data collectors bear a tremendous responsibility to provide full disclosure of
what they plan to do with customer data. But an even greater challenge is the
sharing of data.
For instance, if a consumer grants one company permission to use his or her data,
what rules (if any) will regulate how that information is shared across multiple
companies? Such questions will become one of the biggest sticking points in
terms of trying to navigate the right policies.
10
KEEP CHALLENGING
December 2013
11. Data collectors also need to make it easier for customers to opt in or out of having
their information used, similar to opting into mailing lists or using an “unsubscribe” option to opt out. When consumers feel they’re getting a tangible benefit
for their personal information, their resistance to data collection starts to fade.
Loyalty and rewards programs are a good example of how companies can persuade
customers to reveal more details about behaviors such as shopping habits.
Looking Forward
Leading organizations are already gearing up to create multidimensional customer
profiles using both structured and unstructured data sources. Complete and
continuously up-to-date customer profiles enabled by big data and analytics are
increasingly an essential tool in the arsenals of organizations across industries and
geographies. `
Footnotes
1
“Marketing ROI in the Era of Big Data: The 2012 BRITE-NYAMA Marketing in
Transition Study,” Columbia Business School and NYAMA, 2012, http://www4.gsb.
columbia.edu/null/2012-BRITE-NYAMA-Marketing-ROI-Study?exclusive=filemgr.
download&file_id=7310697&showthumb=0.
2
“Maximizing Customer Lifetime Value with Predictive Analytics for Marketing,”
Aberdeen Group, February 2013, http://www.aberdeen.com/_aberdeen/public/viewlookinside-pdf.aspx?cid=8362.
About the Authors
Sairam Iyer is a Senior Information Management and Analytics Consultant with
Cognizant Business Consulting’s Enterprise Information Management Practice.
His core responsibilities include providing thought leadership in the areas of
business intelligence and analytics, and consulting with clients across industry
verticals. Sairam has nine years of rich experience with Fortune 100 companies,
specializing in CXO and business leader-level workshops to understand business
processes and concerns and convert them into business intelligence and analytics
solutions. As a multidisciplinary BI strategy expert, he has hands-on experience
in executing information management and analytics engagements from concept
to delivery. Sairam obtained his M.B.A. from the Xavier Labor Relations Institute
(XLRI), Jamshedpur, specializing in marketing and strategy. He can be reached at
Sairam.Iyer@cognizant.com.
Vikas Singhvi is a Senior CRM Consultant with CBC’s Enterprise Applications
Services (EAS) Practice. Vikas’s core responsibilities include working on consulting
projects in the sales, marketing and customer service domains across industry
verticals. He has four-plus years of progressive experience in business strategy,
customer relationship management consulting, digital marketing consulting,
sales and marketing process consulting and business development. His consulting
experience includes extensive multicountry project exposure across the hightechnology, retail, manufacturing-logistics, information services and transportation domains. Before joining Cognizant, Vikas worked with Microsoft India as an
APEX (Accelerated Professional Experiences) member, which is a program for highpotential entry-level employees. Vikas received his M.B.A. from the prestigious
Indian Institute of Management at Indore, specializing in marketing and strategy.
He can be reached at Vikas.Singhvi@cognizant.com.
TURNING CUSTOMER KNOWLEDGE INTO BUSINESS GROWTH
11