Analyzing Customer Behavior

3

Peter Hinssen, editor

Analyzing
customer
behavior
Predicting
what happens
next

i
Dear reader
“ Analyzing customer Consumers are no longer what they used to be. Before they buy, they look around. A lot. And

behavior in the age
when they buy, they talk about it. To everybody. They want to be treated as unique. They want
you to listen to them. They want you to talk to them. They want you to thank them. In short,
they are influencers and mini-tyrants who will leave you in the blink of an eye if you don’t do as

of big data” they please. And that’s the easy part. The problem is that they change so fast that you actually
need to know what they want, before they do. Your organization needs to outrun them. It
needs to be agile and fast.
is the third in a series of thought-provoking
The good news is that you have all the information needed to understand, catch and nurture
booklets that Across Technology will be
these empowered consumers. Because they literally leave traces everywhere. Not just inside
publishing this year, sponsored by EMC
your company, nicely stacked in shiny databases, but outside as well, in a colossal mass of
Greenplum. These booklets are elements
unstructured external data. I’m talking about data from Twitter, Facebook, industrial sensors,
in the Data Science Series, which is also
geotracking data, voter registrations, sports statistics, website logs and much much more.
a series of events, and a website:
They are there for the taking. You only need to link the two and unlock their value. And be able
to store them, of course.
www.datascienceseries.com
Not only does Big Data come with real predictive insights but with actions and results as well.
Churn prediction, in-store behavior analysis and recommendation engines are just some of
the cool possibilities that it fuels. It’s customer-centricity on speed.
Publisher:
With this booklet, we want to offer you views on the way markets are changing and how
Across Technology
you can make informed use of these trends to strengthen your company’s position. To keep
Editor: you informed on a constant basis, we have created the Data Science Series website (www.
Peter Hinssen datascienceseries.com), offering you case stories from your peers, valuable insight into
market research and an overview of the Catalyst partners that help EMC Greenplum bring
Executive editor:
the right building blocks to the market. Allowing you to build the right ‘refinery’ for all the
Philippe Gosseye
information that is coming your way.
Contributing editors:
3


ANALYZING
José Delameilleure, Laurence Van Elegem 1 2

CUSTOMER
BEHAVIOR
4


OPEN DATA

INFORMATION THE AGE OF PREDICTING

POWER
WHAT HAPPENS

IS THE DATA-DRIVEN NEXT

NEW OIL MEDICINE SMART CITIES
HOW BIG DATA TURNS EVERY
DRILLING NEW BIG DATA HELPS

Layout:
CITY INTO A DATA CAPITAL
SOURCES REVEAL HIDDEN
OF INNOVATION HEALTH TRENDS
AND BUILD
RISK MODELS

Stijn Van Herck

Make sure you don’t miss the installments of the series. Please contact your local EMC
Greenplum organization to obtain all of these booklets.
SPONSORED BY

3

01
Fast Customers + Big Data
= Big Marketing

It is a plain and simple truth: you need customers, and loads of them, to function (in the worst case)
and thrive (in the best case).

The problem is that, these days, your customers are changing at the speed of light: faster than the
market, faster than individual companies and maybe even faster than you. If that is the case, your
company better become agile and nimble in order to keep up with them … and your competition.
It’s a ‘sink or swim’ kind of situation. There’s no in between.

The good news is that you have all
the information you need - for a 360° view of your
customers - staring you right in the face.
I’m not only talking about the valuable information that resides inside your company – and
which is not always used to its full potential – but also about the massive amount of external
information that can give you superior insights into your customers’ minds and behavior. I’m
talking about the gargantuan quantity of data coming from Twitter, Facebook, industrial
sensors, geotracking data, voter registrations, sports statistics, website logs and much
much more.

And it gets even more exciting when the internal and external bits and bytes are combined
and turned into consumer insights that were just plain impossible only a few years ago.
Gone are the days of intelligence that can only help understanding the past. Say ‘hello’ to
predictive analytics. Say ‘hi’ to speed and actionable knowledge.

4
04 5

01 Fast Customers + Big Data = Big Marketing

02
Meet Big Data. If you think that ignoring it,
will make it go away. Think again. It is here to stay. The tyranny of the
Deloitte values the size of the Big Data market at about $1.3-$1.5 billion in 2012, and it is growing
empowered consumer
fast. In 2009 there were only a handful of Big Data projects and total industry revenues were
under $100 million, while by the end of 2012 more than 90 % of the Fortune 500 are estimated Today’s consumers have evolved beyond being merely ‘buyers’. They have grown into ‘influencers’,
to invest in Big Data initiatives one way or another. and highly critical ones at that.

According to McKinsey & Company, a retailer using Big Data analytics to its full potential can
The customer as influencer
add 60 % to its operating margin. If you don’t, don’t forget that your competition will.
While only 6% of all purchases are done online today, no less than 37% of all offline purchases are
inspired by online research. Today’s consumers know the market and your competition through
Big Data can help you to:
and through. So you better know what they want if you want them to buy what you’re offering.
• Gain a 360° view of customer activity across channels
When consumers do actually acquire your product or service, they are very vocal about it, in
• Discover what influences buying behavior all their networks of intelligence. And their reviews – be it positive or negative – will be sure to
• Tailor your services and products to what your customer wants influence their friends, family and peers who, in turn, will influence others, etc., etc.

• Improve customer satisfaction

• Drive customer loyalty

• Optimize existing customers

• Adapt your prices to the market
“ Honest businessmen
should be protected from the
• Drive the ROI and performance of your marketing campaigns unscrupulous consumer.”
• Cut budget without cutting performance
Lester Maddox

… and all this at the speed of change.

Gone are the ‘markets’ of yore. Say hello to the ‘networks of intelligence’, where consumers
have turned into extremely informed and hard-to-please, networked thinkers.

6 7

02 The tyranny of the empowered consumer

A 2011 survey of the ‘Economist Intelligence Unit’ revealed that - according to 30 % of business
leaders across financial services, education, public sector, IT and manufacturing industries
HAPPY CUST0MERS - by 2020 customers will be the main source of new product or service ideas, with online
communities in second and R&D in third place. Yes, by then we will be a long way from Henry
Ford’s “If I had asked people what they wanted, they would have said faster horses” -approach.

From monologue to dialogue

With consumers of today being such volatile and vocal influencers and leaders, monologue is no
longer an option. Smart marketers engage in dialogue with their target groups, if possible in a
one-to-one relationship.

“ When you focus on the consumer, the consumer responds.”
Alexander Wang

TELL 9 PE0PLE TELL 22 PE0PLE
But there are not only challenges to deal with here. Difficult as they may be to catch and please,
consumers that talk about products and services on social networks, leave traces everywhere.
And unlike word of mouth and casual conversations, their ‘breadcrumbs’ can be found, analyzed
and measured.
According to Allbusiness.com happy customers tell, on average, 9 people about their experience
while dissatisfied ones vent their frustrations to 22 persons. Just one airline passenger
complaint letter republished by the ‘Telegraph’ resulted in more than 1,100 retweets and over
3,000 backlinks.

Today’s organizations are facing a very public cycle of buying and influencing which can turn
into a retailer’s worst nightmare if he is not careful.

Loyalty, so last season

Forget about brand loyalty. Loyalty is no longer relevant. You need to keep consumers
interested, surprised and satisfied or they’ll go elsewhere, in the blink of an eye. And if they stay,
they expect you to reward them, in exactly the way that they want.

Ask not what the customer can do for you; ask what you can do for him. Customer centricity is
Consumers leave Big Data
not just a buzzword. It’s how your organization will survive.
footprints everywhere.
Market leaders have already
The customer as leader
grasped this opportunity
for tracking and understanding
It used to be companies that led the market with their innovations and deciding what was best their behavior and sentiments.
for the consumers. Now the customer is in the driver’s seat of the most successful companies.
Even more so in emerging markets, where their expectation rates are even higher and they are
even more likely to switch brands.

8 9

03
Big Data technology
is remolding consumer- Back to the future:
vendor relationships one-on-one contact is back
into something
personal, dynamic with a vengeance
and dialectic
“Sometimes you want to go where everybody knows your name”,so the song goes.
And it used to be like that when you went to your neighborhood shop: the manager knew who you were,
what you wanted and what you might like.

ConsYOUmerization!

Today’s consumers want that back. They need to be treated as unique. They want a far-reaching
personal experience. Calculating averages will no longer do. In today’s seller’s market it’s ‘Customize
or Bust!’. We are back to virtual one-to-one marketing and hyper segmentation.

Dialogue and personalization are what consumers want. Big Data is how you can achieve that.

Contradictory as it might seem, the digital age and powerful Big Data technology are remolding
consumer-vendor relationships into something personal, dynamic and dialectic.

With all the data you need to understand your customer at hand, you only have to unlock their
intelligence. Today it is possible to make a full digital footprint analysis of purchasers by auditing
and merging internal and external data feeds, turning those into real insight and making real-time
decisions based on individual profiles. Just think of Amazon which generates about 20% more
revenue since it started recommending products to its customers.

“ The first step in exceeding your customer’s expectations is
to know those expectations.”

Roy H. Williams

You need kNOWledge about your customers

Tuning is not just fueled by correct information; it needs to be driven by speed. Your customer
will have changed already if you took your time to react. There is no time to react. Consumers
are all about the ‘Now!’.

10 11

04
03 Back to the future: one-on-one contact is back with a vengeance

If they receive your recommendations or your discounts at the right time, they will be happy. Making new customers
AND keeping them
If the tips come too late, they will just annoy them. In a 2012 survey by the Radicati Group, on
average, 19% of emails received were considered spam by the recipient. So you better hope they
were not talking about you.
McKinsey claims that 55% of the current marketing budget is spent on new customer acquisition
There is a fine line between welcome advice and spam, and only 12% on customer retention. However, according to ‘Leading on the Edge of Chaos’
and it is defined by veracity and velocity. (Emmet Murphy & Mark Murphy), you have only a 5-20% chance of selling to a prospect. The probability
of selling to an existing customer is between 60 and 70%. So money spent on preventing churn, is better
spent than that which is invested in acquisition.

Smart companies are increasingly betting their money on cherishing their existing customers, while
the smartest ones find out which customers deliver the most revenue and focus on those, while
ignoring the ones that are too expensive to keep.
DATA + VERACIT Y + VEL0CIT Y
Big Data can warn you which customers are on the verge of switching to the competition.

How? Among other things, Big Data tools tell you when customers are making negative comments
about you online, when they have partially switched to the competition or when their shopping
basket has changed in content or size. Big Data solutions ring the alarm before the damage has
been done and give you the opportunity to prevent.

Reducing your customer defection rate by 5% can increase your profitability by 25 to 125%.
A 2% increase in customer retention has the same effect as decreasing costs by 10%.

The availability of Big Data is the best and most convincing argument for empowering customer
centricity rather than relying on a one-sided acquisition strategy.
Pretty databases filled with lazy data and lethargic intelligence that takes hours, days and
even weeks be delivered will no longer do. When markets become networks of intelligence Big Data can help you thank customers for their loyalty.
you don’t have that luxury anymore.
Consumers that abide with you expect to be rewarded for that, or they will move to pastures new.
You need to keep moving faster. You need systems that can - NOW - deliver the intelligence you Big Data fuels the world’s most successful customer loyalty programs. It can help create member
need for closing the gap between your promises or intentions and what the consumer wants. micro-segments based on loyalty data and identify which marketing channels, products and
services work best for which segments or evaluate how effective a campaign is or if and how
it needs to be fine-tuned.

And a ‘smart’ loyalty program can have a far-reaching effect. For example, it helped a specialized
US grocery store increase its sales by 58 % in less than 4 years, while its gross margin grew by 3%.
A European mobile phone brand succeeded in reducing customer churn by 30% that way, while
a European telecom operator saw a 14% increase in revenue per customer for those following the
program.

12 13

05
So where does this
BIG DATA hide?

To fully understand the Big Data phenomenon, you need to picture the three Vs that drive its value.

There’s ‘Volume’, of course, as Big Data comes in staggering numbers:

• 2.2 million terabytes of new data created every day
• By 2013 approximately 507 billion e-mail messages will be sent per day (Radicati Group)
• In December 2010 there were more than 267 billion websites (Netcraft)
• Twitter: more than 230 million tweets per day
• Facebook: 100 terabytes of data uploaded daily
• YouTube: 3 billion visitors per day & 48 hours of video is uploaded per minute
• Walmart processes a million customer transactions every hour and incorporates it into
its 2.5 petabytes of data. That is 167 times the information contained in all the books
in the US Library of Congress!

Next up is ‘Velocity’. Because Big Data is fast. It is being generated at an incredible speed and
(should be) collected and turned into intelligence as fast.

14 15

05 So where does this Big Data hide?

Last but not least comes ‘Variety’. Because Big Data is very complex and comes in all types: from There is no getting around it: unstructured data constitutes a mountain of input: emanating from
neatly structured into the column- and row-based databases of your company, to a wildly soaring chat records, text messages, website logs, Twitter, Facebook, LinkedIn, Google+, blogs, the sports
chaos of unstructured, external information. Powerful as they each are, there is no telling what they season, medical records, smartphone analytics, geotracking data, news articles, industrial sensors,
can do when they interconnect. weather satellites, economic data sets, location data, voter registrations, photographs, movies,… I
could go on and on for pages here, but I believe you get my drift.
Internal (semi)structured data
Data is expected to grow over 800% in the next 5 years. No less than 80% of it will be unstructured.
Your company is sitting on top of a goldmine of information, coming from all kinds of different
And you can put them to work to better understand your target group.
sources: point of sales, online transactions, logistics data, sourcing information, market research
CRM solutions, satisfaction surveys, customer service, pre-sales histories, e-mails, call data records
Context is everything
loyalty cards, call and contact center interactions, salesforce automation data, traditional
advertising response, credit card purchases, web clickstreams and many more…

One of the biggest issues is storing the data. A lot of companies do not have the necessary
technological power to keep all current and historical material. So they sift and only keep what is
vital for them now, not always realizing that something that might not be useful now, can prove to
be a real blessing later.

Break the silo-nce and start sharing!

Another challenge is that most of these data residing within companies is unconnected. A lot of
departments and their respective leaders often seem to have missed the life lesson on sharing and
Always put your information in the right context. Finding out how your organization performs
are storing ‘their’ information in completely closed off and therefore – pardon my French – stupid (I
against its own benchmarks is good, but knowing how its performance compares to the ‘real’ world
mean that quite literally) siloes.
is even more crucial. Mix, match and compare data from within and from outside.

Possessing information that you cannot interconnect and compare is the true death of dynamism
A report might show that 50 cents’ profit is made on every item that your company manufactures. It
and intelligence and will cost you a lot in lost revenue.
could also reveal that this has increased year on year, and that volumes have grown by 5%. You would
then probably assume that everything is fine and that profitability will be sustained in the coming
External unstructured data
years.
You cannot understand and engage consumers by looking in the mirror – by which I mean (often very
slowly) analyzing data that reflects your own company – but have to look beyond yourself, through However. If analysis of external data showed that your competitors made 60 cents’ profit on the
the windows of your organization, at the even vaster sea of data that is available out there. same items, that their volumes were growing by 20% making them the market leader, and they were
planning to drop their prices by 70%, then the outlook wouldn’t be so rosy.
Some companies are scared and paralyzed by the sheer number and size of Big Data. They need to
Big Data gives you the context you need to make correct decisions.
realize that Big Data is not the enemy; it is their ally against customer churn and their assistant for
driving revenue.

16 17

06
Big Data, Big Brother?
Before I go any further, we need a moment of reflection on the issue of privacy.

I am not so much talking about compliance with local and federal data regulations, as about striking
the right balance between privacy and utility when using big consumer data.

Disclosing personal information is part of modern life. Most consumers have come
to terms with that, but that does not mean they are happy about it.

Privacy versus enhanced consumer experience

The key is to use their on- and offline tracks for ‘Good’. By enhancing the customer experience
with better advice and less impersonal spam and junk mail. Or through special offers and
loyalty bonuses. It is a clean-cut give and take situation, really.

But do not overstep any boundaries…

Why? Consumers perceive it as intensely creepy when you address them in a manner
that is too personal.

An example is that of a well-known American retailer that sent coupons that were timed to the
very specific stages of their customer’s pregnancy. This was calculated on the basis of behavior
analysis of women that had signed up for their registries. It revealed that they were buying
Consumers allow larger quantities of unscented lotion at the beginning of their second semester, loading up on

the use of their personal calcium and magnesium supplements in the first 20 weeks, buying hand sanitizers just before
the delivery date,…
data if it enhances their
customer experience. Because their customers felt uncomfortable by being approached so directly and correctly, the
leading retailer learned to be more discreet and less ultra-targeted. They sent more general
information along with the focused ads and coupons and everyone was happy again.

Although this is a perfect example of the endless possibilities of Big Data when it comes to
targeting, I also want to use it to stimulate some reflection about the ethical boundaries of
leveraging Big Data.

So when you decide to launch your organization in big analytics, first establish some
governance policies that cover privacy, IT security and other regulatory issues.

18 19

07
CMOs and CIOs as BFFs
In a recent past, most CMOs used to be veterans with impressive track records, longer than ‘Lord of the
Rings’ (back in the days when it still used to be a book). They based their strategy on experience and gut
feeling. Today those in charge of marketing are often a lot younger and more technologically savvy. They
realize that corporate decisions are too important to be based purely on instinct. Marketing is not just
an art, it is a science, too.

Don’t play dice with your company. A 2010 study
from MIT revealed that the adoption of ‘data-driven decision
making’ can result in a 5 to 6 % rise in productivity.
Today ‘data-driven’ actually means Big Data-driven. CMOs cannot afford to miss the boat just
because they cannot comprehend or fear the immense mass of information that is at hand or
the technology that is needed to drive it. There is too much at stake for their organization. They
need to team up with the CIO to make Big Data work for their company.

Gartner predicts that, by 2017, CMOs will have greater control over the IT budget than the CIO.

So CMOs and CIOs have to work together to create a better customer experience and put
the business in a position to succeed. Pretty straightforward as this might seem,
a recent survey by ‘Forrester and Forbes Insights’ revealed that the majority of organizations
are not ready to merge their marketing and IT skills. There is still a long way to go when
it comes to IT marketing alignment.

The solution for the Big Data conundrum comes in the form of the currently very ‘hot’ data
scientist, who is the best possible negotiator between the CMO and the CIO as he has
the ability to speak both technology and business language. Data scientists offer the rare
combination of in-depth expertise in statistics and machine learning and the business skills to
know which data are relevant and which insights can drive revenue.

Currently, however, data scientists are as hard to find as Cerulean Warblers (yes, those are
hard to find, just ask Jonathan Franzen). McKinsey & Company estimates that the “United
States faces a shortage of 140,000 to 190,000 people with analytical expertise and 1.5 million
managers and analysts with the skills to understand and make decisions based on the analysis
of Big Data.” So start looking already.

20 21

08
How can IT help Marketing
master Big Data?
• How can they be filtered by relevance?
• Are we talking living content or historical data?
Clearly, there are five domains for which technology can provide the right answers for • What kinds of external data are at hand and which ones are relevant?
the Big Data challenge and create order out of seeming chaos: • Can they be acquired freely or should we earmark a budget for that?

1. Storage & processing power: making room for the Big Data Data Quality Best Practices boosts revenue by 66%.
2. Data Management: using the right Big Data
3. Linking the data: making sure that all your centers of Big Data are Customer analytic processes are only as valuable as the quality of the underlying data. So
able to communicate with one another a well-funded data management strategy is key. This means getting rid of fragmentation,
4. Analytics: Filtering intelligence from the Big Data duplication or inconsistency. Be sure that the data you have is complete and correct.
5. Visualization: making sure that non-technical business Be certain that the external data sources you use are not out of date and relevant to the
users have easy access to the insights analytics that you want to perform.

There is still a lot of work to perform when we consider that:
Making Room for Big Data
• Fewer than 1 in 5 managers believe they have all the information they need to perform
CMOs are all about improving customer satisfaction, boosting revenue and adding strategic
their jobs effectively and fewer than 1 in 10 believe they have realized the full return from
value to the business. Their priority list is not likely to acknowledge storing, managing and
their technology investment.
analyzing massive amounts of data. But it should.
• The average management report is 30 to 40 pages long, contains 12-15,000 data points,
Organizations today need Big Data to understand and please the consumer. If you cannot yet managers use less than 5% of that information.
store the massive amount of Big Data, you obviously cannot use it. Or even more correct: if you
cannot store Big Data without slowing down all of your IT and business processes, you cannot Linking the Big Data
use it.
There are two challenges when it comes to Big Data and linking: one internal
So, first things first: you need a robust and agile Big Data platform with massive processing and the other external.
power. And it is vital that it is purpose-built to handle all the applications that will turn the raw
Most organizations have a multitude of databases that are disparate and unconnected in
data into understandable and actionable consumer and business information.
silos. Making those modules communicate will allow them to answer many of the growing
challenges that companies face today, such as what products are selling, what do my
Using the right Big Data
consumers look like, what is the marketplace doing or how do I compare to the competition?
Those in charge of the technological side of Big Data – be it the CIO or the data scientists –
The biggest challenge, however, will be to merge the internal data with the colossal amount
should ask some important questions before they start using the data for business value:
of unstructured external information from open source clusters. And it is exactly this
• Which departments and employees are creating data? integration of Big Data with traditional data that will allow you to gain a 360-degree view
• How trustworthy are these internal sources of data? of your customers.

22 23

08 How can IT help Marketing master Big Data?

Unlocking the value of Big Data
In-memory computing enables rapid analysis of large volumes of customer behavior data and
Storing and managing the massive amounts of Big Data is, of course, just a small part of
real-time decision making by allowing you to collect, manage, analyze, provide and finally act
the solution. Obviously, if you want to make the Big Data work for you in terms of real-time
on data in the moment.
marketing, actionable information and a real customer insight, you need powerful, next-
generation analytics to unlock its value. Current analytic technologies do not put the source data – such as the types collected
on Facebook and LinkedIn – in a straitjacket, prior to storage, like traditional BI and data
warehousing systems do. This lets the discovery analytics processes run against raw data and
allows the user to tap into even larger volumes of data and run algorithms and other analytic
processes without going through the I/O bottleneck to access data stored on disk.

In-memory computing is a great combination of real-time and low cost, transactions and
analytics as well as open and native.

Gone are the days of slow-process analytics that were relying on weeks or even months-old
transactional data, that were relatively poor in volume and therefore in content. Say hello to
speed and agility.

Insights that once took days or weeks to develop can
now be produced in hours, minutes or even seconds.
The ‘rear view mirror’ approach has made way for predictive information that can drive real-time
Look forward, not back
decisions and actions.
Analyzing things that are in the past will just no longer do. What you need are predictive analytics
tools, helping you to foresee future actions from customers. Is he on the verge of leaving you? If so, Delivering Big Data insights to the business users
how can you lure him back in? Or what can you recommend to a new customer browsing on your
Big Data insights need to be delivered to those who can turn them into revenue-driven actions.
e-shop that will really appeal to him? These are the questions you need an answer to.
Easy-to-understand dashboards and data visualizations are a big part of that. Keep that in mind
Next-generation intelligence tools are scalable and can handle and make sense of the enormous when looking for your Big Data tool kits.
amount of data at hand. They can accommodate analysis of both structured and unstructured data
Insights that can only be interpreted by the IT department and are too complex for business
from open source environments like Hadoop. Traditional enterprise data warehouse and business
users are a total waste of time and money.
intelligence tools cannot offer you that. They cannot offer an integrated view of complex customer
behavioral data. Business users also need self-service capabilities for working with data so they can use it for
decisions and predictions. They want accurate information, but velocity is as important so
In-memory computing: making Big Data faster they cannot afford to wait until the IT department finds the time to perform and interpret the
analytics. They need access themselves, at all times and as fast as possible.
Moving these large volumes of data around to large data warehouses so that they can
communicate and be compared is a time-consuming and expensive process. This is where a key The question “How can we make sure that the information reaches the right people, … and
feature of high-performance analytics, in-memory computing, comes in really handy. fast?” should be on the minds of IT at all times.

24 25

09
“There is only one boss.
The customer. And he can fire So, concretely, what can
everybody in the company
from the chairman on down,
Big Data analysis do for you?
simply by spending his money
somewhere else.“
Big Data tools are key for customer centricity and a competitive edge. Big Data will enable you to
know your customers so well that you can give them exactly what they want: personalization, dialogue
Sam Walton
and tempting rewards.

The possibilities of using Big Data for analyzing customer behavior are countless:

• Customer micro-segmentation • Determining customer lifetime value
• Finding the influencers • Increasing portfolio penetration per customer
• Campaign performance optimization • Targeting cross- and up-selling opportunities
• Call center monitoring and reporting • Location-based marketing
• Churn prediction • In-store behavior analysis
• Detecting and measuring lost sales • Sentiment analysis
• Loyalty program analytics & targeting • Recommendation engines

But here are some of the most exciting ones:

Campaign performance optimization

Big Data analytics can allow you to monitor and determine the effectiveness of marketing
campaigns, determine their ROI and adjust them for better performance. This can significantly
reduce wasteful spending and increase gains relative to costs. Something which is on every
marketer’s mind.

The digital media company Havas Digital uses Big Data to provide attribution for its clients’
marketing efforts and to understand what was really driving the results that its clients
expected from marketing. For instance, someone was exposed to a video and then searched
for it on Google. Classically, in digital marketing, the last exposure will be considered the most
important while, in this case, it obviously is not the case. So Havas Digital offers sophisticated
methodologies to be able to spread credit for the sales across those exposures. For one client in
the travel sector in the UK, they were able to improve the ROI by 300% with 27% more sales.

26 27

09 So, concretely, what can Big Data analysis do for you?

Big Data can reveal that and let you know if you need to adjust your product selection, pricing
Call center monitoring & reporting
or the way it is displayed in your shops.

“ Courteous treatment will make a customer a walking advertisement.” In-store behavior analysis
James Cash Penney
Big Data insights bridge the online-offline dichotomy. They provide retailers with the possibility
Companies that work with call centers can activate Big Data analytics to monitor and report to mine their databases in real time when a customer is browsing their store and extract (from
the calls with fully automated solutions. These convert voice to text and allow calls to be ranked past purchases, clickstream behavior or activities on the company’s Facebook page) how he
in different areas such as operator friendliness or cross-selling and up-selling opportunities. should best be approached or what he is willing to pay.
Some even offer the additional benefit of being able to look at e-mail correspondence and
Some of the leading-edge retailers apply Big Data technologies to analyze video streams from
online chat.
their in-store camera systems and create mappings of customer foot traffic throughout the
Solutions like these offer obvious implications for performance management, quality stores. This Big Data stream is then merged with sales data which allows retailers to develop
monitoring, risk and compliance optimization, sales effectiveness and service satisfaction. new applications that help optimize product placement in the store.

Churn prediction Sentiment analysis

Big Data technologies enable you to analyze behavior patterns that indicate which customers Using real-time mapping, web data extraction ‘brand listening’ techniques and sophisticated text
are most likely to leave for a competing vendor or service and to find out if they are worth the mining tools can allow you to follow conversations on social media sites, filter them for relevance
money you would invest in keeping them. Action can then be taken to save the most profitable and process the content using natural language algorithms. This sentiment analysis enables you
of these customers. to determine how consumers really feel about your company, brand or services, from a macro-level
down to individual user sentiment.
For instance, you could use speech analytics to monitor the conversations of customers who
were terminating their accounts and identify other at-risk customers by comparing their These insights can even be integrated with retail point of sales (POS) data from tills to better
discourse with those tell-tale keywords and phrases. An online case reported that a company understand how emotions about the brand translate into actual sales or, when matched to news
thus saved some 600 accounts and more than $12 million in revenue in the first year of the stories and weather conditions, to enhance the communication with your clients. They enable you
program. to target people who already like your brand individually in-store or with relevant digital marketing
content.
Mobile companies might learn through Big Data if a subscriber changes their SIM, swaps
between multiple devices or even receive a warning if a customer ventures on the website of As we speak, Big Data technology companies are developing increasingly sophisticated approaches
their competition through mobile clickstream analysis. to tap into the power of the social media. Not only can you eavesdrop on what people really
think about your company, but there are tools that allow you to re-feed this intelligence to these
A European mobile operator was able to identify ‘promotional churners’ by tracking a sudden
networks.
surge of activity on a competitor’s website. The operator was able to immediately launch a
counter promotion to successfully prevent churn.
Recommendations

Detecting and measuring lost sales This is perhaps one of the coolest goodies to result from the marriage between Big Data analytics
and customer behavior.
As important as keeping the most profitable customers and finding new ones, is detecting and
evaluating the lost sales: if you are able to understand why an action did not result in an actual Recommendations are all about customer micro-segmentation and improving their experience.
sale, you can adapt your strategy and up your sales. The best-known examples are, of course, Amazon’s use of real-time, item-based, collaborative

28 29


filtering (IBCF) to fuel its “Frequently bought together” and “Customers who bought this item
also bought” features and LinkedIn suggesting “People you may know” or “Companies you
may want to follow”.

Amazon generates about 20% more revenue
via its recommendations.
Making effective recommendations is a real art, and is not so much improved by using smarter
algorithms as it is by feeding the engine many, many, … many data sources: purchase history,
product ratings, products a visitor looked at but didn’t buy, products their friends bought, responses
to previous recommendations, products bought by people in a similar geographic region, products
purchased by people with similar tastes, etc.

You could even refine the suggestions up to the type of place and time of day the searches and
purchases are done. Or your Web shop might, for instance, suggest ‘Interesting products for today’,
like a warm winter coat because it’s snowing outside or binoculars for a Facebook friend whose
birthday it is and who very recently joined several bird-watching organizations.

You need Big Data to feed the recommendation engine, because when suggestions are not correctly
tailored, they will be considered as spam. Things like “Time to have your car checked?” or “Looks like
you will need a new printer cartridge soon!” walk this fine line between blessing and bother. When
they are spot on, however, they can prove to be a gold mine for the recommender.

Walmart is another company that understood the revenue-driving potential of suggestions and even
took it up a notch with their gift recommendation Facebook application ‘ShopyCat’. Not only does
it use the Facebook profile of its subscribers to recommend suitable products based on the interests
and hobbies of friends, but it will also offer links to the competition if they do not stock a suggested
item in their own stores. Knowing how critical consumers have become, their partially neutral
approach might very well prove to be the future of marketing.

Finding the influencers

You can mine Big Data from social networks to uncover the ‘halo effect’ of influencing consumers
on social media. These are not necessarily your most important customers – which can easily be
discovered by analyzing internal data and checking how much they buy or spend – but they are
equally essential to a company as they are the ones that tend to influence the buying behavior of
others the most.

Product and service innovation & optimization

A 2011 survey by the ‘Economist Intelligence Unit’ predicted that customers will be the main source

30 31


of new product or service ideas by 2020. If they are right, organizations that swear by their R&D
departments for innovation will need a serious change of mindset.

This is especially important in emerging markets, where customers are known to be particularly
engaged in product & service innovation – at double the rate of consumers in mature markets even.

Big Data solutions can help comb through enormous volumes of text-based research and other
historical data to assist in the development of new products. They also scan the Web and social
media to understand how the market develops or customer needs evolve. This will allow you to
adapt the design of existing products or create new ones accordingly.

IS BIG DATA
THE DEATH 0F PR0FIT ?

As with any other business opportunity, not everyone is pleased with the Big Data
phenomenon. Opponents reject it because some market leaders actively scan
their competition online to ensure the lowest advertised price. Such tactics can
induce a downward push on all prices in the market and there is a risk, then, that
companies will see their margins squeezed as they try to compete.
The visions that
New technologies that were introduced together with Big Data have the same arise from big data
effect by giving shoppers powerful pricing, promotional and product information, are actionable. you
frequently in real time. Applications like RedLaser, for instance, let shoppers scan
will be able to turn
the bar code of an item in-store with their smartphones and obtain immediate
price and product comparisons. your business around
immediately
Let’s face it, there is no sense in fighting how the market evolves. It is very much an
‘adapt or die’ situation. Just keep in mind that Big Data offers far more advantages
than it does drawbacks.

32 33

10 02
Clearing the way
for Big Data?

Realizing the immense potential of Big Data technology for understanding customers and driving
your sales is already a big step.

The next one is preparing your company – and yourself – for a matching strategy.
“There is nothing so terrible
as activity without insight.“
Embracing change is always hard. Especially when the insights that it will bring might uncover
that your marketing tactics are not as efficient as you expected or that your company suffers
from a bad image. Just realize that the visions that will arise from Big Data are actionable. You
will be able to turn your activities around immediately. Johann Wolfgang von Goethe

Besides knowledge, Big Data also comes with results.
Next up - like with any IT-driven project – is to convince your organization about the value that
investing in Big Data technology will create. We are still in an age where many people view IT
as a matter for the back office and a cost center rather than a dynamo for business growth, so
this might prove to be challenging.

Just don’t forget that Big Data tools are the future of business insights. If your company does
not change its mindset and is not ready for the speed multiplication that the competition will
take, then dark times await.

Leading companies already understand that Big Data initiatives will be a crucial source of
competitive edge and already have business departments working in close collaboration with
their IT colleagues.

According to Deloitte, by the end of 2012 more than 90 % of the Fortune 500 will likely have
at least some Big Data initiatives under way. Make sure you don’t lag behind!

34

01
‘Analyzing Customer Behavior’
is the third in a series of thought-provoking booklets that Across
Technology will be publishing this year, sponsored by EMC
Greenplum. These booklets are elements in the Data Science Series,
which is also a series of events, and a website:

www.datascienceseries.com

1 2 3 4

Peter Hinssen, editor Peter Hinssen, editor

OPEN DATA

INFORMATION THE AGE OF ANALYZING
IS THE DATA-DRIVEN CUSTOMER POWER
NEW OIL MEDICINE BEHAVIOR SMART CITIES
PREDICTING HOW BIG DATA TURNS EVERY
DRILLING NEW BIG DATA HELPS
WHAT HAPPENS CITY INTO A DATA CAPITAL
SOURCES REVEAL HIDDEN
OF INNOVATION HEALTH TRENDS NEXT
AND BUILD
RISK MODELS

Don’t miss these booklets and events.
Go to www.datascienceseries.com,
register for the newsletter and stay informed.
11/2012 - © Across Technology

SPONSORED BY

036

Analyzing Customer Behavior

More Related Content

Similar to Analyzing Customer Behavior

More from EMC

Recently uploaded

Analyzing Customer Behavior