New segmentation strategies with Big Data

March 2015
A people-based telecom
business
Breathing new life into
segmentation strategies

Contents
Introduction 1
Reach and segmentation in the mobile telecom sector 2
Demographic segmentation as a starting point 3
Reaching mothers at the right time: A moments-based approach 4
Turning the promise into reality 6
Methodology 8
Contacts 9
This report is written by Deloitte MCS in collaboration with Facebook Inc. It utilises Deloitte survey data and anonymised Facebook user
data. The content of this report is not intended to be relied upon and should not be interpreted as specific advice.

Introduction
The principles of segmentation and personalisation – splitting a market or customer base into
segments that behave or purchase differently – are almost as old as commerce itself. While not
performing formal segmentation exercises, the concept was inherent in trading practices of
early merchants – targeting trading partners that place high value on their wares and tailoring
their approach and pricing appropriately.
Though the principles remain true today, the terminology and techniques have evolved considerably. Organisations
have grown to serve increasingly large customer bases due to macro-trends of industrialisation, consumerisation and
globalisation. In response, segmentation has become critical in many sectors and is now practiced routinely by the
professionalised disciplines of strategy and marketing. Typically, practitioners must exercise judgement in balancing
the depth of segmentation and personalisation against customer reach and associated costs of customisation.i
For digital goods, services and marketing the associated costs of customisation are considerably lower. As a result,
digital companies may be able to move closer to the ideal of mass customisation –”producing goods and services to
meet individual customer’s needs with near mass production efficiency”.ii
Internet portals of the 1990s were some of
the first to allow this degree of customisation but the possibility exists for all companies with digital assets.
In an age when so many people are online, there is an opportunity for operators and manufacturers to personalise
experiences in a meaningful, privacy-safe way. And as connection gateways to the digital world, mobile
manufacturers and operators could be some of the first to benefit from the next phase of segmentation’s evolution.
This report explores the role of traditional demographic segmentation and the potential of a moment-based
approach, whereby people are categorised according to who they are and what they are experiencing in life, rather
than characteristics largely determined at birth. We illustrate this through a closer look at the mothers segment,
and highlight the steps operators and manufacturers (OEMs) can take to reach them and seek to increase lifetime
value.
... as
connection
gateways to
the digital
world, mobile
manufacturers
and operators
could be some
of the first
to benefit
from the
next phase of
segmentation’s
evolution.
1A people-based telecom business Breathing new life into segmentation strategies

Reach and segmentation in the mobile
telecom sector
As people increasingly use mobile connectivity in their lives the ability to determine behaviour based on their mobile
usage becomes all the more valuable. Last year, the top 10 operators made $305bn in revenues by providing their
customers with over 2.8bn connections.iii
As a result, Mobile network operators (MNOs) have a significant amount
of customer data at their fingertips, such as age, gender, billing address, data plan and device details. MNOs can
use such data to segment their customer base and thereby develop exceptional user experiences, earning customer
loyalty and continued spend in return.
Segmentation is a fundamental part of corporate, product and marketing strategy. However, it has become all the
more important as markets have become more complex, and digital channels, such as mobile, have rapidly evolved.
People can now choose from a wider range of operators (many of them virtual), bundles of services based around
broadband or entertainment packages, and of course an array of smartphones, tablets and phablets at various price
points.
OEMs and network operators are already investing in retention and acquisition of high value customers. We believe
they would benefit from re-evaluating how they use the wealth of data they possess, and focus on developing more
sophisticated approaches, to better segment, target and reach existing and potential customers.
Unfortunately, following industry expansion and consolidation, these data are often stored on legacy systems and
compartmentalised. Extracting insight retrospectively or in real time can be very difficult. Equally, achieving mass
reach has frequently been favoured over customisation. For this reason, MNO customer segmentation still tends
to use data to confirm a high-level assumption about how consumer behaviour differs by demographic. Such an
approach historically provided relatively little in the way of explaining consumer behaviour, offering segments too
broad to be of any actionable value.
Demographic segmentation aggregates and assimilates people from disparate life stages into a handful of broad
buckets based on shared characteristics. It does not explain why customers behave in the ways they do. In any
large population with normal distribution of behaviours, the segment ‘mean’, or persona, is likely to be a diluted
representation of the majority; and is unlikely to provide a meaningful way to increase relevancy. Importantly,
valuable customers (with high ARPU and low cost to serve) may occupy the same segment as low value customers
due to some shared characteristics.iv
Retention and acquisition efforts applied equally across the segment would
therefore result in over-paying for customers that provide inadequate return, as well as under-investing in retaining
and attracting high-value customers.
Moreover, people change; not necessarily because of an attractive handset deal, but because their life changes.
They leave home, they fall in or out of love, they relocate, have children; the list goes on. These might seem like
characteristics too granular and subjective upon which to base segmentation, but in truth segments need not break
up the entire population, just identify and focus on those likely to change usage or purchasing habits. This is a
moment-based segmentation and the basis of this report.
… segments need not break up the entire population, just identify and
focus on those likely to change usage or purchasing habits.
2

Demographic segmentation as
a starting point
This is not to say that demographic segmentation is without merit. Having a broad view of market segments can help
spot overarching market trends. Take the recent preponderance of large-screen smartphones, for example. In the UK,
92 percent of customers surveyed by Deloitte did not consider a large screen to be important when choosing their
next phone, and this is true across all demographic groups and both sexes. Large screen phone sales have increased
in recent years suggesting that customers are purchasing for other reasons.v
Such insight can be used to focus
strategic, growth-related investment in product development and marketing.
A deeper inspection of sub-groups can reveal differences that significantly impact a business’ ability to tactically
execute in the market. For example, price is often assumed to be a key factor for customers when choosing their next
phone. But younger men – often thought of as having little disposable income and thus being more price sensitive
– were actually less likely than other groups to say it was their key consideration. Even more importantly, 18-24 year
old men in higher socio-economic groups were actually more price sensitive than those in lower groups. Such results
begin to challenge traditional thinking about ‘high value’ customers, in which socio-economic status, tech savviness,
ARPU and customer value are assumed to go hand-in-hand.
The challenge is strengthened further when looking at mobile data use. 18-34 males with a higher socio-economic
status – who we might expect to be constantly on their phones, with a salary to afford it – report using less data
than their lower socio-economic status counterparts, with a lower average usage as well as fewer reporting having
unlimited data tariffs.
There are also significant differences between price insensitive customers and early adopters; two groups that are,
logically, often conflated. Historically, networks have charged a premium for 4G services, although this has eroded
somewhat over time.vi
Though lower socio-economic groups feel less concerned about price, they are also less likely
to be on 4G than their higher socio-economic counterparts.vii
Moreover, affluent 18-24 males are also more likely to
switch to 4G in the next 12 months.viii
Identifying what criteria people value can also serve as a trigger or indicator as to possible churn behaviour. For
example, 16 percent on average highly value internet connectivity, but this rises to 25 percent in higher social-
economic group men aged 35-44. Quality of customer service is highly valued by around 14 percent of people in
general; 22 percent of higher social-economic group men 55-64 found it important. Different parts of the mobile
offering are prioritised differently according to age and socio-economic status.
Isolating and acting upon different values of demographic segments can have demonstrable impact on return on
investment. The ideal segmentation for campaign execution on digital platforms, which now comprise 50% of operator
marketing expenditure, is one that enables decisions to be made in real-time, at a more granular level.ix
For example, a
segment of 18-24, London-based, early adopter, low spender, multiple device owner, high socio-economic group of
men, interested in quad play service. These data exist, but to date marketers have been largely limited to executing
digital advertising programs in the same way as with traditional channels; with segments that are too broad to
achieve cut-through and thus do not deliver the highest possible impact.
Such results
begin to
challenge
traditional
thinking about
‘high value’
customers, in
which socio-
economic
status, tech
savviness, ARPU
and customer
value are
assumed to go
hand-in-hand.
… to date marketers have been largely limited to executing digital
advertising programs in the same way as with traditional channels;
with segments that are too broad to achieve cut-through and thus do
not deliver the highest possible impact.

Reaching mothers at the right time:
A moments-based approach
Parents are an interesting group in this regard. Becoming pregnant, perhaps giving up work for a period of time,
having one or more children, all represent significant life changes. They often accompany other significant events
such as moving house or getting married. Segmenting parents, and mothers in particular, presents an interesting
example of moment-based, people-centric segmentation. Successful targeting depends on real insight into their
needs at different points, in a journey that can take decades.
Deloitte’s survey data show that mothers are 10 ppt (percentage points) more likely to have changed operators
over the last five years than women of the same age without children. On the whole, mothers are less likely to be
early adopters; less likely to own or plan on owning a 4G-capable handset in the near future, deem screen size
less important and have fewer devices than women in general. Despite their relative conservatism, price is a less
important factor for mothers when choosing their next phone. Insights gained from such granular segmentation
suggest vastly different advertising strategies for mothers and non-mothers.
But this isn’t enough to tell us about the moments where change is most likely to occur and thus where products,
services and marketing should be targeted. To get that insight we partnered with Facebook to provide actual
examples of behaviour of mothers, related to their phones.
With 1.39bn monthly active users, Facebook has a bigger population than India.x, xi
This vast resource enables us to
look far beyond binary questions and into individual behaviours. A continuous time series of data allows for isolation
of trends, influences and behaviours in specific groups to build a portrait of comparative behaviour. The results were
immediately apparent, confirming our initial survey findings.
Mothers appear to have different behaviours online than women without children. We found mothers to be
significantly more active on Facebook than non-mothers and possibly more influential. Mothers uploaded more
photos, left more posts and connected to more Facebook friends than non-mothers. For example, while women who
were not mothers posted an average of six photos during a 3-week period in 2014, mothers posted an average of
twelve. And while non-mothers had on average 227 Facebook friends, mothers had 318.xii
Moreover, as an important additional data point, mothers watched more video. This is relevant for telecom
companies for two reasons. First, it indicates that mothers may use more data than the average customer across
cellular and WiFi or wired networks. With a potential interest in higher tariffs, operators may consider expending
more effort retaining these mothers. Second, video advertising is becoming a staple part of reaching customers.
Facebook data implies that mothers, a valuable customer segment, will be receptive to such communication.
Identifying churn patterns is a crucial point for telecoms providers. Currently some of the more sophisticated
operators are trying to develop customer lifecycle models based on length of contract, handset type and broad
segment category. This is an attempt to personalise communications, improving relevancy and customer experience
throughout their relationship. However, even these lifecycle models are prone to broad strokes, which let many
customers fall through the gaps and, worse, can push customers to churn when they receive the wrong type of
communications, at the wrong time.
There are moments in life when everything changes; and people change everything around
them, including the type and volume of product and services they consume. These include
beginning higher education, getting a first job, moving in with a partner, or having a child. It is our
contention that marketing directed to people in these moments can markedly improve return on
investment as it offers a rare opportunity to capture people with high total lifetime value, rather
than habitual churners. Lifetime value models already exist within some mobile operators but we
believe they are likely to be far more effective when combined with moment-based marketing.
4

Deloitte analysis suggests that mothers have churned more frequently than non-mothers; Facebook data indicates
that the age of children appears to correlate with a mothers’ propensity to churn. It suggests that mothers with
young children (0-10 year-olds) are almost 50 per cent more likely to churn than those with older children. There are
likely to be many factors relating to a mothers’ lifestyle that contribute to this, but child-related reasons may range
from upgrading handsets for enhanced photo capability; changing tariffs to suit a different life routine; replacing
handsets damaged by inquisitive toddlers; or even changing operator and handset as an old device is passed down
to her child.xiii
Understanding these root causes and spotting early intention to churn could allow operators to devise
more successful retention strategies.
In the Deloitte survey data, we identified that more active customers, who traditionally might be seen by network
operators as prized customers, also had a higher likelihood to churn. Facebook data seems to indicate a similar
pattern suggesting that networks that wish to keep customers paying the highest tariff will need to work harder to
maintain their loyalty, not least because all their competitors are most likely targeting the same audience. However,
customer value is much harder for operators to quantify.
Sophisticated lifetime value assessment may need to balance ARPU, customer advocacy and influence against
consideration of cost to serve and propensity to churn at various life stages. Other than ARPU, these are all difficult
elements to measure. Therefore any attempts to understand the moments that drive cost, advocacy or churn
can only strengthen operators’ assessment of customer value. Many technology and telecoms companies focus
on traditional early adopting 18-24 males, however we suggest that the overlooked mothers segment may be
exceptionally valuable due to their high level of influence within and outside the family, and potential for advocacy.
Additionally, the major life-changing stages of motherhood may provide a suitable starting place for predicting and
reducing instances of churn.
… networks that wish to keep customers paying the highest tariff will
need to work harder to maintain their loyalty, not least because all their
competitors are most likely targeting the same audience.
… the major life-changing moments of motherhood may provide a
suitable starting place for predicting and reducing instances of churn.

Turning the promise into a reality
There is a difference between a business being content with inertia and one actively encouraging retention.
Adapting to a people-based segmentation, based on life stages, available carrier or third party data, can illustrate
triggers for certain types of behaviour. In other sectors such as gaming, 1,000 segment categorisation of specific
groups is not uncommon, providing agile and profitable approaches to offer excellent experience for customers, by
treating them like individuals. These granular segmentations are used in conjunction with algorithmic approaches to
media and service interventions, to stimulate activity in the right context to inform an execution.
The concept of such data mining might initially make consumers wary and should be handled with care and
transparency. Data providers can anonymise and aggregate data to a level such that buyers of advertising only know
they are targeting a segment more likely to be interested in seeing new offers from them. In this way, it is similar to
the type of segmentation that classifies an individual as likely to be receptive because they are female and
25 years old; but is in fact a good deal less patronising and more accepting of likely personal priorities. The potential
upside for the consumer is receiving communications that are more relevant and interesting to them.
In order for mobile operators to improve segmentation, provide a personalised customer experience and target the
associated business benefits, we suggest several steps for consideration:
There is a
difference
between a
business being
content with
inertia and
one actively
encouraging
retention.
1. Identify loyal customers with high potential total lifetime
value that are most important to retain.
2. Build a deeper understanding of the life changes that
cause these customers to churn.
3. Construct segments based around those life changes and
map the customer’s life journey.
4. Identify in-house and external data along that journey to
enable personalised segmentation
5. Create tailored offerings that bring more value to these
customers.
6. Reach out and tell likely churners about these offerings
via the channels they prefer.
6

Steps 1-3 require operators to interrogate their own data sets, make use of existing lifetime value models, and speak
in depth with their customers. Focus groups and interviews will yield greater insight than the most comprehensive
survey. Operators regularly conduct research by these methods but less frequently focus on people’s life-changing
events.
Step 4 is the critical change in transitioning from demographic segmentation, or lifecycle segmentation, to one of
moment-based segmentation. Moments must be significant and detectable, such as a child starting school. And
leading indicators, in this case childbirth will provide greatest opportunity for the operator to act.
Triggers could be developed and refined that inform the business when a customer is most likely to require an
intervention to discourage churn. Depending on the trigger, different products or services may be more appropriate
and over time operators can learn when to deploy each tailored offering. Triggers may indicate when a “surprise and
delight” call (e.g. free upgrade to a higher data tariff) may be most effective, when a customer might be willingly
upsold (e.g. a new handset), when they may be receptive to additional services (e.g. phone insurance) or when they
might appreciate and take advantage of cross-sell approaches (e.g. family bundles).
Step 5 suggests taking this further as greater customer understanding allows for greater tailoring of products and
services. In order to take full advantage of moment-based segmentation, operators need to have more flexible
pricing, purchasing options and channels, as well as a more sophisticated approach to management that allows for
a more tailored personal communications experience.
Operators may reach their customers through a range of channels, either directly through operator-consumer
channels (e.g. email, online accounts or SMS) or through third party publishers. Step 6 requires operators to assess
which publishers can provide sufficiently granular segmentation of users to be effective. And publishers should be
aware of higher demand for increasingly specific targeting; requiring them to develop capabilities for capturing
segmentation data and being able to make this readily available to those wishing to advertise.
Such a drastic departure from traditional segmentation models conjures up images of major investment and year-
long initiatives to redesign processes and manage change. But this is not necessary; indeed not advisable. In a new
and inherently uncertain venture a small-scale, lean approach offers an efficient way to test hypotheses and learn
critical lessons before scaling. Operators can adopt an experimental approach by identifying just one ‘moment’
relevant to some of their customers and defining a segment around it. Running a pilot using existing offerings
and one publisher will be sufficient to begin learning and will establish a process appropriate for testing additional
moments, tailored offerings and a portfolio of publishers.
An operator with 1,000 segments may not exist today, but the first is most likely to emanate from those that
embrace a people-based segmentation approach and successfully combine it with traditional demographics. These
operators could capitalise on a differentiated level of customer understanding, enhanced ability to serve and retain
those customers, and the ability to earn greater returns for the business. In turn, lessons learnt maintaining the
customer base can be focused on customer acquisition; and those quickest to master the emerging art are well
placed to capture market share growth.
… operators
could capitalise on
a differentiated
level of customer
understanding,
enhanced ability
to serve and
retain those
customers, and
the ability to earn
greater returns for
the business.

The Deloitte data used within this report was collected as part of the Global Mobile Consumer Survey 2014.
Data cited in this report are based on a 22-country online survey of mobile phone users around the world.
All research has been undertaken via online research. Fieldwork took place between May to August 2014.
37,000 responses have been included in the study.
For the US, UK, Norway, Finland, Sweden, Netherlands, France, Germany, Italy, Spain, Australia, Japan, Singapore
and South Korea, respondents were surveyed online and data was weighted to reflect the general population.
The questions for this survey were written by Deloitte member firms. The multinational online research program
was managed by Ipsos MORI and On Device.
Additionally Deloitte conducted analysis on anonymised data for over 50m people using Facebook in eight countries
(France, Germany, Italy, Netherlands, Norway and the United Kingdom). Random sampling was used for statistical
analysis of the data set. No weighting was applied to this data during the process.
Methodology
8

Authors Contacts
Deloitte
Matthew Guest
Director, Digital Strategy
+44 (0) 7786 691970
mguest@deloitte.co.uk
Tom Struthers
Manager, Digital Strategy
+44 (0) 7814 352313
tstruthers@deloitte.co.uk
David Llewelyn-Jones
Senior Consultant, Digital Strategy
+44 (0) 7825 853284
dllewelynjones@deloitte.co.uk
Facebook
Jane Schachtel
Global Head of Mobile and Technology Strategy
+01 818 292 2527
janes@fb.com
Contributors
Robert Dagger and Michael Baines,
Digital Strategy, Deloitte
Deloitte
Royston Seaward
Partner, Digital Strategy
+44 (0) 7788 747232
rseaward@deloitte.co.uk
Facebook
Jane Schachtel
Global Head of Mobile and Technology Strategy
+01 818 292 2527
janes@fb.com

Endnotes
i Segmenting to a lower level allows for more precise personalisation; however each tailored message, product or
service applies to a smaller set of customers.
ii Tseng, M.M. and R. J. Jiao, 1996, “Design for Mass Customization”, Annals of the CIRP, 45/1:153-156
iii Published September 2014. 12-month mobile revenue, Q2 2013 to Q2 2014. https://gsmaintelligence.com/
analysis/2014/09/operator-group-ranking-q2-2014/444/
iv Average Revenue Per User
v Large screen phones or “phablets” are typically defined as smartphones with 5.0-6.9 inch screens
vi http://www.pcpro.co.uk/news/broadband/387832/ees-14-mth-tariff-marks-end-of-4g-price-premium
vii 23 percent, versus 19 percent, respectively
viii 37 percent versus 32 percent, respectively
ix Based on UK figures. http://www.telegraph.co.uk/finance/newsbysector/mediatechnologyandtelecoms/11418000/
Digital-to-account-for-half-of-UK-ad-spend-for-the-first-time.html
x Facebook Inc. Q1 2015 Earnings call http://files.shareholder.com/downloads/AMDA-NJ5DZ/4016806580x0x805520
/2d74edca-e02a-420b-a262-bc096264bb93/FB_Q414EarningsSlides20150128.pdf
xi Published January 15, 2015. Population figure of 1.27bn. http://www.indiaonlinepages.com/population/india-
current-population.html
xii Average number of photos posted and friends calculated as an arithmetic mean, inclusive of outliers
xii In August, 2013 a UK based online and telephone comparison and switching service reported that nearly one in
ten children get their first phone at the age of five. The figure is likely to have risen since then.
10

Notes

New segmentation strategies with Big Data

Deloitte refers to one or more of Deloitte Touche Tohmatsu Limited (“DTTL”), a UK private company limited by guarantee, and its
network of member firms, each of which is a legally separate and independent entity. Please see www.deloitte.co.uk/about for a
detailed description of the legal structure of DTTL and its member firms.
Deloitte MCS Limited is a subsidiary of Deloitte LLP, the United Kingdom member firm of DTTL.
This publication has been written in general terms and therefore cannot be relied on to cover specific situations; application of the
principles set out will depend upon the particular circumstances involved and we recommend that you obtain professional advice
before acting or refraining from acting on any of the contents of this publication. Deloitte MCS Limited would be pleased to advise
readers on how to apply the principles set out in this publication to their specific circumstances. Deloitte MCS Limited accepts no duty
of care or liability for any loss occasioned to any person acting or refraining from action as a result of any material in this publication.
© 2015 Deloitte MCS Limited. All rights reserved.
Registered office: Hill House, 1 Little New Street, London EC4A 3TR, United Kingdom. Registered in England No 3311052.
Designed and produced by The Creative Studio at Deloitte, London. 42158A

New segmentation strategies with Big Data

More Related Content

What's hot (19)

Viewers also liked (16)

Similar to New segmentation strategies with Big Data (20)

More from Amalist Client Services (20)

Recently uploaded (20)

New segmentation strategies with Big Data