1 Making ‘Mass Customization’ a Reality Authors: Chandan Agarwal Social & Business Analytics Lead, Unilever, South Asia Rahul Kishore People Relationship Marketing Lead, Unilever, South Asia
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Making ‘Mass Customization’ a Reality
Authors:
Chandan Agarwal Social & Business Analytics Lead, Unilever, South Asia
Rahul Kishore People Relationship Marketing Lead, Unilever, South Asia
2
Table of Contents
Topic Page No.Abstract 3
Research Objective 4
Introduction to brand Lakmé and Lakmé Fashion Week 5
Social Media Command Centre 7
People Relationship Marketing 8
Persona, a social profiling tool 9
Social Media Command Centre, PRM and Persona creating magic 10
LIVE Social Listening Recommendations during the event 11
Business Results 13
Challenges 13
References 14
3
Abstract
by
Chandan Agarwal, Unilever
Rahul Kishore, Unilever
Big Data is today’s most (ab)used word in the marketing and research industry. Big Data is supposed to
enable marketers to go beyond segmentation and do micro marketing – so that you can customise the
communication (and products/services) at an individual level. But how far are we from realizing it? This
paper outlines how this was achieved for communication using technology to get Single View of
Consumers (SVoC) when information about them comes from different sources and to target them with
customized content.
Significant number of profiles were identified for the brand (both online and offline) and when they were
served the same campaign the engagement rate for them was twice as compared to the Facebook affinity
data, thus providing the proof of concept. This enabled the brand to get higher engagement and access
to the bull’s eye TG which means they could build relationships with their consumers with much higher
ROI than ever before.
4
Research Objective:
Using big data analytics to develop highly targeted audience clusters and provide customised content to
increase reach, and hence prove the concept of Big Data enabling marketers to take a step closer to the
dream of mass customisation.
5
Introduction:
In today’s world, one of the biggest question that marketers are trying to answer is how to use Big Data
analytics to be able to achieve mass customization. This paper aims to demonstrate how we had achieved
mass customization by deploying cutting edge technology and hence improved the ROI of the campaign
significantly be reaching out only to the bull’s eye TG of the brand.
The framework below illustrates how different tools & technology, along with data sources, were used to
achieve the same.
In the sections to follow, there is an introduction to Lakmé and Lakmé Fashion Week along with how social
media listening through a command centre has been helping to achieve a higher reach. Further there is
an introduction to PRM, the in‐house people database and CRM solution of Unilever and one of the tools
(Persona) that was used in the project. Following that, is the methodology of how the project was
executed, the business results and the challenges.
Command Centre‐ LFW 2015 Command Centre‐ LFW 2016
6
About Lakmé:
Lakmé continuously innovates to offer a wide range of high performance colour cosmetics,
skincare products, and beauty salons.
Combining international cosmetic technology with an in‐depth understanding of the Indian
woman’s needs, Lakmé offers its consumers a comprehensive beauty experience through its
products and services.
About Lakmé Fashion Week:
Lakmé Fashion Week (LFW) is jointly organized by Lakmé, the No.1 cosmetics and beauty services
brand in India and IMG Reliance, the global leader in sports and entertainment event marketing
and management.
LFW has been conceived and created with a vision to “Redefine the future of fashion and Integrate
India into the global fashion world”.
Two seasons of Lakmé Fashion Week are organized every year – ‘Summer Resort’ & ‘Winter
Festive’.
Lakmé strives to create Brand Love by being at forefront of fashion in India and also launching
their super‐premium product range during the event.
15 years of Lakmé Fashion Week:
Lakmé Fashion Week Summer/Resort 2015 marked the completion of fifteen glorious years of
fashion.
The anniversary was a landmark depicting achievements, memories and magnification of Indian
fashion which is currently at par with the global counterparts.
7
Command Centre for Social Media Analytics
A social media command centre is a dedicated area where a company can monitor and engage in social
conversation around its brand and market.
The enabling setup contains multiple LIVE screens for real‐time monitoring of social media trends backed
by analytics that can help:
bring together brand marketing teams, data scientists, media buyers and creative agencies
strategise content development and release during the event
promote fresh & engaging content based on the sentiment of netizens, and closely monitor the
performance of content deployed during the event
give executives a glance at the social health of the brand
stay connected with the consumers in real time
serve as the platform for crisis communication plan
Layout of a Command centre
8
People Relationship Marketing (PRM)
Unilever leadership recognizes that data is the goldmine that, if mined well, can be our biggest asset to
tap into the connected world of new age consumer. Our pursuit is to put data at the heart of our marketing
initiatives. Our Data Centre and Relationship Marketing initiatives are designed to create owned
audiences and personalized communication strategies for Unilever by capturing and analyzing people data
from all owned digital marketing channels. PRM is a global CMI initiative that enables the business to start
maximizing all data sources, through a mix of technology capability and analytical/marketing expert
support which provides:
Secure and privacy compliant data warehouse where all local data can be consolidated and
captured to give total visibility
Team of analysts to brief in and answer any brand or business question using PRM consumer and
campaign data, and build repeatable models to deepen understanding of consumers
Capability to deliver accurate, sharper segmented campaign data for Marketing teams to better
target and engage consumers
Further capability to optimize channel selection, creatives and incentives thereby enhancing
campaign performance and consumer engagement
Unilever PRM architecture is powered by our Technology Services and is built on a robust Big Data
infrastructure. We use sophisticated data storage and warehousing platforms from Amazon Web Services
with analytical layer supported by Teradata and Hadoop. We are extremely sensitive about the security
of our consumer PII data and AWS S3 provides us a highly secure technology for data exchange. Our data
governance principles make sure that we are 100% compliant with Unilever data protection guidelines as
well as country regulations.
India, by sheer population strength is one of the largest PRM databases within Unilever. We have close to
130 Mn consumer records within India PRM. Apart from contact information we look to capture additional
information like e‐mail address, city, age, gender, brand interactions etc.
In India, unlike other markets, mobile has proven to be the most effective medium of contact vs. e‐mail.
As a result, most of our re‐marketing campaigns have been mobile led & now foraying into the digital
world. However, with the explosion of internet & digital media in the country, digital campaigns have
started taking precedence. Using PRM data, it has been observed that PRM audiences on social media
have always shown greater engagement than any affinity audiences provided by publishers.
The post‐campaign response data also flows back into PRM. This data leads to further enrichment of
consumer profiles as well as enables analytics & modelling to derive insights on campaign execution,
segmentation & targeting.
9
Persona‐ A trademark social profiling tool
We, at HUL, partnered with a market intelligence agency (BlueOcean) to help us with profiles of audiences
who converse about our brands on social media. For this purpose, a social profiling tool called Persona
was used.
Persona is an outside‐in deep learning customer segmentation and insights product. It provides a single
view of the customer by bringing in the first party data from clients infrastructure; and further enriching
it with data from third party digital sources. Third party data enables insights in behaviours, preferences,
attitudes and psychographic information about an individual or a segment. Persona utilizes artificial
intelligence techniques for actionable intelligence and insights at an individual level.
An advanced identity resolution engine powers the identification and differentiation of each individual
customer, which enables Persona to “stitch” both online and offline attributes of the customers. Multi‐
dimensional variables like psychographic, experiential, behavioural, affinity and engagement data is
algorithmically analysed and mapped, for each individual entity. Subsequently, the taxonomies are also
implemented at an individual level to identify segments from digital body language.
Persona differs from other enterprise segmentation tools by its modus operandi of capturing intent data
and behavioural signals from across the digital entities and marrying them intelligently with internal data
sources.
10
Social Media Command Centre, PRM and Persona creating magic
In 2016, HUL identified that enhancing reach on social media does not necessarily helps in building brand
equity or increasing sales unless the activation is done on right audience groups. Using Facebook, helps in
reaching out to a wider audience however the ‘spray and pray’ approach is not ROI efficient.
Thus, at HUL, we wanted to find a way where we can use social media effectively and reach out to the
Bulls Eye TG and get highest ROI.
This is where PRM and Persona were incorporated in Social Media Command Centre to identify the right
audience – ‘makeup junkies’.
Identifying the right audience to be targeted – pre‐event
The identification of target consumers was done by a two‐fold strategy:
1. Top Down
2. Bottom Up
Using top‐down strategy, we listened to people on social media using our social listening tool
and scanned their social media accounts, profiles, interest areas, etc. (all publically available
information only) and segmented audience profiles. People who had engaged in fashion
conversations in the last 12 months were profiled. Using Persona we were able to identify all
social accounts and interest areas of the audiences. This helped us get access to around 300K
people who were interested in fashion events.
Further, using bottom‐up profiling, we used the existing database of users of different brands
and filtered them on various parameters like life style measures. This is where we used our in‐
house people database, People Relationship Marketing. From PRM we got access to around 2.4
million people who had visited Lakmé Salon in the last 12 months. However, we realized the most
of these people had visited Lakmé Salon for basic hair cutting service only. So, removing them,
we identified 800K people who had taken higher order services in Lakmé Salon.
Further, using Persona we identified their social profiles and interest areas which enabled us to
refine the audience understanding. Stitching the social profiles and first party data together
ensured we had a Single View of our Consumers (SVoC) and hence the right audience cluster.
We finally identified 220k profiles to be targeted on Facebook and Instagram.
11
LIVE Social Listening Recommendations during the event
While SVoC provided us a database of bull’s eye TG to be used for LFW campaign, we still needed to
optimize our creatives and communication to ensure maximum impact.
This was achieved through our listening capabilities deployed live on social media channels (Twitter,
Facebook and Instagram) by setting up the Command Centre. The insights derived from this data would
help the brand to react to the trends at a near‐real time basis.
Insights and Action:
During the event, command centre utilised social media monitoring and analytics to provide real‐time
recommendations to the collaborating teams.
Makeup trends: In order to appeal to the target audience, popular trends and hashtags of different
makeup categories were identified. These hashtags were coupled along with the hashtags of Lakmé
Fashion Week to increase the reach of content multiple times.
1> Among Lipsticks, Red Lips were trending the most in discussions on social media, so creatives were
released on Red‐lip look of celebrities.
2> Around Lips, most trending hashtags (#lips – 19.6Mn, #lipgloss – 1.8Mn, #lipstick – 13.5Mn) were
suggested, which were then used in all subsequent posts to reach out to maximum people.
12
3> Around Nails, it was suggested that more content be released around Nail Art (#NailArt – 24.6Mn,
#Nailswag – 6.2Mn, #Nailstagram – 4.9 Mn) as it gets tagged with most trending discussions
around nails.
4> Around Eyes, it was suggested that the most trending hashtags (#eyeliner – 6.1Mn, #Eyeshadow
– 5.6Mn, #Eyemakeup – 1.4Mn), which were used in all subsequent posts to reach out to
maximum people.
5> Most trending General hashtags around makeup like (#Beauty – 128Mn, #DIY – 18.7Mn,
#MakeUp – 91.9Mn, #MakeUpArtist – 18.5Mn) were also suggested, which were immediately
used for all subsequent posts for maximum reach
6> The discussions on Facebook suggested that the Smokey Eyes effect was trending and it was
suggested to the creative team to come out with relevant content on this for Lakmé brand which
will appeal to the makeup junkies.
Celebrity/Designer: To benefit from celebrity presence at the Lakmé Fashion Week and the work of top
designers, the command centre team suggested the following actions
1> Since the hashtags for top designers of the day were trending the most, it was suggested that they
are clubbed with all relevant content coming on digital media for that day
2> To reach all major fan‐groups of celebrities and increase overall reach of the campaign, it was
suggested to tag the Fan‐pages on the pictures
13
Business Results
Using big data analytics enabled Unilever to do ‘mass customization’ for the first time.
The key results are as follows:
1. The audiences reached through SVoC approach had more than twice the engagement rate as
against that of Facebook affinity data.
a. Further, due to its high engagement rate, we also reached out to PRM look alike audiences
(170K audiences) on Facebook and even they had twice the engagement rate
2. Purchase Intent of Lakmé for these audience groups went up by more than 4 times as compared
to the previous event.
3. This has also led to acquisition of 200k+ audiences who are interested in fashion and with whom
we can continue to build relationships and thus grow brand equity of Lakmé.
The project proves that we can use analytics to reach out to specific audience group with a high ROI and
not rely on ‘spray and pray’ approach.
Challenges:
Currently we have two challenges before we scale this up further:
1. Lakmé is a fashion brand and people like to talk about fashion on social media. The same might
not be true for other categories like laundry
2. Though we identified 200K+ audiences with whom we can create relationship but for FMCG
products in a country like India 200K audiences is not enough scale.
14
References:
1. BlueOcean – Audience Profiling Partner
2. Capgemini – Social Analytics Partner
3. Gain Theory – Data Strategy Partner
Related Documents
