2. Analytics have become one of the most powerful tools Marketing Mix Modeling
available to retailers, and are being used to enable fact-
based, insight-driven decision making to manage their
Problem Definition
strategic, operating and financial performance, and create
shareholder value. Retailers today are searching for ways The task of measuring returns on the marketing mix has
to derive more customer intelligence and operational become somewhat more complex as media have
insights from their overflowing databases proliferated, and as alternative explanations of marketing
lift have to be ruled out.
The Opportunity and the Challenge “Half of the money I spend on advertising is wasted; the
trouble is I don't know which half.” The old complaint, no
Today, there is a compelling need to provide the longer has to be true - retailers can find out whether
right information, at the right time, to the right advertising works, and how it compares to other
decision makers, using the right technology. marketing tools.
Solution
The aggressive adoption and exploitation of analytics has
led to competitive advantage among some of the world's HCL leverages Marketing mix modeling (MMM) to unearth
most successful retailers. Retail Analytics can collect, the driving forces in the marketing environment in order to
process and analyze a wide variety of data on retail stay profitable.
performance.
MMM defines the effectiveness of each of the marketing
elements in terms of its contribution to sales-volume,
The Breadth of Analytical Options for a effectiveness (volume generated by each unit of effort),
smarter enterprise efficiency (sales volume generated divided by cost) and
ROI. These learnings are then adopted to adjust
Leading retail executives believe they can achieve true marketing tactics and strategies, optimize the marketing
competitive advantage with retail analytics technology by plan and also to forecast sales while simulating various
using enterprise-wide approach that involves product, scenarios.
customer and functional boundaries. That is the reason
why Analytics is the centre of focus for any profitable It helps an organization's efforts to measure the change in
enterprise today. sales and attribution of the change to various marketing
mix elements such as Trade, TV, FSI, Print amongst
Retailers recognize analytics as key to business others.
transformation and competitive advantage
Benefits
To sustain and lead in a complex and constrained
marketplace, companies have to resort to mathematical Being able to know what investment works for the brands
tools and techniques to make informed choices from will potentially save a lot of marketing dollars. Experience
terabytes of data available. HCL can customize analytics shows by deploying the models, 7-10% of promotion
offerings to retailers to enable them to do their business dollars can be saved or reallocated to more hard working
profitably like never before. marketing buckets.
Based on analysis of what all business problems retailers
are plagued with today, HCL has come up with a set of
often-asked problem definitions, and attempted to map
against those what could be palpable solutions to the
defined problems. HCL is well equipped to work with the
retailers and implement these solutions.
3. ANALYTICS SERVICE OFFERINGS
Targeting
?
Frequently Asked Customer Segmentation Analysis
?
Customer Loyalty Analysis
?
Questions Customer Analytics
?Basket Analysis
Market
?I target precisely and
How do
Marketing Mix Modeling
?
customize offerings? Marketing Spend Advertising effectiveness
?
?I customize offers to stores?
How do Optimization Awareness planning
?
Service Offerings
?I justify my marketing spends?
How do ? attribute drivers
Product
?I optimally price throughout the
How do Assortment Replenishment analysis
?
Optimization Quantifying ‘lost sales’
?
product lifecycle? Pack size simulation
?
?I decide on my assortment
How do Optimization
?
composition so that I can minimize lost Price Analytics Price/Cross Elasticity Modeling
?
sales? Item Cannibalization Analysis
?
? Simulation
Pricing
?I optimize
How do ? Scenario Modeling
What-If
Inventory Analytics
inventory/replenishment and ? Management performance
Inventory
transportation costs? Analyzing Inventory Replenishment Policies
?
Evaluating optimal quantity to order to
?
?I manage demand?
How do Test & Learn minimize total variable costs
?I know whether my retail
How do
? store renovation prospect
Test out
innovation will work? ? effect of any store intervention on
Finding
sales and profits
ISOLATING FACTORS THAT ARE IMPACTING
INCREMENTAL AND BASE SALES
MODELING SALES
BASE VARIABLES INCREMENTAL VARIABLES
90.56 * Distribution 6168.18 * Trade
-686.35 * Price 943.95 * Print
SALES =
32.43 * Seasonality 5187.29 * TV
-0.59 * Competition 0.08 * Radio
SALES DECOMPOSITION BY DRIVERS
INCREMENTAL
TV
?
Print
?
Radio
? MEDIA BASE MARKETING MIX MODELS
Trade
?
Outdoor,
?
Price
? OUTPUTS
Distribution
?
etc.
Competition
?
Elasticity
?
TRADE Long-term
? ? each marketing vehicle
ROI of
impact from Marketing activity which drives volumes
?
marketing Marketing activity driving the consumer
?
OUTDOOR behavior
? of operational factors
Impact
? planning
Financial
4. Customer Analytics - Propensity to buy Some of the questions answered are:
Modeling/Response Modeling 1. What impact will the program have on key
performance indicators if executed across the network
Problem Definition or customer base?
Decision makers today believe that getting a clear view of 2. Will the program have a larger impact on some
customer preferences and customer behavior with stores/customers than others?
effective Predictive Analytics and Data Mining tools, to
identify the customers with the highest propensity to buy 3. Which components of the idea are actually working?
new products and services, is imperative for accurate and
better customer segmentation. Retailers are striving hard Solution
to personalize offers and hence need to identify the
targets very precisely. HCL can help in carrying out test on smaller set of
subjects (stores or customers) and results can be
Solution extrapolated to the entire population.
Business intelligence technology that produces a Benefits:
predictive score for each customer or prospect hence
targets the most likely prospects of a marketing Feasibility study at a low cost
?
campaign. The offers can be based on extrapolating from
? provide an ongoing test laboratory facility for
HCL can
past behavior in an ad-hoc manner but a more scientific
various tests to be carried out – the results can be
way to target would-be customers is to put probability
used for future reference when similar innovations are
scores to each customer from a customer base of millions
envisaged
and choose the most probable ones.
Benefits
Selection of best target customer base for customer
?
acquisition campaigns
? higher response rate and reduce marketing
Achieve
Cost
Maximizing ROMI (Return on Marketing Investment) on
?
campaigns
Understand the demographics of specific product
?
(category/brand) buyers and design promotions
accordingly
Test and Learn for Stores and
Customers
Problem Definition
Test and Learn is followed by retailers for randomized
testing, to test ideas in a small number of locations or
customers to predict impact of retail innovations on a
small scale.
Large retailers with multiple stores are uniquely positioned
to employ the “test and learn” analytical approach, in
which a relatively small sample of stores is used to test
whether a particular change or intervention delivers the
desired result.
5. PROPENSITY TO BUY MODELING – PROCESS FLOW
Scope to increase category base Profile Modeling
Sizing the opportunity buyers and non-buyers
? variables
Candidate
? variables
Significant
Combine Data sources ? type (Step wise regression etc)
Modeling
Analysis sample creation Create Analysis sample
Model performance
?
Create Training and validation set
? custome
Scoring
Targeting strategy
Combination of business unit buyers
Buyer definition $ value of purchases
Time frame of buying
TEST AND LEARN - UNDERLYING PRINCIPLE
Test and Learn TEST AND CONTROL METHODOLOGY
Test and Learn is a set of
techniques carried out by retailers PRE TEST POST TEST
and consumer-focused
companies to test out hypothesis
Locations/groups where Amount of change in
that holds business promise.
test is performed test group
Tests are carried out on a smaller
set of subjects (stores or
customers) and results Control group – no Amount of change in
extrapolated to the entire activity performed control group
population. These techniques
essentially are built on the
foundation of ‘Design of NET CHANGE : TEST VS. CONTROL
Experiment’ theory in Statistics
RESULT BASED ACTION TEMPLATE
RESULTS ACTION
Result:
> Goal Rollout
?
Sales Lift > 5%
Result based Result:
= Goal Re-test
? the same hypothesis
Actions Sales Lift = 0~5%
Result:
< Goal Stop/Re-design
? the test
Sales Lift < 0%
6. Pre Pack Optimization Market Basket Analysis
Problem Definition Objective
Fierce competition in today's global markets, the With the introduction of electronic POS data, retailers have
introduction of products with shorter and shorter life cycles at their own disposal an incredible amount of data. MBA
and the heightened expectations of customers have forced (Market Basket Analysis) is one tool that leverages this data
business enterprises to quickly respond to customer needs and helps retailers understand underlying hidden pattern in
without an increase in costs. This has necessitated the customer transaction and use that information profitably.
streamlining of business processes by cutting down
unwanted activities that add to cost and lead time across Solution
supply chain.
Market Basket Analysis unfurls the science that goes
Solution behind why certain products are bought together.
Business enterprises are using more and more of The application of market basket analysis is generally
information technology tools to optimize their supply facilitated by the use of the data mining tools. HCL Analytics
chain. One such tool is Pre-pack Optimization which has the expertise to take the regular MBA analysis to the
takes a system-wide perspective in identifying and next level by identifying the most profitable baskets,
reducing cost and thereby enhancing the supply chain differentiate between the natural rules that are inherent to
profitability. the stores and patterns induced by promotion, normalize
the effect of store attributes like size, footfalls etc to sales
HCL has devised an innovative process to address this and margins and compare stores.
issue by treading a middle path between optimization of
'predetermined ratio of sizes combination' and 'pick a Benefits
pack' options.
In retail, affinity analysis can be used for purposes of cross-
HCL has undertaken broad steps to follow in order to selling and up-selling, in addition to influencing sales
identify optimum size ratios, framework to quantify promotions, loyalty programs, store design, and discount
'unnecessary replenishment”, calculate percentage of plans.
units sold on clearance, zero in on stores with highest
clearance sales and which had the biggest negative
impact on margins, find ways to reallocate the
unprofitable units to stores that had stock outs, develop
and continuously refine size skew customer look-alike
model(s).
Benefits
? optimization takes optimization beyond
Pre-pack
the retail enterprise to improve efficiencies of its
vendors and the retailer's stores
Retail planners and replenishment analysts can rely
?
on an automated and sophisticated solution to
determine pre-pack configurations
Boost the ability to get the right sizes to each store
?
while decreasing the amount of excess, fringe-sized
assortment
? to create a lost sales model and incorporate
It helps
into current processes
7. PRE PACK OPTIMIZATION
Predetermined Ration of Size
Combinations with a Pack
Is there a third option that is more viable?
Create and manage a few more packs, (Pack = Size)
Automate decision process
Pick-a-Pack
MARKET BASKET ANALYSIS
Market Basket Analysis analyses customer buying habits by finding associations
and correlations between the different items that customers place in their
“shopping basket”
TRANSACTIONAL DATA ANALYTIC ENGINE
Milk, Milk, Milk, Milk,
Eggs, Eggs, Eggs, Eggs,
Sugar, Sugar, Sugar Sugar
Bread Cereals
Bread
Customer 1 Customer 2 Customer 3 Customer 4
Market Basket Analysis
X ===> Y
RESULTS
SL BSKcount LBSK RBSK Total Support Confidence Expected Lift Median Median
(X,Y) Basket Confidence S S Support
?
X(Rules Y(Rules
1 RC L R T S C E L Basket) Basket) Confidence
?
2 =RC/T =RC/L =R/T =C/
3 EC Association rule
?
4
5
discovery (Simple and
6 complex rules)
7
8
9
10
8. Hello, I’m from HCL. We work behind the scenes, helping our customers to shift paradigms and start revolutions. We use digital engineering to build
superhuman capabilities. We make sure that the rate of progress far exceeds the price. And right now, 90,000 of us bright sparks are busy
developing solutions for 500 customers in 31 countries across the world.
How can I help you?
www.hcltech.com
For more information contact us at: retail.solutions@hcl.com