Smart   Consulting


SMART RETAIL SOLUTION

Analytics Solutions for Retailers Worldwide

What is SMART RETAIL SOLUTION?

Smart Retail Solution is a suite of modular analytics products for retailers worldwide.

Recommendation Engine

Customer analytics helps to provide personalized offerings, services and promotions to each customer.
Data Mining algorithms helped the UAE based ratail chain to know their customer better.

    Customer analytics enable retailers to do targeted marketing

  • The right product for a customer.
  • The right time to make an offer.
  • Which customers are likely to return purchase
  • What is the best offer to each customer
  • Many More.....

Collaborative Filtering

    Score the affinity of customers for SKUs based on

  • Quantity of purchase
  • Recency

    LDetermine a similarity score between consumers

  • Cosine Score

    Cluster users according to similarity

  • Recommend to consumers what other consumers in their neighbourhood are consuming.
  • Target pregnancy case

    SKU based approach

  • Group together similar SKUs, i.e. those that tend to be purchased together
  • Market basket analysis
  • Recommend to consumers items similar to those they bought

Pricing & Promotion Optimization

An analytics solution was developed for Fmart to manage the pricing and promotion strategy
  • Which are the profitable lines ? The best and worst customer ?
  • How much money are you loosing through inappropriate rebates?
  • When should promos be offered ? What is likely impact ?
  • Price products based on demands and inventory
  • Pricing of new products and variation of existing products
  • What type of promotions to be offered to earn maximum revenue ?

Demand Forecasting

    Demard is a function of

  • Price - log (Dmd)) = α + β (Price)
  • Seasonality
  • Special events
  • Tread
  • Promotions

    Methods | Guidelines for use

  • Moving Average
    More or less constant level, No trend or seasonality
  • Exponential Smoothing
    Slowly fluctuating level, No trend or seasonality
  • Holt's Method
    Trend (possibly slowly changing)
  • Holt Winter's Method
    Trend and seasonality
  • linear Regression
    Causal variables affect demand more or less linearly
  • Log linear Regression
    Demand varies with price in a non-linear way
  • Croston's Method
    Erratic or lrregular demand
  • Arima
    Sophisticated method only for experts

Unconstrained Demand

    Out of stock

    Distinction between demand & sales

  • Demand - what customers are willing to buy
  • Sales is constrained by available stocks, Demand is not
  • Essential to forecast unconstrained demand correctly to avoid stock outs which apart from being an opportunity loss can help competitors to penetrate your market.

Promotion Optimization

    Promotions undoubtedly increase sales

  • However there should not be too many promotions of any SKU.
  • Promotion fatigue - sales decline immediately after a promotion because of stockpiling.
  • Dilution of store / brand image due to too many promotions.
  • It is also necessary to choose right discount percentages.

Inventory Management

A Business intelligence solution developed for Fmart to ensure

    Right Product at the right place at the right time

  • When to re-order ?
  • How much to re-order ?
  • How much safety buffer to keep to avoid stock out ?
  • Allow for uncertainty in demand and lead times.
  • Manage items with different demand patterns (for exmaple : intermittent sales)

Fraud Analytics

    We developed data analytics technology to prevent and detect fraud and theft

  • It analysis the stock movements,sales & purchases data very carefully to identify possible fraud.
  • It analyze the claims and refund to identify possible frauds
  • It analysis PO, Invoices and payments to identify possible frauds.
    • Fraudsters are everywhere !!

Benford Analysis

Digit Position in number
1st 2st 3st
0 NA .11968 .10178
1 .30103 .11389 .10138
2 .17609 .10882 .10097
3 .12494 .10443 .10057
4 .12494 .10443 .10057
5 .07918 .09668 .09979
6 .06695 .09337 .09940
7 .05799 .09035 .09902
8 .05155 .08757 .09864
9 .04576 .08500 .09827
Benford's Law
We show below a sample report for benford's two digit test.
First Two Digit Count Actual Proportions Expected Proportion Difference Absolute Difference Z-Statistic
10 10439 0.058724256 0.041 0.01772426 0.0177243 37.6806645
11 10306 0.057976069 0.038 0.01997607 0.0199761 44.0443526
12 5966 0.03356154 0.035 -0.0014385 0.0014385 3.2935997
13 5756 0.032380192 0.032 0.00038019 0.0003802 0.90403573
14 4462 0.025100837 0.03 -0.0048992 0.0048992 12.1017092
15 5591 0.031451989 0.028 0.00345199 0.003452 8.81503382
16 4391 0.024701428 0.026 -0.0012986 0.0012986 3.43303913

Expected Proportion - Expected to follow Benford's Law

Z-Statistic - Statistic to check if the distribution of each digit conforms to Benford's Law

Chi-Square Statistic & K-S Statistic mean Absolute Deviation - Statistic to check if overall distribution of digit conforms to Benford's Law

Loyalty Programs Objective & Implementation

Objective of Loyalty Programs
  • Customer Acquisition
  • Customer Retention
  • Increased Spend
  • Customer Engagement
  • Data Acquisition
  • Personalized Recommendations & Promotions

Loyalty Programs Timeline

    Initiation

  • Set Strategic Objectives
  • Choose Technology
  • Setup Administrative
  • Decide Rewards Scheme & Tiers
  • Test runs

    Launch

  • Enrolment
  • Awareness Building
  • Coupons
  • Direct Mailers
  • Mobile & Online Marketing

    Maintenance

  • Acquire and Validate Data on an Ongoing Basis
  • Use data for Analytics
  • Tweak Program as Necessary
  • Link with Credit Card
  • E-Commerce

Loyalty Programs Analytics

  • It is important to identify customer preference while determining product assortment.
  • Sometimes stores decide to eliminate low frequency items.
  • However these items may be bought exclusively by a certain number of buyer who also buy other items.
  • The elimination of low frequency items may drive away such buyers to other stores.

Loyalty Programs Rewards Structure

  • Value Club Membership
  • Earn 1 V point for every dollar spent at any store
  • 1 V point = $0.01 rebate
  • 2500 V point = 2 - Year membership renewal
  • Immediate redemption at any store
  • May be good for customer acquisition and retention but does it encourage member to increase spend ?
  • Different into Multiple tiers
  • Higher Tiers on Accumulating Grerter Number of Miles
  • More privileges for higher tiers (Extra Mileage, Lounge Acess, Priority Check in, Extra Baggage etc.
  • Encourages Addiitional Spending