How to Identify Right Shopper for Promotions / How to Categorize Shoppers based on Loyalty

Identify Profitable Shoppers using Statistical Modeling

Not all shoppers are profitable - the statement sounds bit confusing. But a careful analysis shows that there are some who buy only during Mark Downs, select low margin products in discounts etc. In the previous posts What are the steps in NBO (Next Best Offer) to give Personalized Promotion and Find Right Channel we get an idea of how to select a right product for promotion and what would be the appropriate time for sending the promotion/offer

Well before attaching an offer to the customer, one needs to select if the customer is profitable for the loyalty program

There are many methods to check if the customer is profitable. This is is based on the past transaction history and applying methods / models on them

CLV and NPV are some techniques to use in the current scenario. Segmentation of customers based on the above would provide a good insights

This is a challenge for many US Retailers whose customers enroll in multiple loyalty programs (Average US household enrolls to 23 loyalty programs) and hip hop to other retailer based on Sale/Discount. Price plays an important role in this

According to a study by +McKinsey on Marketing & Sales targeting High Value customers is one of the way to improve the Basket Size (both items and Value). One of every retailer's goal should be to Increase the Basket Size by appropriate promotions

How to Get Sales Leads from Analyzing Web Data

Web Pattern Analysis for Identifying Prospects and Products

Web is an abundant source of information. The trails left behind by online shoppers / browsers are as part of Web Log, Click Stream data etc.

This data along with Social Media is fed to the Analytical model which analyses the trend and identifies potential customers / products that can be promoted

There are many companies /retailers that are using Special algorithms / tools to crawl the web and derive useful insights.


Some specialized companies like Palo Alto based Infer (https://www.infer.com/) offer an analysis software that can be linked to CRM system like SalesForce (http://www.salesforce.com/)a and rank the customer



Infer's ranking system uses more than 150 signals , which includes many data feeds - Census, Job Boards, Tweets , Facebook Likes and Comments



There are many algorithms that help in analysing the Facebook data

OpenGraph API is one of them


 See Also:

How to Segment Social Media Users (Influencers / Detractors / Recommenders)

How to measure Price Sensitity of Customers

Segmentation Standards / Framework

How to use OLAP for interpreting Cluster Results

How to Intepret Clustering Results / K-means Cluster output using Data Warehouse

OLAP and Clustering ? Strange combination isn't it? OLAP - we know what to do, what fields to aggregate, slice and  dice, what report to show and what fields to show. Clustering on the other hand is looking for groups (with or without a particular goal ). How OLAP can then help Clustering. Please Read on. You might be convinced to try

K-means (or any other method) clustering partitions the data into groups that are homogeneous with respected to the cluster variables.

The output mostly would be the primary variable, and its associated cluster along with other cluster / segmentation variables

Cluster Profiling / Interpretation

Though cluster output and their scatter plot provides some interesting findings, it's the cluster profiling that helps the user in a better way

OLAP and Data Warehousing

OLAP  is an BI (Business Intelligence) approach to answering multi-dimensional analytical queries in real or near-real time

OLAP consists of Facts, Fact Tables, Dimension tables, Cubes that contains the precise data for a given requirement etc.

OLAP and Clustering

If the cluster output along with the interpretation variables and the cluster score (index) is fed to the OLAP and analyzed, it will provide a faster way of Cluster Interpretation and through many interesting insights

In similar way decision tree results can also be fed back to OLAP (+Oracle , +SQL Server , +Terradata AG )

Widely used Segmentation Products / Solutions

Free Segmentation Frameworks / Products

We had an indepth analysis of various segmentation frameworks for profiling customers in Segmentation Standards / Framework. Here we will look in more detail on some readily available Segmentation products that will accelerate segmentation/clustering.



Prizm's market segmentation comes-up with beautiful segments for a particular Zipcode. If you want to open a Store in the locality ZIP Code 65231, Auxvasse, MO. Here are the results from the search (http://www.claritas.com/MyBestSegments/Default.jsp?ID=20#)

PersonicX is a household-level segmentation system including 70 clusters and 21 life stages. The PersonicX suite helps to know and anticipate customers’ demographics and buying behaviors, conduct market analysis, plan customer acquisition strategies, and create cross-sell, up-sell and retention campaigns that are truly targeted, personalized and powerful.





Forrester's Social Technographics data classifies consumers into seven overlapping levels of social technology participation. Based on their proprietary Consumer Technographics survey data, they can
share with how social participation varies among your consumers globally and help plan a targeted social technology strategy.

+Monetate's LivePredict is an automated segment discovery product that automatically identifies valuable customer segments and the attributes that define them. This gives marketers the ability to
instantly take action by targeting offers and content specifically to those segments

+OfficeMax is using LivePredict to identify highest and lowest performing segments at both brand and campaign levels

Customer Analytics - Relative Price as Data Point

Relative Pricing in Customer Behaviour Analytics

We had some discussion on Price Sensitivity earlier. This was used for determining the Right Price of offer for that customer. However, the price sensitivity is more related to Demand Elasticity

Relative Price is the price of a product/service with respect to another product/service. The latter one is usually termed as the base product/service.

How to use Relative Price to understand Customer Behaviour

Relative price provides very much useful insight about the customer's behaviour. The following would be the typical values for Relative price

• High
• Medium
• Low

The relative price is applied to each item based on a base item from that category. For example a segregating a customer based on the
different price preferences (Twinnings Tea might be a High valued one, while Tetley might be Medium priced; depending on the purchase behavior of each individual across category the following Relative Pricing Table is created)

Cust ID	Category	Premium	High	Medium	Low
RF0000123	OVGT120980	0	6	16	0
RF0000227	OVGT120980	11	3	2	0

Relative Pricing Customer Table

)Looking at the above table, one could distinguish between the customers - (for a particular category) - the second customer mostly shops for Premium items, while the former is a bit conservative.

The above methodology can be used and the following can be used as Segmentation variables

• No of High Value Items
• No of Medium Value items
• No of Low Value items

Please try this out and let us know the difference it brought to your segments

How to measure effectiveness of Switcher Campaigns in Retail 

Switcher campaigns are the one that disrupt brand loyalty and tease an otherwise loyal customer to switch over to the competitor. Tetley and Brooke Bond are competitors - what would be the impact of providing an offer on Brooke Bond tea for a Tetley fan.

dunnhumby's analysis on offers finds Switcher campaigns are always the worst performing. According to the reporthouseholds tend to buy more of their favorites, such as Coke products, or a new Coke product, if offered an incentive rather moving to Pepsi for example

Case Study: Customer Analytics for Telecom Operator to Cross-Sell and Up-Sell

How to find products/services to Cross-Sell / Up-Sell using Customer analytics / NBO

The Telecom operator has a good customer base and wants to drive more sales using existing customers. The operator decided to craft a NBO strategy for increasing sales from existing customers (this is not just retaining the customer - by reducing Churn)

As with any NBO - the first step is to determine the Objective

Step 1: Define The Objective

The objective here is - Increase sales from existing customer through cross-sell or up-sell

A customer with two or more services is four times less likely to Churn than a customer with a singl service - this means that cross-sell not only increases the incremental sales but also increases Loyalty. The strategic objectives like these should be driving the NBO objective

With the goal being defined very clearly, the next step is to collect required data for the NBO

Step 2: Collect Data

The Telecom operator has data on the services/products that are used by each customer, the plan, the usage history, subscription data,  the CDRs , Customer Demographics data and Internet usage data (for the subscribers) . This data holds wealth of information and needs to be mined.

The entry channel (the channel through which the customer was acquired) would be a vital information.

The Location data also plays an important role in understanding the customer, Location history would explain if the customer is a frequent traveller, when and where he/she makes uses the service.

If the internal data collection is done, one can look for Syndicate data or External Data (Payment Card data)

Step 3: Analyze, Model  and Execute

Once the Data is available, it needs to be cleaned and normalized. Then the customers are segmented based on their purchase behavior, subscription plans etc. There are also industry specific segmentation standards like Prizm. This should be a good input for rolling up data

Understanding the customer is a first stage in NBO. The customer's behavioral attributes are
analyzed using statistical analysis (clustering, link analysis, decision trees), predictive modelling etc.

It's just not about understanding the customer, the Telecom company needs to analyze the product/service that can be offered to a particular segment. Understanding the Service/Product Offering is the next stage .

For example,

• Wireless devices
• WiFi Services
• VOIP Call services
• Local call services
• Broadband Direct connection
• loud hosting

For example, a propensity model can be created for knowing the probability of the customer for purchasing/trying a particular product / service. For example, a Broadband plan for a customer who is not using the services from the operator (and using the one from a Competitor) with competitive pricing / value add would entice the customer to switch-over.

The propensity of a customer who is currently using the service is set to Zero. This will avoid irittating the customer with the product/service she already uses.

The propensity of each service is ranked along with the best suited time of  sending the offer. The Offer is then sent to the customer through his prefered Channel and at the appropriate context.

Context is always the key - if someone has a 4G-enabled phone, he has a fair chance of upgrading to a 4G network. Knowing the user's device information is helpful here.



Tools like SAS and SPSS can help in achieving this faster. There are Open Source Software - like R, which is catching up fast

Step 4: Measure and Recalibrate

Every offer that is sent is an Test Offer. An offer can be accepted/ignored/rejected. The best NBO strategy is to measure the response of the offer and take the learning back to Analysis stage.

The offer response data is fed back to the database. If there is any need for further data points, it needs to be collected. The model is then refreshed based on the new analysis and deployed.

Next Best Offer Solution Framework

This is a continous cycle and the role of Statisticans, Business Leaders play an important role

Is RFM Model Outdated (Emergence of RFM+ Model)

Recency, Frequency and Monetary (RFM) is the standard model used by many retailers/banks and financial institutions to segment their customers. This model is a time-tested one and has never failed.

 

So

 

The RFM is slowly becoming outdated.

Emergence of Social media, Mobile devices has made the researchers to go with the RFM+ (RFMPlus Model). The RFMPlus includes social media information, Customer Lifetime Value (CLV), Last Visit / Latency / Frequency between visits.

Segmentation Standards / Framework

Nielsen is the market leader in Segmentation. It's framework/system - PRIZM has more than 66 distinct groups/segments (mostly for US Customers) based on their demographic, geographic and purchase behaviour. They have catchy nicknames, images and behavioral snapshots that     help the marketers better understand the customers

P$YCLE is a segmentation system that evaluates consumers using key demographic factors that have the greatest effect on their financial behaviors, such as income, age, presence of children, home ownership and Nielsen’ proprietary measure of IPA. This is mostly used in Finance and Insurance industry

The advantage it is to identify consumer segments who have the resources and propensity to purchase specific financial products and services

Advantages of Segmenting - A case study

Women control over 80% of purchases of most products and services. Many companies have yet to fully capitalize from marketing and selling effectively to women. Insights in Marketing is dedicated to help increase the effectiveness of your marketing efforts towards women!

What are the steps in NBO (Next Best Offer) to give Personalized Promotion

Personalized Promotion / Next Best Action (NBA) in Retail

Increased Personalization is a trend that is common in all industries , particularly, the Retail one
According to a Research, a personalized upsell or cross-sell offer is 30% more effective when delivered within two seconds of initial product selection

The steps, models or algorithms in NBO (Next Best Offer) to give Personalized Promotion are shown below:

The combination of these models would provide a great insight into the customer behaviour and his product and offer preferences, which in turn can be used to provide him context and location senstive appopriate product as the promotions


Big Data and Personalized Promotions

With the advent of Big Data - the Analysts can store huge amount of data and analyze them using Big Data Analytic tools. One such is +SAP HANA , which can be used collect personal customer information and provide optimized offers based on their individual histories and preferences

There are many open source softwares like 'R' and +Hadoop that are spear heading the personalized promotion algorithm development

With the cost of big data hardware on the decline this trend is definitely going to speed up

How Recommendation Engine Works

Recommendation engines are a craze these days. +StumbleUpon  recently fired half of their marketing staff and hired Data Analysts to build a recommendation engine that predicts accurately (or more accurately) the links that it provides to the customers

Recommendation engine work on the same principle of other statistical engines - they are sometimes called personal promotional engines. The key here is to

1) Analyze Data
2) Create Segments / Groups
3) Create Models (Statistical)
4) Send and Measure Offers
5) Re-Calibrate the Model
6) Re-Issue Offers

How to Create Personalized Direct Mail Offers based on Shopping Patterns

Kroger - one of the world's leading grocery chain creates personalized offer on individuals (not segments).
Kroger tracks each customer as an individual. The quarterly mailers it sends contain 12 coupons specific to an individual household and are carefully designed, thanks to dunnhumby's insights. Each part of the coupon is carefully popluated with product choices from the customer's previous
shopping patterns. The last two coupons are for experiments, such as adjacent products — a purchaser of baby food who doesn’t buy diapers might see an offer for diapers.
Apart from suggesting the offers for the customers, dunnhumby also provides insights on placement of these products with the help of data analysis

CVS issues weekly sales circular using a feature called myWeekly Ad. 

This feature uses insights generated based from customer data and transactions to arrive at appropriate offers