Big Data and Product Affinity

56 %
44 %
Information about Big Data and Product Affinity
Business & Mgmt

Published on February 19, 2014

Author: MichelBruley



Analytics can offer a key possibility identifying which products are sold together, because that information can be used to influence targeted communication efforts, store layouts, and in-store promotions.

Product Affinity Extract from various presentations: CRS, BUS 782, Aster Data … January 2013

Product affinity analysis is one of the basket analysis techniques Assortment analysis & Management Customer Analysis & Marketing Promotion evaluation & management • • Traffic Builder ID Traffic analysis • • Frequency of visits Consumer Penetration • Av. Basket Metrics • • • Item Contribution Trx Builder ID Price Point Contribution • • • Value, av. Purchase Discount behaviour Customer Modelling • Promotional Evaluation Av. Dist. # of Items & Depts • • • Trx Builder ID Item contribution “Variety driver” • Purchase Variety Behaviour Customer Modelling • • • • • Item Deletion Cross Sell Lost Sales Prevention Potential overstock • Co-marketing opportunities by customer • • • • Cherry Picker • • Item deletion Item contribution • Cherry Picking Behaviour Consumer Profitability • Price Point Identification Price Elasticity No. Of Baskets Price Point • Transaction analysis • Product quality (returns) • • In-Store activities • In-Store Activities Payment Type • • Local Store Assortment Pricing by segment • Store Performance evaluation Promotion Evaluation Promo Item Selection Event Strategy Cross sell opportunity • Vendor Participation • Co Merchandising Opportunities (visual merchandising) • • Promotion Evaluation Promo item selection • • Margin Protection Vendor participation Promotional Pricing • • • Price integrity Fraud detection Local pricing • • Fraud Detection Cashier productivity • • Manpower planning In-store activities Payment type relevance • Store layout & visual merchandising Payment influence Trx. Profiling Item Performance Fraud detection Time of Day • • • Traffic Analysis (manpower planning) • • • • Consumer Penetration (with Cust.ID) Traffic Builder ID Traffic builder Opp. • Store Operations Promotion Evaluation (w/Cherry Picker) Affinity Analysis Vendor & Supply Chain Management • • • • Customer Profiling Customer Modelling Propensity to Buy • • Event Strategy (“Early Bird” opportunities) Promotional Evaluation (behaviour change) • • Item Performance Consumer/retailer relevance of item

Product Affinity Definition Identify which products are sold together and use that information to influence targeted marketing efforts, store layouts, and in-store promotions Product Affinity enables an organization to detect product/service purchase patterns, linkages, and cross-sell opportunities in order to increase revenues. Results from this application will enable the organization to identify, with a high degree of accuracy, those customers most interested in specific products, services and product/service groupings

Affinity Analysis Affinity Analysis is a modeling technique based upon the theory that if you buy a certain group of items, you are more (or less) likely to buy another group of items. The set of items a customer buys is referred to as an item set, and market basket analysis seeks to find relationships between purchases. Typically the relationship will be in the form of a rule: Example: – IF {beer, no bar meal} THEN {chips}

Product Affinity and Cross- Selling For instance, customers are very likely to purchase shampoo and conditioner together, so a retailer would not put both items on promotion at the same time. The promotion of one would likely drive sales of the other A widely used example of cross selling on the internet with market basket analysis is's use of suggestions of the type: – "Customers who bought book A also bought book B", e.g.

Product Affinity Analysis Process Historic market basket data and analyzes are used to build more effective marketing programs: – Past customer purchase data is used to identify which products/services are acquired by which customer groups – Predictive analytics is applied to this data to discover profiles of customers most likely to buy the products in each group – These profiles are used to target those customers most likely to respond favorably to specific cross-sell campaign – Pair-wise product associations are also determined to enable the constructed of offers featuring the purchase of these pair products – Customer product dislikes are also identified so that company does not promote unwanted products Benefits that can be realized from utilizing this solution: – Improve customer knowledge allowing company to better understand what their customers are likely to buy and not buy. – Increase revenue and decrease costs by identifying those customers most likely to respond to cross-sell campaigns

Behavior Prediction This uses past consumer behavior to foresee the future behavior of their customers. This analysis includes several variations. 1. Propensity-to-buy analysis- understanding what a particular customer might buy. 2. Next Sequential Purchase- predicting the customers next buy. 3. Product Affinity Analysis- Understanding which products will be bought with others. 4. Price elasticity modeling and dynamic pricing- determine the best price for a given product.

Product Affinity = Link Analysis Aims to establish links (associations) between records, or sets of records, in a database There are three specializations – Associations discovery – Sequential pattern discovery – Similar time sequence discovery

Link Analysis: Associations Discovery Finds items that imply the presence of other items in the same event Affinities between items are represented by association rules – e.g. ‘When customer rents property for more than 2 years and is more than 25 years old, in 40% of cases, customer will buy a property. Association happens in 35% of all customers who rent properties’.

Link Analysis: Sequential Pattern Discovery Finds patterns between events such that the presence of one set of items is followed by another set of items in a database of events over a period of time. – e.g. Used to understand long-term customer buying behaviour

Link Analysis: Similar Time Sequence Discovery Finds links between two sets of data that are timedependent, and is based on the degree of similarity between the patterns that both time series demonstrate – e.g. Within three months of buying property, new home owners will purchase goods such as cookers, freezers, and washing machines

For Analytics: SQL or SQL-MapReduce Teradata SQL Aster SQL-MapReduce SQL is better for: SQL-MapReduce is better for: •Standard transformations across every element in a table •Custom Transformations •Standard aggregations using GROUP BY on tables • sum(), max(), stddev() •Dimensional Joins •Set Filtering • Lookups, data pruning to limit a table to a subset. •Presentation formatting • For example, “get me top K counts only” • e.g. unstructured data, log extraction, conditional manipulation •Custom Aggregations •Inter-row Analysis, like time-series •Layered queries • Nested queries, sub-queries, recursive queries •Analysis that requires reorganization of data into new data structures • Graph analysis, decision trees, etc.

Time Series Analysis discover patterns in rows of sequential data Sales Transactions {user, product, time} Purchase 1 Purchase 2 Purchase 3 Purchase 4 Aster Data SQL/MR Approach • Single-pass of data • Linked list sequential analysis • Gap recognition Traditional SQL Approach • Full Table Scans • Self-Joins for sequencing • Limited operators for ordered data

Identify common product baskets of interest Cross-Channel Transactions 43M Customers Online Alone! Teradata Aster solution In-Store Transactions Challenge • Identify correlations between purchases made over time Impact • Move beyond “people who bought this also bought” to time-based recommendations EAN Author Store time 15682817 823201 JK Rowling 100 12:00 PM 16816193 123101 Shakespeare 105 1:45 PM 19825996 182191 Rick Riordan 201 3:00 PM 15528047 With Aster Data • SQL-MapReduce for market basket analysis indicates correlations between products userID 823201 Walter Isaacson 100 4:20 PM item_no type EAN 12334 book 823201 Product 13345 music -- Catalog 21456 periodical -- 82673 toy -- Online Transactions EAN Author time 823201 John Grisham 12:00 PM 123101 Dostoevsky 1:45 PM 182191 Obama 3:00 PM IPAddress 823201 Stephen King 4:20 PM

Basket Affinity: Retail Business Need Overview: For most retailers, Market basket affinity is a well known tool for cross-promotions and marketing. However, there is very little affinity known “outside” the basket. For example, there are many cases where the consumer will return to the store to get the additional item(s) they did not purchase initially. Examples: Electronics retailer (Best Buy, Radio Shack, Fry’s): – A Blue-Ray player is purchased online on a given date. The same customer visits the store next week to buy HDMI cables and a B-R disc. Fashion Retailer (Target, Macy’s, J Crew): – A customer purchases a dress and hand bag one week. Returns within a month to buy matching shoes. With this sequential affinity analysis, the retailer can send very specific and timely email marketing, to drive traffic and increase revenue.

Overview of Cross-Basket Affinity Challenge • Difficult to do in a relational DB due to the sheer size of the combinatorial permutations of the various purchasing sequences. Cross-Channel Transactions X Customers X Marketing Campaigns Transactional DB Customer Loyalty Requires good customer recognition via a credit card database or a customer loyalty card program. With Teradata Aster • Use nPath/Sessionization to identify “super” baskets within a time window. Tighter time window implies higher affinity. TransID UserId Date/Time Item UPC 874143 10001 11/12/24 83321 543422 20001 11/12/28 73910 632735 • 30002 11/12/24 39503 452834 10001 11/12/30 Run Basket Generator to identify the most frequent affinity items & subcategories. Impact • Enables more accurate targeting of customer needs; reduce direct marketing spend, increase revenue yield. Address Phone 10001 10 Main St 555-3421 20001 24 Elm st 232-5451 30002 49019 534 Rich 232-5465 Retail EDW Product/Item Hierachy Item UPC Category Dept 83321 Heels Shoes-Womens 73910 Handbags Accessories 39503 • UserId Dresses ApparelWomens 49019 Perfumes Cosmetics Marketing/Promotions Date CampaignID UserId 11/12/24 3241 10001 11/12/28 2352 20001 11/12/24 3241 30002 11/12/30 2352 10001

Cross-Basket Affinity Example UserId Aster MapReduce Platform Address Phone 10001 10 Main St 555-3421 TransID st UserId Date/Time 24 Elm 232-5451 Item UPC 30002 534 874143 Rich 10001 232-5465 11/12/24 83321 543422 20001 11/12/28 73910 632735 30002 11/12/24 39503 452834 Prepares multi-structured data 20001 10001 11/12/30 49019 •Stitches rows together by customer in a timeordered view Scans all records to produce complete set of sequences •No need to define patterns in advance •Fully parallelized for scalable performance using MapReduce where not feasible with SQL/SAS Step 1: nPath/ Sessionization to identify “super” baskets. TransID UserId Date/Time Item UPC SuperSessi on SeqNum 874143 10001 11/12/24 83321 101 1 452834 10001 11/12/30 49019 101 2 Summarize sequential affinity output for business exploration •Rank order the most popular sequential purchase paths. Step 2: run Basket Generator to identify frequent affinity items. ProductU pcB Support Confiden ce Time Window Sequentia lOrder 83321 49019 0.10 0.30 14 days true 73910 Product UpcA 83321 0.11 0.25 7 days false

Identifies the Cross-Basket Affinity Products The frequent sequence of purchased items identifies products B & C which are likely to be sold when a customer buys a certain product A. – Leverage this Cross-Affinity analysis to run more targeted marketing campaigns; increase affinity purchases – Personalized email offers yields higher customer retention and loyalty, and reduces churn. Aster SQL-MR functions nPath/Sessionization and Basket Generators are key algorithmic differentiators; this process cannot be done in a scalable manner in a relational DB and/or SAS

Affinity Use Case 1/3 Analyzing item price movements and its impact on: – Basket size over a long duration (6-10yrs) will provide key insights into halo impact and affinity contribution for items – Basket composition over a long duration (610yrs) will provide key insights into price bands for items Analyzing Affinity of items over a long duration (6-10 yrs) will provide key insights into running better promotions, planogram and price planning of around affinity items

Affinity Use Case 2/3 Affinity Analysis •Analyzing Affinity of items over a long duration (6-10yrs) will provide key insights into running better promotions, planogram and price planning using items affinity •Time Frame: 8 Years, 1 Banner - Data Set: Transaction Data, Product hierarchy Consumer Migration •Analyzing declines in consumer segments over large timeframes. •Time Frame: 3 Years - Data Set: Transaction Data, Segment Data, Competitor Data, Pricing Data Pricing Affinity •Analyzing item price movement and its impact on basket size and affinity of items over a long duration (6 years) •Data Set: Transaction Data, Price data - Time Frame: 6 Years Competitor Impact •Data Set: Transaction/Consumer/Competitor/Pricing Data, Unit_Inf - Time Frame: 8 Years Social Media •Integrating consumer online data (Social Media - Facebook) with existing transaction data and understand impact on consumer loyalty. •Data Set: Should be collected by vendor

Affinity Use Case 3/3 Data – ~ 8 years of transaction data (2004 up to Sep-2011) – 15 Billion baskets (or transactions) – 225 Stores – 367K Unique UPCs – 12 Categories: Alcohol, Cereal, Frozen – Ice Cream, Laundry Detergent, Cheese (Shredded/Sliced/Chunk/Other), Paper Towels, Pizza & Shelf Stable Juice Solution – Aster SQL-MapReduce: Collaborative Filter – Query Runtime: 48 minutes (4 Workers using Columnar)

Add a comment

Related presentations

Canvas Prints at Affordable Prices make you smile.Visit http://www.shopcanvasprint...

30 Días en Bici en Gijón organiza un recorrido por los comercios históricos de la ...

Con el fin de conocer mejor el rol que juega internet en el proceso de compra en E...

With three established projects across the country and seven more in the pipeline,...

Retailing is not a rocket science, neither it's walk-in-the-park. In this presenta...

What is research??

What is research??

April 2, 2014

Explanatory definitions of research in depth...

Related pages

product affinity | IBM Big Data & Analytics Hub

About; Contact; Content By Type. Blogs; Videos. All Videos; IBM Big Data In A Minute; Video Chats. Analytics Video Chats; Big Data Bytes; Big Data ...
Read more

Product — Boomerang Commerce

Product affinity. Traffic. ... Vendor costs. Competitor pricing. Product reviews. Units per ... Utilize market data, machine learning, and big data ...
Read more

How To Identify Fake Big Data Products - Forbes

How To Identify Fake Big Data ... by naming itself a Big Data product when that is ... value by claiming affinity with Big Data.
Read more

Big Data's Power to Innovate Content and Enhance Reader ...

How educational and non-educational book publishers can redefine product ... with big data analytics. Digital ... better understand readers’ affinity for ...
Read more


Big data is unharnessed in terms of its velocity and volume. ... TRANSFORMING BIG DATA INTO ... upgrade strategies, product development, ...
Read more

The Database of Affinity Can Bring Discipline To Brand ...

Nate Elliott's Blog ... affinity data remains ... Brilliant thoughts on how big data and social offer more than what is thought of ...
Read more

What is Predictive Analytics and Big Data? Webopedia ...

Predictive Analytics and Big Data. Predictive analytics is an enabler of big data: ...
Read more

Big Data Meets Presidential Politics | CSC

... the Obama campaign successfully employed big data ... Big Data Meets Presidential Politics; ... CSC calls ‘affinity ratios ...
Read more