Mining through droves of data for consumer shopping patterns and market insights can make or break a retail company. Market Basket analysis using IBM's Point of Sales System helps smart retailers deploy effective targeted product offerings that generate significant revenues. Find out more on how IBM clients drive higher returns from more precisely targeted campaigns.
IBM Retail | Turn Insight Into Action:Point of Sales Systems Analytics
1. Data Sheet
IBM SPSS Retail Market Basket Analysis
Leverage POS data to understand sales patterns, customer
preferences and buying patterns to create targeted and
profitable promotions
Overview
Solution description “ Multiple purchases
Bulk transaction data as a
within Category W are
high-value asset
Retail operations generate a vast
common, and customers
In today’s highly competitive world,
retailers struggle to differentiate amount of Point-of-Sale (POS) who make these
themselves via their product offerings transaction data. The sheer bulk of purchases also tend to
and by how they promote products to this transactional data – recording, at buy Product T.”
customers across all channels. Smart the item level, every purchase through
retailers deploy effective targeted stores, on-line store fronts, and other SPSS algorithms are highly scalable
product offerings which can generate channels – makes it very hard to and able to process huge amounts
significant revenues; poorly chosen understand the repeated patterns of of POS data efficiently. Our solution
ones, though, waste money and purchase that give insight into customer includes powerful tools to let retailers
opportunity. behavior and preference. browse and explore the associations
discovered and decide which are the
Retail operations generate huge In Market Basket Analysis, smart most relevant and potentially lucrative
amounts of transactional algorithms analyze huge quantities to apply in their business.
information, holding a wealth of of transaction data to reveal
detail on product purchase “associations” – the patterns which Turning Insights into Action
patterns. The sheer volume of data, show the linkage between products These patterns can drive decisions
however, makes those patterns typically purchased together. Typical on how to differentiate assortment,
obscure and impossible to detect by insights might be: merchandise stores and to develop
manual inspection. combined offers of multiple products,
“ Products A and B tend and within and across categories, to
If retailers have insight into these to be bought together.” drive sales and profits. They can be
patterns, they can ensure the implemented across an entire retail
products they offer and promotions chain, by channel, or, if the data is
“ If Products C, D and
they run match shopper preferences analyzed at the store level, specific
and behavior, and give maximum
F are purchased, it is offers can be formulated and rolled out
return on their marketing spend. extremely likely the at a local level.
Retailers able to link purchases to customer will also buy
individual purchasers can take this Product E.”
even further, tailoring offers to
specific customer segments and
driving higher returns from more
precisely targeted campaigns.
2. Offers might be implemented through • Interaction information that The high level of personalization of
in-store displays. For example, two describes how they interact with the these offers, and the fact they are
products showing a strong tendency retailer – for example, their use of based on robust predictive models,
to be purchased together can be on-line facilities (e.g. e-commerce delivers high conversion rates and
stacked together, with a special “buy and “loyalty club” websites) and increased revenue per shopper and
both for…” offer encouraging combined order or service hotlines. per visit.
purchases. • Attitudinal information that
describes why they do what they Empower decision makers
Alternatively or additionally, coupons do, such as satisfaction scores or SPSS Market Basket Analysis is
based on these offers could be sent out net promoter scores, or ecological delivered to business users in the form of
in a weekly mail drop to all households or other attitudes captured in dashboards, reports, alerts, and analysis
in each store’s catchment area, surveys and in feedback at points of provided by IBM Cognos 8 BI. This
encouraging store visits as well as the interaction. allows marketing and merchandising
desired shopping behavior. executives to understand sales
Apply predictive models to identify patterns, customer preferences and
Getting Personal the most appropriate offers for each buying patterns so they can improve
These offers, however, are “one size fits individual customer product sales and margins, and create
all”. While they may generate significant SPSS algorithms analyze the whole targeted and profitable promotions.
revenues by leveraging common range of customer data, relate that to The combination of IBM Cognos 8 BI
purchasing tendencies, they don’t previous purchases, and build predictive and SPSS provides a complete view of
differentiate between customers who models that for any customer and historical performance coupled with a
may have different preferences and potential offer can be applied along with predictive view of the future.
propensities to respond to different offers. business rules and other analyses to:
A common integrated platform allows
Retailers who have a closer relationship • Decide whether that offer is valid for for a single version of the truth to be
with individual customers – for example, that customer disseminated throughout the retailer,
through a loyalty card scheme or • Predict the probability that the unlocking data from silos within the
on-line shopper registration – can refine customer will respond to that offer organization. This allows chains to
this approach. In this case, they can • Calculate the value to the retailer of seamlessly combine information that
combine analysis of all customers’ the customer accepting that offer was previously time consuming to pull
shopping patterns with information on together, error prone, and in many
the individual customer’s shopping Applying these models across offers cases led to decisions made on intuition
habits. This behavioral information can enables retailers to select the best set as opposed to objective and data
then be combined with any other data of offers for each customer or customer driven, which most retailers
held on the customer, for example: profile. These are then delivered to the drive toward.
customers in the most appropriate way.
• Descriptive data, combining self- For example, loyalty card holders might These capabilities enable retailers to
declared data with any bought-in receive several coupons enclosed with slice information in meaningful ways
geo demographics based on ZIP or their monthly statement. in order to isolate specific challenges
postal code or identify areas of opportunity.
2