Monday Jul 20, 2015

Unlocking the Secrets of Big Data with Science

Buzzwords abound in the tech world, and mobile, social, big data, and cloud are the four horsemen.  Of those I think big data is least understood, but possibly the most powerful as exemplified by my favorite quote:

“Data is the new oil; analytics the new refinery.” -- unknown

The retail industry has always been dealing with large volumes of data; it's just that now the velocity and variety have dramatically increased. So the problem for retailers is two fold: first, how do they convert all that data into meaningful information? And second, how do they make the information relevant and actionable to employees?

Oracle Retail's answer comes in two families of cloud services: Oracle Retail Advanced Science and Oracle Retail Insights.  Continuing our momentum in delivering cloud services, these products simplify implementations allowing retailers to realize value faster and with less IT effort.

Oracle Retail Advanced Science Engine

We have a long history of employing data scientists that cull through donated data to find useful insights that can be productized in algorithms.  Over the years we've led the industry in areas like demand forecasting, markdown optimization, and localized assortments.  To make it easier to bring new algorithms to market, we've built our Advanced Science Engine, a platform tuned for analyzing data at scale and exporting the results.  Over the past six months, we've adjusted the architecture to provide this in a cloud deployment.  The specific science being offered is packaged into three separate cloud services:

Advanced Clustering

This solution provides insight into how store clustering best benefits the business.  It helps to answer questions such as:

  • What categories or merchandise classifications benefit most from clustering?
  • At what level of product or location hierarchy should clusters be created?
  • What product/location attributes should be leveraged?

The solution automatically selects the best clustering method depending on the clustering approach selected, and allows for what-if scenarios that help explore the data.  Is scores clustering approaches for comparison, and recommend the optimal number of clusters.  This fosters consumer-centric assortments that ultimately increase sales.

Customer Decision Trees & Demand Transference

CDTs map the decision process made by shoppers to purchase items.  They help a retailer understand if they have the right variety of sizes, flavors, colors, etc. in their assortment.

Suppliers often provide consumer decision trees that help retailers understand the impact to assortment decisions based on "generic" consumers.  But when a consumer is identified, they become a customer.  Our solution focuses on customer decision trees, which reflect the actual customers in your stores.  The solution even allows a side-by-side comparison of imported consumer decision trees alongside the calculated customer decision trees.

Using this science helps to reduce duplication in assortments, and prevent dropping unique items to which customers are attached.  Demand transference helps forecast if customers will switch to alternatives so that the number of variety of items offered can be optimized.  This solution often works closely with Category Management.

Assortment & Space Optimization

This solution helps to identify the optimal targeted assortment withing the space constraints.  It understands shelves, pegboards, and freezers and incorporates various business rules and visual merchandising standards.  The what-if analysis is very helpful in exploring options while seeking to maximize profits for any given space within the store or store cluster.

Integrated Planning

These cloud services can be used with existing planning products, or with Oracle Retail planning solutions such as Category Management Planning & Optimization, Macro Space Optimization, Retail Demand Forecasting, and Advanced Inventory Planning.


Tuesday Jun 23, 2015

Perspective from NRF Protect 2015: Adidas Uses Oracle Retail XBRi to Reduce Fraud at the Point of Service

Analytics and exception-based reporting, made available across all stores brings Big Data-style science to loss prevention

In advance of NRF Protect, here is a look at what some of our customers are doing to reduce and respond to fraud in stores. This is the first in a two-part series. To learn more, be sure to visit us at the Oracle Retail Booth #1227 at #NRFProtect this week in Long Beach, CA. 

Retail loss prevention professionals are well aware that employee theft and employee-related fraud account for the biggest single segment of shrink. According to the November 2014 Global Retail Theft Barometer, employee-generated shrink accounted for just over 40% of the previous year’s $128 billion total, even more than the one-third generated by shoplifting and organized retail crime.

Given these facts, retailers have a compelling interest in understanding and curtailing employee-generated shrink. The conundrum, however, is that no retailer can effectively investigate every single transaction in every single store. Fortunately, employees who commit fraud tend to follow specific patterns. By using tools that apply science to the problem, retailers can shift this challenge from a Big Data problem to an opportunity for insight.

One of the most important loss prevention tools is exception-based reporting, using advanced algorithms to constantly monitor point-of-service (POS) activity, identify potentially fraudulent transactions, and alert specialists automatically. Trends, outliers and “red flags” can be measured and tracked by region, store, or individual employee. By providing essential data to multiple levels of staff – from individual loss prevention specialists in the field to regional managers – an organization can effectively empower their team to root out fraud, and act quickly to resolve it. Doing the same thing manually is impossible when transactions multiply over dozens or thousands of locations. 

For adidas, the global designer and manufacturer of athletic shoes, clothing and accessories, it was nearly impossible to consistently identify the causes of shrink and fraud in its 2,470 stores worldwide. The company was unable to perform loss prevention exception reporting and faced operational challenges including lack of data protection, multi-system misalignment, difficulty adjusting to time zone and language variances, and system failures resulting in non-compliance issues.  In a recent Chain Store Age article, adidas shares how it reduces fraud in employee and administration losses following its implementation of Oracle solutions. Adidas shared their experience at Oracle Industry Connect. You can download the presentation adidas: Measuring and Managing Loss to Preserve Profit from the Oracle Retail RACK. 

Now available as a cloud service, Oracle Retail XBRi Loss Prevention Cloud Service captures all POS transactions and then administers advanced business analytics that apply a laser-focused look at key loss patterns. Designed to be completely agnostic to the POS solution and source data, XBRi integrates with both Oracle and third-party POS solutions – even multiple solutions – giving retailers flexibility and freedom of choice. The cloud service shifts funding from a potential capital investment in software and IT infrastructure to an operational expense. 

To learn more, be sure to visit us at the Oracle Retail Booth #1227 at #NRFProtect this week in Long Beach, CA. 


Friday Jan 30, 2015

A New Kind of Credit Card Fraud

Fraud analysts at credit card issuers (banks) pour though transactions looking for credit card fraud. Patterns in this data are what usually lead investigators to find breaches. Find a bunch of transactions involving stolen numbers, then work backwards to find the commonality and you've got your breach. But there are lots of additional patterns in that data as well.

Two enterprising fraud analysts at Capital One decided to look at purchase patterns for public retailers, then use that information to place bets in the stock market. Why not, right? All of Capital One's credit card transactions are sitting in the database for them query. As you can see in the heavily redacted SQL query below, it took some serious analytical skills.

For example, they compared this quarter's volume of Chipotle transactions to last quarter's and seeing that sales were strong decided to buy call options just before the earnings announcement. Earnings were good so the stock went up, and their $100,000 investment netted $270,000 in profits -- for basically three days work.

Over three years, they made around $2.8M which worked out to about a 1819% return on their investment. Of course this is considered insider trading and they were eventually arrested. So the guys that were looking for credit card fraud were actually committing a different kind of fraud. There's gold to be mined in that data. See the complete story over at Bloomberg View.

Tuesday Oct 28, 2014

Oracle OpenWorld 2014: The Pace of Change for Retailers

At the center of change in retail are the 9 Billion devices now connected to the Internet – a number predicted to go to 50 Billion in fairly short order.  Beacons and other location devices are a disruptive force for the retail industry, and they completely change how retail experiences are built. The proliferation of mobile devices, among shoppers and store associates, opens up new ways to tell customers where inventory is, where can they get it, at what price, and it invites a whole new set of competitors.

Welcoming retail executives, partners and industry experts to the Retail Experience @ Oracle OpenWorld 2014 in San Francisco, Oracle Retail Senior Vice President and General Manager Mike Webster said that the rapid pace of change being driven by mobile and other influences will not slow anytime soon. It’s one of many reasons retailers are suddenly looking to accommodate a higher velocity of data in a variety of different formats.

“Big data is not big news in retail but we are having to solve problems around the velocity and the variety of data,” said Webster. “How do we bring in social interactions and marketing interactions together, to give you a more unified view of the entire customer engagement. We are a mobile world, with 6 billion mobile subscribers.”

In retail today, there are tons of investments across social, mobile, analytics and cloud. Seven out of ten companies don't know their current stock position. Retailers must return to the basics. The biggest item on retail balance sheet is inventory. Transparency is the key to shift inventory closer to customers to impact the bottom line and satisfy the consumer.

To help retailers succeed, Oracle “spends more on R&D than any other solution provider in the industry, and the most basic element of what we are creating is to make sure you reach customers where you need to, that you are able to hit the basics and innovate. Our focus is building the best solutions for retailers,” said Webster. During his keynote, Mike Webster took the opportunity to share the highlights built into our upcoming release coupled with the unique capabilities that MICROS adds to the footprint.

Our success is measured in terms of customer results. Oracle Retail saw great success with vanilla implementations and this trend reflects all of the work done to fine-tune retail functionality across the Oracle Retail suite of applications. With the introduction of version 14 and the work with world-class partners, we have allowed customers to focus on the business opportunity with less complexity, customization and integration from the implementation process with best practices built into the solutions.

Customers including Hot Topic, Kohl’s, Gordmans, and Zenni Optical are just a few of the retailers benefiting from recent implementations of Oracle’s robust, mature retail solutions. Customers should continue to expect us to take out complexity and take out cost, Webster added.

The Retail Experience @ OpenWorld 2014 presentations are available in the Oracle Retail virtual community. Log in to the RACK to review the presentations from the retail track.  

-  Commerce Anywhere: Retail Innovation
-  ULTA Beauty: Improving the Customer Experience with Oracle Commerce
-  Inventory Management for Commerce Anywhere with Dubai Duty Free
-  Running Oracle Retail Applications on Oracle Systems with Kohls
-  TOMS: Oracle Commerce Case Study
-  Retail Analytics: Creating Value from Insight
-  How Two Brazilian Retailers Linked Shopping Across Channels with Oracle Commerce (Part 1)
-  How Two Brazilian Retailers Linked Shopping Across Channels with Oracle Commerce (Part 2)
-  Retail Trends: An Oracle Perspective

Tuesday Oct 16, 2012

What's Old is New Again

Last night I told my son he could stream music to his tablet "from the cloud" (in this case, the Amazon Cloud).  He paused, then said, "what is the cloud?"  I replied, "a bunch of servers connected to the internet."  Apparently he had visions of something much more magnificent.  Another similar term is "big data."  These marketing terms help to quickly convey topics but are oversimplifications that are open to many interpretations.  At their core, those terms are shiny packages holding recycled ideas.

I see many headlines declaring big data changes everything, but it doesn't.  Savvy retailers have been dealing with large volumes of data since the electronic cash register was invented.  But there have been a few changes to the landscape that make big data a topic of conversation:

1. Computing power has caught up to storage volumes. Its now possible to more thoroughly analyze the copious volumes of data retailers have been squirreling away.  CPUs are faster, sold state drives more plentiful, and new ways to store and search data are available.  My iPhone is more powerful than the computer used in the Apollo mission to the moon.

2. Unstructured data is everywhere.  The Web used to be where retailers published product information, but now users are generating the bulk of the content in the form of comments, videos, and "likes."  The variety of information available to retailers is huge, and it's meaning difficult to discern.

3. Everything is connected.  Looking at a report from my router, there are no less than 20 active devices on my home network.  We can track the location of mobile phones, tag products with RFID, and set our thermostats (I love my Nest) from a thousand miles away.  Not only is there more data, but its arriving at a higher velocity.

Careful readers will note the three Vs that help define so-called big data: volume, variety, and velocity. We now have more volume, more variety, and more velocity and different technologies to deal with them.  But at the heart, the objectives are still the same:

  • Informed decisions
  • Accurate forecasts
  • Improved optimizations

So don't let the term "big data" throw you off the scent.  Retailers still need to execute on the basics.  But do take a fresh look at the data that's available and the new technologies to process it.  The landscape will continue to change and agile organizations will always be reevaluating their approaches.  You just need to add some more weapons to the arsenal.

About


David Dorf, Sr Director Technology Strategy for Oracle Retail, shares news and ideas about the retail industry with a focus on innovation and emerging technologies.


Industry Connect


Stay Connected
Blogroll

Search

Archives
« July 2015
SunMonTueWedThuFriSat
   
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
22
23
24
25
26
27
28
29
30
31
 
       
Today