 |
|
|
By the numbers
|
 |
Company Companhia Brasileira de Distribuição–Grupo Pão de Açúcar, São Paulo, Brazil
Year founded 1948
Current banners Pão de Açúcar, Comprebem/
Barateiro and Sendas supermarkets, Extra hypermarkets and Extra Eletro electronics and home appliances
Market penetration Companhia Brasileira de Distribuição–Grupo Pão de Açúcar is the leading retail chain in Brazil, with 560 stores and 10 distribution centers located throughout 11 Brazilian states; 15% total market share
Technical leadership Largest B2B platform in Latin America; pioneered CRM and supply chain automation in Brazil
Total selling space 10.5 million square feet
2003 sales US$3.84 billion
Total employees 58,000
Corporate objectives To increase selling space an average of 10%–15% annually over the next five years and to assure the best shopping experience in every store
|
|
Silvio Abrahao Laban
|
 |
What do you do?
I'm technology and processes director, Companhia Brasileira de Distribuição-Grupo Pão de Açúcar.
What is your educational background?
I graduated in 1985 from the University of São Paulo with a degree in naval engineering. I'm completing my Ph.D. in business administration from Fundação Getúltio Vargas with an emphasis in marketing.
Where have you been?
I focused on marketing and consumer goods as a senior consultant with Andersen Consulting in the mid- to late-1980s. Then,
I worked for the Americas Region at Carrefour as information technology director and technology coordinator. I subsequently
headed up Wal-Mart Brazil's information systems division and was director of public relations.
How involved are you in industry initiatives?
I serve on the executive board of the Global Commerce Initiative. I have been co-chair of ECR Brazil since 2001 and co-chair of ECR Latin American for the past year.
What do you do for fun?
I enjoy comics, traveling and reading.
|
|
Behind the solution
|
Teradata Warehouse powered by: Teradata Database V2R4.1, 2-node 5250 NCR Server, 4-node 5300 NCR Server, 1-node development system
Storage: 8TB
Operating system: UNIX MP-RAS
Teradata Utilities: Fastload, FastExport, TPump and Multiload
Tools/Apps: MicroStrategy Intelligence Server, VP-ASP
|
|
|
TERADATA Q&A
First in line
Grupo Pão de Açúcar uses technology to keep products and service at the peak of freshness.
by Len Lewis
Centuries ago, early Portuguese explorers known as bandeirantes used São Paulo, Brazil, as a staging area to search for gold, emeralds and other precious stones along the Amazon. Today, this sprawling megalopolis is home to Grupo Pão de Açúcar, Brazil's top supermarket chain and the staging area for expeditions in search of data mining's jewels.
Over the years, Grupo Pão de Açúcar has seen unprecedented growth, with 22 separate retail acquisitions, the renovation of about 400 stores in the past six years and the development of a distribution network that supplies about 80% of
the merchandise to its stores throughout 11 Brazilian states. Since its inception as a small family-owned pastry shop in 1948, this 560-store chain has been a pioneer in the Brazilian retail food industry.
The road hasn't always been paved with gold. Over-expansion and internal struggles nearly pushed the chain into bankruptcy in the early 1990s until its chairman, Abilio dos Santos Diniz, a member of the founding family, took the reins and turned the business around.
New technology was partly responsible for this rebirth. Grupo Pão de Açúcar was one of the first chains in Brazil to equip stores with POS systems. It also was among the first to use a computerized system for automatic inventory replenishment, which led to significant reductions in labor costs at the company's supermarket and hypermarket divisions.
Now, improved handling is lowering shrinkage, stock-outs and delivery cost. At the same time, technology is spearheading improved operating efficiencies at distribution centers where cross-docking—loading multiple stores' trucks at the same time from the same distribution center—is resulting in lower inventory and savings in working capital.
Teradata is central to Grupo Pão de Açúcar's IT restructuring and its entry into the brave new world of single-source data warehousing. At present, the chain is in the process of migrating from an old Oracle system and moving reports out
of its mainframes and into a Teradata Warehouse.
Teradata Magazine spoke with Silvio Abrahao Laban, technology and processes director, Companhia Brasileira
de Distribuição–Grupo Pão de Açúcar, to discuss the chain's progress with the new platform and its plans for the future.
Q. Has Grupo Pão de Açúcar's focus on technology been key to its success?
A: Exactly. We take a realistic approach, so it's not just about data. Technology encompasses many things that are important to our business. In the 1990s we developed a number of initiatives, including investments in customer knowledge and automation in order to speed up checkout lines. For us, information technology was the beginning of scientific- or fact-based retailing.
Q. Has that technology push continued?
A: Yes. We believe our company is built on three main pillars. The first is people, because even the best technology is worthless without the right people. The second pillar is capital structure, because a sound financial structure is essential to fueling our expansion. Third is investing in the right things—and that is technology in a broad sense, beyond information technology.
Q. In what IT areas do you invest?
A: Supply chain management, category management, store layout, construction and site research are areas in which technology is embedded. And information plays a major role across all these areas. So this is the emphasis we give to information technology.
Q. When did Grupo Pão de Açúcar develop its first data warehouse?
A: You must understand that Brazil is very different from the United States in terms of IT development. The first
bar code was read in the United States in 1974. In Brazil, the first (bar code) was not read until the mid-1980s.
Data capture at the point of sale is the beginning of the information chain, and if you don't have it, data warehousing
initiatives are worthless.
Additionally, Brazil's government closed this market to imported IT goods until 1992. The production of IT goods was considered strategic and should be done locally. But it prevented the companies from even buying PCs from IBM or servers from HP or NCR.
Q. What was the first step?
A: In 1994, Companhia Brasileira de Distribuição tried to move forward with information gathering and data analysis. By 1995, all stores had bar coding and computer-aided ordering. And as the business got more complex, we decided it was time to develop a data warehouse. We evaluated many technologies at that time. Teradata was included in the selection process, but we started with Oracle.
From 1995 until 2000, everything was working OK. But there were some issues with the development of the data warehouse that had to do with the evolution of the company in terms of numbers of stores, sales, logistics and the complexity of our distribution network. Our growth was escalating at an incredible speed and so was the amount of data (we collected), but we needed to extract more value from the data warehouse application.
After five years of operating it online, we had very few users, and information (such as) product inventory wasn't readily available. And when we started a loyalty program in 1999, we decided to do it on a separate database because of internal problems with scalability and availability for all users. So we had the company data warehouse with all the inventory, replenishment and stock information, and we had another with customer information. We needed to integrate into a single data warehouse.
Q. Is that when you chose Teradata?
A: More or less. By 2001, we decided to re-evaluate the situation. Our vision was to have a single data warehouse for the entire company that was available to the majority of users. We were not sure that our existing technological platform would fit this vision. We invited Oracle to evaluate the present status of our data warehouse and develop a revamping plan. At the same time, we spoke with Teradata in Brazil, told them where we wanted to go, and asked them to develop a roadmap that would enable us to have a world-class business platform and tell us what it would cost in terms of investment and data migration.
Q. Was scalability an issue at the time?
A: It was a major concern. Between 1996 and 2002, our sales increased more than 250%. We had 349 stores, and SKUs had grown significantly. We also made 20 acquisitions during that period, and our market share had jumped from 8% to 15%. With this kind of growth, you can't afford to have scalability issues or information bottlenecks.
Q. What other problems had your company been experiencing?
A: We began to have some performance issues due to the integration of different platforms and the new analyses that were required by end users. We handled it but knew if we kept investing in the existing platform we'd experience problems. We realized that we needed to revamp Oracle or change to a new platform. We decided to replace the existing data warehousing infrastructure.
Q. Why did you consider Teradata?
A: The company has a remarkable presence in the retail industry. Scalability was the main reason to choose Teradata. In consultations with Oracle we didn't get good enough answers.
We also wanted specific knowledge of the supermarket industry. Since Teradata is related to NCR, there's a strong historical relationship with the industry. Also, Teradata has a lot of retail decisions already embedded in its solutions. If you don't have to develop them from scratch the conversion is faster.
Q. Did you expect these solutions to provide more information?
A: Yes. We had IBM mainframes with a major database, and a lot of information was migrated there rather than our data warehouse platform. Our intention was to fix all the architectural issues by having a platform that could grow with us in the future. For example, at the beginning of the 1990s we operated just one banner.
Now we have five operating in 11 Brazilian states. We didn't have a supply chain network, and now we have 10 distribution centers across the country.
Q. When did you finally decide to go with the Teradata platform?
A: The decision was made at the end of 2001. Then we spent part of 2002 redeploying the entire system. We are currently migrating existing information onto the Teradata Warehouse. During the process we also acquired two more chains. Because we operate with a very lean IT structure, this generated some turbulence in terms of resources. The project was designed for a certain projected growth. But normally you can't project acquisitions—they just happen. We've worked with the local Teradata group and managed to rearrange the schedule with little impact.
We're going to turn off the old repository by June. Right now, we're finalizing the migration of all reports that existed on the old platform. The problem is not with the data—it's the volume. A lot of reports were developed for the old platform and we need to adapt all of them to the new data model. We're on target. But moving data from the mainframe is a huge effort.
Q. Are store-level associates currently part of the user group?
A: At this point it's only corporate. We're getting more information on things like out-of-stocks, but instead of a mainframe report being sent to the stores, specific out-of-stock information is sent to them directly. This gives them more timely and accurate information. We want to be capable of acting on data. But to act on it we need to put information closer to the problem or to people at the stores capable of solving the problem. Stores need to take care of customers, products and the store itself. We don't want them distracted by too much data. But by having all the information on the same repository and giving them selected access, we will facilitate the process of translating data into information and then into action.
Q. At this point what results have you seen from the system?
A: Customer data won't be migrated to Teradata until later this year. But our service levels have already improved for things like inventory stocks, replenishment, slotting allowances and financial results. But at the same time, we're gradually eliminating one of the bigger problems in legacy systems by migrating toward a single source of data. Each user group used to have its own data mart, and if there was a change in basic information, you couldn't be assured that all data marts were properly updated. With the single source of data all transactional data moves to the main repository and is then fed into other data marts. The result is unified, more consistent information.
Q. Do you expect to see more benefits from the system over time?
A: Yes. Because it's a more appropriate platform, we can start migrating applications to the data warehouse that are now implemented on the mainframe. We've also been working with Teradata in the United States to establish a roadmap for the future—a compass for where we should be heading in order to extract maximum value from the system. We need to prepare the infrastructure for the future by analyzing the return on investment for each application.
Q. What applications does this roadmap include?
A: We already have a lot of solutions like price simulation, competitive analyses and computer-assisted ordering in all stores. But there's an opportunity to improve the accuracy of our replenishment process and extract more benefits. For example,
I have a lot of competitor information on the mainframe, which can be extracted to another platform. But I still need more flexibility and in-depth analysis in a more cost-effective manner.
Generally speaking, in 2004 we're talking about more forecasting and promotion analysis. In 2005, the applications will focus on the customer and financial sides as well as on fraud and detection. Toward the end of 2005 and into 2006, we will embark on more analytical customer relationship management, which we already have on other platforms.
Q. Will the system eventually help with marketing and merchandising?
A: It already has in other platforms. But sometimes we need to generate a specific report, like an assortment review, within the IT group. We intend to put this information at the internal user's fingertips so he or she doesn't have to depend on the IT group to get it out of the mainframe.
Q. What will that do to the relationship between IT and marketing?
A: In our company, before the specific function, everyone is and must act as a merchant. So, I could say that we will replace the part of the relationship that is still based on marketing people putting demands on the IT group. A more accessible system would enable these demands to be met by the marketing group. We'd then work with the IT group on a consulting basis to look for new ideas, approaches and analyses that the data warehouse could carry out.
Q. Do you have any advice for others considering implementing the Teradata solution?
A: I only have some general advice when it comes to IT. Of course, it's important that for cost reasons we keep our IT infrastructure as homogeneous as possible. But beyond that, you need to be sure that every IT initiative is properly aligned with the business. If a company is only growing organically and at a controlled rate, things tend to be less complex. If that was the case we might have kept the other platform. But because of the strategic path we were taking we decided to implement a technology that could offer greater benefits.
Len Lewis, editorial director of Lewis Communications, Inc., has covered the retail industry for many years and currently contributes to several industry publications.
© Teradata Magazine-June 2004
| RELATED ARTICLES: |
|
|
|
|
|
Printable version
Send to a colleague
