The secret weapon of data storage. (Retail Technology: The Details are in the Data).
DATA STORAGE IS CONSIDERED THE SEcret weapon for grocers of all sizes to compete with the discount giants--and each other. As the technology continues to evolve, retailers have new storage options that hold more data. These new options also feature integrated decision support and analysis tools that unlock the knowledge needed to develop promotions and product placement combinations that spark incremental revenue.
Storing data in a dedicated repository is a fairly new concept. During the 1970s, retailers typically flowed data directly into corporate mainframes or onto large tapes. By the 1990s, retailers began moving this data into boxes that could hold several gigabytes of data, and then applied decision support tools to understand the recorded item movement. Gradually, these boxes have expanded and now hold several terabytes of data (1 terabyte equals 1,000 gigabytes, while 1 gigabyte is approximately 1 billion bytes) and have integrated analysis tools that help retailers efficiently sift through volumes of files of sales data.
What is pushing retailers to store data in these centralized boxes? One regional supermarket started loading point-of-sale (POS) data into a repository due to growing competition in the supermarket industry. "You need to understand your business to best compete and go to market, and to do that, you need to learn what your customers are buying," says a technology official at the chain.
"Between the consolidation that makes the big players even bigger and the competition imposed by Wal-Mart, regional players are doing an effective job of driving different service levels," says Rick Schultz, vice president of industry marketing at Dayton, Ohio-based Teradata, a division of Atlanta-based NCR.
These service levels are reliant upon two types of POS data -- item movement data and customer-specific data. Retailers are storing massive amounts of this information to learn about the correlations between the two, as well as to examine what items customers are choosing, when they are being purchased and where they are being bought.
Yet, in the 20 years the supermarket industry has been scanning at the front end, most retailers did not rush into leveraging the information. "Originally, scanning was added so retailers could speed tip checkout throughput and gain more accurate pricing," says Schultz. "Storing and analyzing that data had a slow adoption, and in the last 10 years there has been a strong focus on getting summary financial information, tracking item sales results and using results of market basket data."
The amount of data collected at the front end is endless. Retailers only need to look at the simple receipt customers receive following a transaction. "That is a printed record of the market basket. The cashier's ID, all item detail, all promotional information and coupon usage are captured within a transaction," says Schultz.
The other information captured, which is transparent to the customer, is the correlation of purchased products. If customers have loyalty cards, retailers are able to attach transactions to a specific customer. "This data gives retailers the ability to understand specifics for each customer, identify various segments, and later, analyze targeting strategies to understand purchase trends within a segment. Leveraging this data has caused an explosion in storage growth," he adds.
Simply storing the data is not enough, however. In the past, retailers could afford to wait to see yesterday's statistics to make business decisions. But as the industry struggles with consolidation and stronger competition, even day-old information is just not good enough. "The mantra was, 'We need to analyze yesterday's sale today.' Now what is happening is more customers are updating their data warehouse throughout the day," explains Schultz. "This gives companies the flexibility to look at trends and make sure fast moving items are in stock."
Real-time data allows retailers to also leverage customer data. "By identifying a customer, retailers can truly leverage data not only for marketing but by gaining insight into what a customer buys at an individual store location, and at different times they can align merchandising efforts and pricing efforts," he adds.
Various storage options exist depending on what is being stored and what role it will play during later analysis. Databases support operations and transaction systems, such as specific information on the record that needs to be accessible at the front end, back office or headquarters. For example, this data could support a price look-tip (PLU) application at the front end.
The next version of databases is the relational database that is designed for quick record access, according to Schultz. For example, this type of information storage could be used for access to specific customer or employee records, or price information. "These [databases] support the transaction process system and provide quick updates for a specific business function or operation, and aid in strategic decision making," adds Schultz. "They are not designed to support cross-function business process analysis, or in making decisions that impact another area of business."
That is where data warehouses come into play--combining databases across an entire enterprise. Data warehouses require decision support and querying tools to give users a view into specific functions and the detail and history around the operating systems supporting that function. For example, the data warehouse can capture all POS functions at the chain level, or at specific stores, for a given point in time or particular transaction. They could also be comprised of data marts, which are usually smaller than databases, and focus on a particular subject or department, such as frequent-shopper data or POS.
Retailers are finding the importance of merging databases and data warehouses to analyze affinities, or correlations, between specific items based on the buyers' or shopper clusters' habits and preferences. By identifying individual customer or segment behavior, retailers gain a better understanding of cross-transactions.
The real work to populating a repository occurs after the data is extracted from POS by a dial-up or high-speed network. It is housed in a staging area where item file and product movement files are merged, and then the data is scrubbed, or cleaned, to remove erroneous and duplicated data lines, "The cleansing ensures accuracy and integrity of the data over time," says Schultz.
"If the staging commands are not properly executed, the entire data file could be lost," notes Carlene Thissen, president of Retail Systems Consulting, based in Naples, Fla.
Loading data is a continuous process that happens a minimum of once a day. Other retailers poll POS units multiple times throughout the day. For Wal-Mart, data is pulled in 15 minute increments. It is this quick polling that helped WalMart know what stores were being visited and what was being purchased immediately following the attacks on Sept. 11, according to Schultz.
"Wal-Mart has instant updates of sales and traffic, giving them the ability to analyze different stores in different markets, and stock them with the right product and best stock levels, still remaining ahead of the competition," he says. "This is a good example of the importance of immediate, real-time data feeds."
While not as often as Wal-Mart, grocers tend to poll their store POS registers several times each day and the data is pulled to headquarters through a combination of satellite, frame relay and integrated services digital networks (ISDN), depending on data volume and what network is available in different areas. After polling stores for item movement and sales data, the grocer loads this information into an operational data repository. Once the movement and operational data is combined, the new data is loaded into a data warehouse.
In the early 1990s, one regional supermarket created its first corporate data warehouse that focused on POS, warehouse shipments to store and direct-store-delivery data. By 2000, the grocer upgraded its Redbrick repository running on servers from Sun Microsystems, based in Palo Alto, Calif., and created a new decision support front end, allowing users to more easily access existing data. Today, the grocer is again evaluating more robust options that allow more complex analysis.
A maturity level has been attained with storage products and analys is capabilities. In the past, companies focused on the transaction and not on the analysis. Now, there are options that more tightly integrate storage and analysis products together. This combination leverages the vendor relationship as opposed to having a retailer create its own storage and build its own front end.
With its new storage and analysis system, the grocer will strive for more flexible reporting and more end user power. "Today users cannot schedule reports to come to them, and instead manually run reports after the data becomes available," adds the technology executive for the regional retailer. "In the future, we want to push reports to people on the schedule they want to see it. They can create parameters to run reports on specific data and receive it via email or a dedicated Web page. There are tools available today to do this, and this is one requirement for our new system."
Opting for more powerful, enterprise data warehouses is a strong wave in the supermarket industry. Recently, London-based Sainsbury's Supermarkets replaced its Oracle data warehouse with a Teradata data warehouse. The repository will also support the Teradata's customer relationship management (CRM) solution, enabling Sainsbury's to drive more targeted and personalized communications, improving decision-making throughout the company.
The data warehouse and CRM solution being installed at the London-based IT center at Sainsbury's is outsourced through Chicago-based Accenture. The rollout is expected to be completed early this year. "The food retail sector has undergone fundamental change in the last decade. Today's consumer demands and receives a more varied selection of produce and a more convenient and personalized shopping experience," says Stephen Vowles, Sainsbury's director of customer and category marketing. "Sainsbury's wants to be able to play a proactive role in taking the shopping experience beyond current levels of customer expectation and, with Teradata, it has the technology and experience to do this."
Supervalu is an example of a company that is encountering a challenge similar to the one expressed by Vowles. While seeking new ways to maximize its IT investments, Supervalu also added Teradata's enterprise-wide data warehouse, which uses a decision-support application from MicroStrategy, based in McLean, Va. Now, Supervalu can analyze daily POS transaction logs and apply findings to various corporate operations, including merchandising, pricing and category management.
Albertsons, a new member of the Teradata family, also opted for the enterprise data warehouse to consolidate a variety of its chainwide information. "This new data warehousing system reaffirms our companywide focus on technology," says Bob Dunst, Albertsons' executive vice president and chief technology officer. "Teradata has a great track record of data warehousing success at some of the world's largest retailers, and we look forward to our new partnership."
Albertsons is one of the world's largest food and drug retailing companies, with stores under the banners of Albertsons, Jewel-Osco, Acme, Say-on Drugs, Osco Drug, Albertsons-Osco, Albertsons-Sav-on, Max Foods and Super Saver. Such a vast enterprise forces the company to add technology solutions with a proven return on investment. The chain opted for Teradata to gain a data warehouse and analytical applications that could meet their needs now and in the future, according to Darryl McDonald, vice president of retail enterprise solutions for Teradata.
Even Family Dollar Stores, a discount retailer that sells groceries, switched to an 11-terabyte storage unit, the Enterprise Storage System from IBM, based in Armonk, N.Y., to consolidate all store-based data into one location. The box supports data backups during operating hours, rather than waiting until the store is closed to transfer and store this data.
Adding data to a repository does not guarantee its safety, however. Retailers need to make concerted efforts to add security into the structure of the database or data warehouse, note observers. "Typically users should be password-controlled and users should be tiered, or given access to specific data," says Schultz, depending on the department's relevance or usage of the information.
While there have been significant improvements in data security, retailers are still vulnerable. The U.K.'s Safeway customer database, which holds details about 25,000 shoppers, came under attack in August 2000, when hackers penetrated website firewalls. Approximately 1,000 customers received an email, signed by the "Safeway team," that prices would increase by 25% and they should shop at competing supermarkets. It was unknown whether the customer database was tampered with by an internal or external source, but to be on the safe side, Safeway U.K. temporarily shut down the site, safeway.co.uk.com, while upgrading its Web security. A company spokeswoman claimed the incident did not affect sales.
|Printer friendly Cite/link Email Feedback|
|Date:||Feb 1, 2003|
|Previous Article:||Dial-up is dead. (Retail Technology: The Details are in the Data).|
|Next Article:||POS data provides insight on future sales. (Retail Technology: The Details are in the Data).|