Real-time Technologies

Guide to E-Commerce Technology

Imagine collecting data about customers and their activity across retail channels in real time and having the ability to analyze conditions as they happen.

Success stories include Overstock.com, an online retailer using fresh data to refine marketing campaigns, personalize services in real time and, ultimately, better serve customers to drive additional business. Demand-driven merchandising is the new model for success: having the right product available at the right time (and price), and knowing who will buy it. Real-time technologies have arrived, giving forward-thinking retailers a winning I.T. strategy and an important competitive edge.

The Internet enables retailers to maintain a dialogue with customers. In the last several years, retailers have also been working toward collecting more complete and timely customer data from their store point-of-sale as well as online order management systems. They’re doing so to leverage critical customer data and fine-tune the total customer experience across multiple channels.

Without complete and accurate transactional data coming from in-store POS, as well as from online and catalog channels, no investment in better business intelligence systems or tools will make a difference. Systems must be capable of not only capturing accurate product information but also customer information related to each transaction to better match consumers and the products they want, and when and how they want them.

The good news is technology solutions are available to help retailers know and keep customers. For example, most point-of-sale vendors support an XML standard—developed by the Association of Retail Technology Standards XML committee of the National Retail Federation—to facilitate the real-time exchange of POS data with upstream transactional and business intelligence systems. Many retailers use internal messaging subsystems for the asynchronous exchange of messages in XML format to other systems as a method of integrating different generations of business applications across channels.

Unlike their tightly bound predecessors, modern POS systems use many of the same technologies as those found on web-based applications, with browser-based or browser-like thin presentation layers, an application layer, and data layer, and with continuous XML messaging capabilities. This enables retailers to use the same set of business rules and data regardless of the presentation.

These systems also provide data that enables the retailer to relate a customer to a store market basket: not only does the retailer know what gets sold, but what gets sold together with other products, and who the customer was. This information is an invaluable new piece of business intelligence because retailers now can understand the affinities between items purchased, and also begin to understand who their best customers are and what they buy. Combined with information gleaned from online and call center channels, retailers gain powerful insights into true customer demand across multiple channels.

As timeliness of data becomes critical, more organizations are moving toward continuous feeds into business intelligence databases so they can eliminate data and analytic latencies to access fresh data and respond more quickly to changes in the selling environment. Survey responses from a July 2007 RSR Research benchmark study, “The Next Generation of Business Intelligence: Driving Customer Insights across the Retail Enterprise,” show many retailers are moving toward continuous feeds into a data warehouse.

Retail winners (those companies that outperform their peer group) are more aggressive, with 48% indicating they now collect data continuously from the selling environment. Winners are clearly aware of the opportunity to improve customer satisfaction, top line revenue, and gross profit profitability by using technology to proactively respond to conditions in all channels.

Real-time customer insight

First and foremost, retail winners have a focus on the multi-channel customer dimension, in addition to the traditional dimensions used for merchandise planning (product, location and time). It is the integration of data, technology and process that sets them apart. These winning organizations understand demand because it is based on actual customer data gathered from the shopping experience, and they demonstrate shared best practices, which include:

l A focus on the quality of data. Legacy POS hardware and software can only take a retailer so far. But a well-architected and modern POS system that can accurately capture customer-related data with the market basket, and is integrated to multi-channel business intelligence capabilities, will provide accurate and timely information that can be used to customize the online shopping experience to meet the true demands of consumers.

l A focus on real-time access to real-time information. Fresh information gathered from the shopping experience is invaluable and drives decisions about what products and services to offer, when to offer them, how to promote and price them, how to present them most effectively, and even who the target customers are. And the faster a retailer can get at the information, the faster it is able to fine-tune marketing and stock to effectively meet customer demand.

l A focus on a single view of the customer. Technology can address the challenges of gaining the single view of the customer across channels (online, phone, store) either by having one physical repository of customer and sales history data or by creating one logistical view of customer and sales history data through data integration. Recognizing that retail data sets can be huge—measured in many terabytes—the problem won’t get any easier by waiting. Retailers of any size should address this now to avoid negative impact to business.

Moving to real-time data

What differentiates any successful retailer is not its products but its services, and these are increasingly driven by the information asset. What does it take to deploy a real-time business intelligence system? First, some homework.

Many organizations are hesitant to move to a real-time architecture for fear of exorbitant cost and complexity. Yet recent cost studies have shown that deploying real-time technologies and continuously streaming data may be no more expensive than their batch or ETL extract transform load counterparts. (ETL refers to the use of a single tool to extract, transform and load information between multiple databases.)

Retailers need to answer questions about their own customers and business to deploy technologies in the right way. For example, what data is critical for immediate analysis? This is typically customer transaction and product movement data, as well as page clicks from the online channel. This type of data analysis is often action-oriented, triggering alerts to real-time conditions in the selling environment. What data can be analyzed once a day or less often over time? Typically this type of data analysis is more retrospective in nature. A real-time architecture can support both of these scenarios and better support future requirements.

Retailers also need to decide who gets the data. Today data is informing many groups across a retail organization, but not everyone will require the data so quickly. For example, in addition to business analysts and executives, marketing and retail operations staff typically need real-time data in order to evaluate how customers are reacting to promotional campaigns as well as conditions in the selling environment itself.

For example, store managers can be alerted that a hot selling item needs to be replenished from the back room within the next hour based on an analysis of the rate-of-sales compared to the calculated amount of inventory on the shelf. An online merchandise manager, alerted to spikes or downturns in web sales, could make any necessary adjustments to online product displays or marketing campaigns.

What are the potential pitfalls to avoid? Organizations must take legacy systems into account, particularly more mature retailers, where customer data has been collected over a long period of time. “Stovepipe” systems are all too common, where the store, web and catalog are operating independently, each with their own set of technologies and vendors.

Integration can be a challenge, but is critical. It’s a good idea to consult with vendors, systems integrators, consultants or other experts with experience in retail and legacy systems integration. Careful planning and execution will ensure that integration and implementation of real-time business intelligence and data warehousing solutions goes smoothly without impact to customer-facing systems or the customer experience.

Online is real time—and technology is finally catching up to the industry’s needs. The most successful Internet retailers are thinking ahead and recognizing that real-time customer and product data provides significant competitive advantage. Continuous feeding of data to business intelligence systems enables real-time, or right-time, decisions that impact stock, pricing, marketing and customer experience.

Hottest items

Business intelligence can detect the hottest selling item (as well as the weakest) as it happens and enable a business to react. Traditional retailers are catching up to their online counterparts by using the same technologies to improve the time-to-action by vastly reducing the data and analytic latencies inherent in their older legacy systems.

Overstock.com is an example of a company that uses real-time business intelligence to create competitive advantage. As part of its winning strategy, the retailer has been able to create a real-time, single view of the customer to better understand purchasing habits, refine marketing efforts and more effectively drive business to the web site.

Overstock.com’s innovative approach uses GoldenGate Software to pull transactional data from online production sites and ultimately feed the data in real time to a data warehouse from vendor Teradata. Overstock leverages this real-time customer data to produce more accurate reports and improve analytics, enabling them to very quickly determine the results of marketing campaigns and promote personalized offerings to its customers.

When a retailer learns something isn’t working a day after the fact, it is a day too late. Retailers who implement real-time technologies for business intelligence consistently outperform their counterparts.

Now, imagine the possibilities created by knowing: What value proposition drove a customer to the purchase? Which promotional campaigns are working? Which customers are responding? What are they buying? What are their buying patterns? Is there an opportunity to cross-sell or up-sell?

Today’s consumers have virtually unlimited choices. They are driven by the currencies of information, time and value. These expectations have extended to all retail channels. Retail winners understand this, and the need to be able to respond more quickly to consumer demand. And when the alternative could mean losing a customer, can a company really wait?

Brian Kilcourse is a managing partner at Retail Systems Research LLC, former CEO of Retail Systems Alert Group and former CIO of Longs Drug Stores. He can be reached at bkilcourse@rsrresearch.com.