Insight-Driven Retailing Blog

You gotta love WholeFoods!

View larger image

The best all-time pranks are at Museum of Hoaxes. I especially like the Taco Liberty Bell.

When I worked for Schlumberger, my boss was a Frenchman based in Paris. He told me about the Minitel, a precursor to the Web launched in France back in 1982. It was a dumb terminal connected to the phone line via modem with a simple keyboard. Phone customers were given the device for free instead of white pages. The Minitel could search for phone numbers, get stock quotes, make purchases, and chat.

When he was a bachelor, my boss (Jean-Marc) used the Minitel to manage his shopping list. Then once a week, he'd hit "send" and the order would go to a nearby grocery store where they would pick and deliver his food. It was a good system until one day he didn't notice the store changed all it's product numbers. He was quite surprised when the diapers, kitty litter, and tampons were delivered.

Nevertheless, you can't beat the convienence of a small terminal in the kitchen, which is exactly what Verizon Wireless is hoping for with their new Verizon Hub. Its a fairly standard VOIP phone with desktop widgets and interesting services which are better reviewed elsewhere. Its target segment is Verizon Wireless customers that will trade their landline for a VOIP phone with tight integration to their wireless phones.

When I first saw this device, I thought it would be a perfect next-generation Minitel, where my family could easily access information in the kitchen without a PC. But this device doesn't have a browser or an email client, and that makes it really tough to buy groceries.

There are two strategies for leading SOA initiatives. When IT drives the strategy, SOA is typically used as an integration tool, making products work together with minimal coupling. When the initiative is driven by the business, SOA becomes an enabler for agility, allowing greater flexibility for the processes. Both approaches are fine, but I think the most value comes when everything is derived from business processes.

The business-process approach is more top-down, and therefore requires much more up-front work to document the details of the business processes. But the end result produces highly valuable data objects, services, and events that enable the processes. Most retailers want to be given a base, best-practice process that they can tailor. This allows them the freedom to innovate and differentiate from their competitors.

That's generally the approach that Oracle takes with its Application Integration Architecture (AIA). The AIA process produces Foundation Packs that contain the data objects, services, and events necessary to create processes. These are great for implementation partners and retailers that want to develop their own business processes. The Process Integration Packs (PIPs) take it a step further by including the implementation of one or more related business processes. PIPs include all the tech and application patches necessary to make the processes work out-of-the-box.

There was a press release yesterday that announced more Foundation Packs and PIPs along with the following quote:

Oracle knows that the accelerated delivery of more Oracle Application Integration Architecture products is critical to our customers’ need to innovate cost effectively while reducing the time to deploy end to end composite business processes,” said Oracle Vice President of Application Development and Product Strategy for AIA Jose Lazares. “The releases of Oracle AIA Foundation Pack 2.3, Oracle AIA Foundation Pack for Communications and six new Oracle AIA Process Integration Packs provide even further advantage for the enterprise to quickly define a given business process and then rapidly deploy a standardized composite business application that can evolve as business needs demand it.

It took a while to gear up, but these process-based integrations are gathering momentum, including the release of a new one for retail that provides Financial Operations Control.

I visited Legal Seafood in Boston last week, and they used "pay at the table." Instead of handing over my credit card, they brought the terminal to me. This particular model, from Ingenico, used 802.11b wireless networking to transmit back to a base station, which in turn connected to the network. That's the first time I've seen that technology used in the US, although I've used it in Europe lots.

Ten years ago I worked on a similar system from Schlumberger, a French company that specialized in smart cards (amoung lots of other things). Our product line was called MagIC, which supported magnetic stripe cards as well as integrated circuit cards (i.e. smart cards). I thought it was a clever name. The MagIC series ran an embedded real-time operating system, an application manager, and one or more applications. (The CPU was an 8088!) When a card was swiped or inserted, the application manager determined which application was best to process the transaction. That way you could run separate applications for debit, credit, stored value, loyalty, etc.

After working on various smart card projects in the US, I finally gave up. The US networks (phone and internet) are just too efficient and cheap to worry about procesing transactions offline, which is one of the main cases for smart cards. They are great for vending, which makes them popular on campuses like universities, resorts, and ships. But they will never take off as general purpose payments in the US, although they continue to be popular in Europe and Asia.

Back to Legal Seafood. The "pay at the table" payment allows restaurants to simplify payment. Instead of running the credit card for the meal then adding the tip in a second transaction, its all done in a single transaction. This is one less trip for the waiter, skips the need for keying tips into the terminal, and saves embarassment when a card is over limit. Plus if the terminal is also certified as a PIN pad, it can accept debit cards too. I hope other restaurants give it try.

Remember when you pumped your own gas then stood in line to pay at the little booth? (This doesn't apply in NJ where the state mandates full-service. It takes twice as long to get gas but keeps people employed.) Then pay-at-the-pump was introduced and more control was handed to the customer along with a few more steps to perform. This was a great trade-off and win-win for both the retailer and the customer. On the rare occurrence when I stop at a station that doesn't have pay-at-the-pump, I just keep driving.

A few years back I had a similar experience in a Swedish grocery store. When you walk into the store, you insert your loyalty card in a machine that unlocks a scanner. As you add things to your cart, you scan them. Then there's really no need for a checkout (although there are random audits) and time is saved.

Retailer benefits: less labor, opportunity to offer electronic coupons, less need for price tags
Customer benefits: running total, coupons, faster checkout

Again, it seems like a win-win. Self checkout really only helps the retailer, since the customer has to scan each item during checkout. The customer doesn't really save any time. But with a portable scanner that goes with the cart, there are more benefits.

Unfortunately, portable scanners never took off in the US. One might think LP stopped its adoption, but I don't think shrink is really the problem since the service is only offered to loyal customers and random audits tend to keep people honest. As reported by the Hartford Courant, Stop & Shop is experimenting with portable scanners, but the reviews are mixed.

In today's market, I think retailers will need to offer multiple ways in which to shop. For grocers, this means supporting traditional checkout, self-checkout, portable scanners, and home delivery. This increases the burden on IT to keep it all working, but should lower labor costs, and increase customer satisfaction.

Working on technology strategy for Oracle Retail is a role that brings new opportunities and challenges. Our retail customers live in an ever-changing world and Oracle is moving quickly to continue bringing the best. Not long ago, we added BEA to the arsenal of technology and solutions we bring to bare, and now today we know Sun is on the horizon. Stay tuned!

Update: Q&A with Larry Ellison regarding Sun's hardware, with lots of Exadata mentions.

Everyone wants to know if the economy has hit bottom, and we're all looking for the signs. Retailers have been hit particularly hard by the recession, with many reporting sales slumps not seen in years. Several retailers have recently announced the worst is over. They stop short of saying the economy is back on track, but at least its not getting worse. Stability is the first step toward recovery.

Tesco, Britain's largest retailer, saw signs of stability in Asia and Europe, but their results were dragged down by their new US-based chain, Fresh and Easy. Nevertheless, they saw a worldwide sales increase of 15%. Coach, Macy's, and Target are also no longer experiencing deteriorating sales. Of course calling the bottom is not the same as starting the recovery. We could move sideways for a while as employment improves, credit frees-up, and consumer confidence builds.

Mike Duke, Walmart's CEO, spoke with Matt Lauer on NBC's Today Show (see video below).

As they toured a new Walmart in King of Prussia, Pennsylvania they discussed signs of the economy. Mr. Duke mentioned several measures to watch:

* Sales of expensive cuts of meat are down. Presumably when people have more money, they will buy better steaks.

* Adult apparel sales are down, but not childrens' clothing. Parents are willing to make sacrifices for themselves, but not for their kids.

* Consumer electronics are still strong. This may be because families are forgoing vacations and saying home with the HDTV and Wii.

* Vitamins are selling well, possibly because workers can't afford to get sick and miss work.

* Cash sales have increased as a percentage of total sales, probably because credit limits are tighter.

Although Walmart's Duke would not say he saw the recovery in sight, he gave us some interesting signs to watch.

"Cloud Computing" has to be today's most hyped technology. Using Google Trends, one can see that the first mention occurred around mid-2007 with peaks occurring around September 2008 and now (April 2009). I think the recent buzz eminates from several large technology firms getting into the business of clouds. This has led me to ponder how cloud computing might impact the retail industry.

McKinsey just released a study that I found interesting because it stuck to the numbers. Its starts by defining the term as:

Clouds are hardware-based services offering compute, network and storage capacity where (1) Hardware management is highly abstracted from the buyer; (2) Buyers incur infrastructure costs as variable OPEX; (3) Infrastructure capacity is highly elastic (up or down).

Using this as the foundation for calculating and comparing costs, the report states that cloud computing is less financially attractive to large businesses. This is consistent with what we see in the market: the primary customers are small and medium-sized businesses. Those that can afford the initial capital outlay can usually operate their servers more cheaply. However, clouds are still relatively new so I'm sure they will get cheaper over time.

The report goes on to say that clouds do not meet the 99.99% uptime service level agreements required by many CIOs. It also notes that a 99.99% SLA may be overkill in many of those cases. For retailers, some systems are certainly more critical than others.

McKinsey concluded that cloud computing was over-hyped and in danger of hitting the trough of dissillusionment, to put it in Gartner terms. While I agree the term is over-hyped, I'm not pro- or anti-cloud. I'm simply searching for the killer application of clouds for retail. The Association of Retail Technology Standards (ARTS) is collecting comments on cloud computing as they consider researching and assembling a white paper on the topic. I guess others are interested as well.

What do you think about cloud computing? Could it benefit the retail industry?

Update: Bob Parker expressed some interesting views in an article over at RIS News.

An important function of IT at any retailer is to turn data into information. Sounds a little cliche, but its important now more than ever. Most retailers collect every transaction from every store, track every movement of goods, and record every customer service interaction. So there's no shortage of data, but how does one translate all that data into actionable information? How can information be used to make better decisions?

If you don't have a data warehouse, or you have multiple siloed data warehouses that need to be combined, then Oracle Retail Data Model (ORDM), which was announced in a press release today, deserves consideration. It runs on Oracle's database and easily makes use of Oracle's business intelligence suite (OBIEE+) on the front-end (although other tools are also possible).

ORDM incorporates all the latest technology including OLAP cubes, data mining, and optional use of Exadata, Oracle's data warehouse appliance. This gives retailers access to advanced analytics such as forecasting out-of-stock situations, understanding hidden patterns for loss prevention, contribution and market basket analysis, product affinity, customer clustering and segmentation, halo impact and promotional lift.

If you recall the previous posting on Inmon vs Kimball, then ORDM follows Inmon's designs. It provides 3NF tables, based on ARTS standards, into which raw data is loaded. Then internal ETL aggregates the data for faster reporting. This approach gives retailers access to both summary and detailed data.

For those that are wondering about RDW, its still alive and kicking too.

On the heels of my posting regarding cloud computing for retail, I found an interesting article on A.C. Moore's hosted merchandising implementation. A.C. Moore, an arts & crafts retailer with 136 stores and 60,000 SKUs, knew it needed to improve operations with a merchandising system, but also knew it didn't have the IT staff to support one. So they turned to OLR to host their systems.

Not only did the hosted approach help accelerate the implementation of the applications, but it also enabled A.C. Moore's information technology group to stay focused on other strategic initiatives vital to grow the business. The model also helped eliminate A.C. Moore’s need to make capital intensive purchases for its data center and offered a leveraged support model that mitigated the risks associated with finding, attracting, and keeping application support resources.

OLR uses IBM servers running Oracle Enterprise Linux and has a dedicated T1 from their data center to A.C. Moore's headquarters. Clearly software-as-a-service (SaaS) has a place in mid-market retail. But how does a CIO choose the right model? There are three choices:

Traditional - retailer's hardware and software maintained by the retailer. High initial cost with medium ongoing costs.  Has the most control (less risk from an outage and security perspective).

Cloud/IaaS -- vendor's hardware but retailer's software maintained jointly. Avoids capital expenditure for hardware, but allows for customized software. The cloud vendor maintains the hardware, OS, middleware, and database but the retailer maintains the applications.  This is what web hosting has been doing for years.  Very popular with start-ups b/c they don't have capital.

Hosted/SaaS -- vendor's hardware and software maintained by the vendor.  This is basic outsourcing of applications made famous by SalesForce.com.  Low startup costs, but possibly larger ongoing fees with less control regarding upgrades, etc.

I don't think there's an easy formula as the best solution depends on the starting point. But in general terms, the table to the right represents a good beginning. And of course, I'm interested in other people's perspective -- so leave a comment.