Insight-Driven Retailing Blog

In retail, the roles of merchandising and marketing have most often been separate, but in today's hyper-competitive landscape it's growing more important for the roles to collaborate, if not merge altogether. Merchants must always focus on maximizing gross margins, and effective use of promotions is a crucial lever. As you can see from the graphic below, advertising is an uncertain investment, so it's vital retailers make sure their advertising dollars are well spent. The recent announcement of DemandTec's acquisition of Connect3 reinforces that strategy.

Oracle's Promotion Planning and Optimization product already encompasses the features envisioned by this new marriage, so there's no need for retailers to wait. PPO uses promotional intelligence to measure past promotions in order to predict the effectiveness of future ones. It considers promotional lift, cannibalization, and halo-effect using well-tuned algorithms developed in our Cambridge office.

PPO's design allows merchants to perform what-if analysis to better understand and evaluate the results of a planned offer based on entry level price point strategies, the offer, or the marketing vehicle. Embedded item and market basket analysis ensures users have a combined view of the forecast in one solution. In addition, the solution provides a collaborative, centralized platform, which gives sharing capability between marketing and merchandising departments –enabling better and faster planning between internal teams. Merchants and designers also have the ability to graphically view ad space to determine best usage of content, and layouts to maximize profits per page. Forecasting the overall impact to department goals provides operational teams with a better upfront investment plan, to ensure all organizations are aligned before the promotional campaign is approved.

Software that both facilitates collaboration and maximizes promotion profitability is one way for retailers to squeeze more out of their promotions budget.

The closing of Circuit City was bad enough, but the ripple effects that will be felt across the US are only just beginning. This ABC News article looks at the side-effects of such a large closing, especially with respect to real-estate.

Ironically, real-estate may be one of the reasons Circuit City finds itself in this position. CC had tremendous growth in the 1980s and the first half of the 1990s. At the time, the electronics market was fragmented, and there was no strong competitor. This allowed CC to consider itself a destination location. That is, people will travel specifically to the store as opposed to driving to a shopping area. So CC built stores in "B" locations that were cheaper. Best Buy, on the other hand, focused on locating their stores in the best "trade areas" -- large shopping centers, preferably near highways.

Poor store locations, combined with a vast number (many from acquisitions) of stores in need of remodeling, started to seriously impact performance in 1995. You can learn more about Circuit City's demise in this excellent Shopping Centers Today article from late 2004. I also built a CC timeline in this posting.

Today there are lots of options for retailers to better analyze their real-estate portfolios. Ubiquitous mapping software combined with demographics can create interesting mash-ups that help determine which areas are under-served and which are over-served. For example, the Oracle Site Hub can work with Google Maps and Oracle Spacial to deliver meaningful data in order to make better decisions about store placement. See my past posting for more info.

I wish the best for all those at CC and hope the slowdown in retail doesn't get any worse.

It costs a lot to maintain all the different systems necessary to run a retail business. To complicate matters, the systems are distributed so in addition to all the computers at headquarters, there are servers located in the stores, distributions centers, and warehouses. The CIO has to keep these systems up-and-running, tuned to perform, and secure from both internal and external threats. Its a tall task made even more complicated by various regulations.

Staples recently took a step in the right direction by selecting Oracle Enterprise Manager to ease adoption of ITIL and automate compliance controls for PCI and SOX. Retailers that don't already have a comprehensive systems management implementation are probably overspending on IT support.

A good systems management product will monitor systems for hiccups, provide the ability to tune for performance, manage configurations, and collect audit logs -- all from a central console. Today's system administrators need to "manage many as one" if they are to be efficient. So as retailers look to squeeze costs out of their existing infrastructure, they should give serious consideration to systems management.

When you call CapitalOne, they identify the caller (using CallerID) then determine the most likely reason for the call. For example, if you were just assessed a late fee, chances are you're calling to complain about the fee. This prepares the customer-service rep and helps speed the call along. And while you're on the phone, they determine which offers you're most likely to be interested in, so they may offer you some steak knives, which can be conveniently billed directly to your credit card. CapitalOne was a pioneer in applying real-time analytics to customer interactions back in the 1990s, and it continues to help them grow.

Today eHarmony tells us who to date, Netflix recommends movies to watch, and Pandora tells us what music we're likely to enjoy. Using contextual data (who you are, time of day, where you came from), history (your past interactions), plus some interesting algorithms (e.g. regression) can enhance customer interactions in real-time. Retailers like Amazon have been getting pretty good at this, like recommending past products you've browsed, or items related to past purchases. Imagine the possibilities if other sources of data were incorporated, like from social sites...

1. David arrived to Amazon from a Google search on digital cameras
2. David's brother Mike has a birthday coming up (from Facebook).
3. David only purchases things on sale.
4. Mike has a Canon digital camera on his wishlist.

==> Offer a high-end Canon digital camera on sale

Oracle has a strong track record with its Real-Time Decision product in several industries, including retail where its use in e-commerce seems to be gaining momentum. It provides three key capabilities:

Improve Business Responsiveness
+Optimize customer experiences with cross-channel real-time decisions at the point of interaction
Maximize Interaction Value
+Learn from each interaction and take the best action with embedded predictive analytics and rules
Enable Self-Adaptive Processes
+Integrate business intelligence into interaction processes with closed-loop predictive analytics automation

Combined with business intelligence and data mining, real-time decisions have the potential to streamline customer interactions and increase sales; both are welcomed in retail.

Those in the retail industry probably know the significance of 1962. For those that don't, that's when Walmart, Kmart, and Target first opened their collective doors. Since then, those three chains have changed the landscape of retail forever.

Dayton-Hudson (which was found in 1902) opened the first Target store in Roseville, Minnesota. S.S. Kresge (founded in 1899) opened the first Kmart discount department store in Garden City, Michigan. Sam Walton (owner of several Ben Franklin Stores) opened the first Wal-Mart Store in Rodgers, Arkansas. The rest, as they say, is history. Comparison chart.

I've mentioned the ability to map store locations in previous posts, but FlowingData has gone a step further and mapped store openings over time for both Target and Walmart. These animated maps illustrate the spread of Target and Walmart across the US as if they were two strains of a highly contagious virus.

Target Store Openings
Walmart Store Openings

I wonder who has the momentum now. Kohl's?

As continued proof that retail is taking it on the chin, About.com published their list of retail bankruptcies for 2008 and 2009. It contains well-known brands such as KB Toys, Linens 'N Things, Circuit City, and the latest victim Ritz Camera. Some of these companies are down but not out. Filing Chapter 11 just means they need time to reorganize and require a temporary break from creditors. Of course simply restructuring debt may not be enough. Sometimes the organization needs some upheaval as well. Filing is a chance to change the way business is done. Those that don't embrace this chance are on a track toward Chapter 7 and eventual liquidation.

So what should retailers be doing now?

At the beginning of the year, Hung LeHong of Gartner suggested that letting IT take the year off was a bad idea. He suggested four small BI-related initiatives with big perceived value:

1. Use advanced visualization capabilities such as map-based interfaces to spice up online reports and dashboards.
2. Provide more advanced users the powerful, yet simple-to-use, new analytical capabilities found in in-memory, columnar and other next generation BI tools
3. Merge unstructured consumer data found in reviews, blogs and social networks into reports - imagine being able to see sales trending alongside average customer review scores.
4. Create the ability to send key reports and metrics to your managers' smart phones.

Below are some Oracle-specific ideas that fall into Hung's categories...

1. Oracle's acquisition of AVT, a visual space planning company, along with our work with WebCenter fit into this category. Using meaningful graphics is an important way to help users find the data they need to make decisions.

2. Oracle's latest database, 11g, is full of new and interesting features. But the biggest news this past year was the announcement of Exadata, the data appliance built with HP. As Kevin Closson explains, Exadata can really speed queries up using a combination of optimized hardware and software.

3. Retailers have been embracing social media, and its becoming an important part of brand image. Using applications like Siebel to both mine and publish via social sites will be a growing trend.

4. Mangers are on the go, so they need information to follow them. Fortunately, the Oracle Business Intelligence Enterprise Edition platform is able to deliver information to a multitude of channels, including phones.

These are all great ideas that retailers should be considering. Some other "quick wins" that have big paybacks and can take as little as 12 weeks to implement are:

1. Merchandise Financial Planning
Part of the Enterprise Performance Management cycle is planning the merchandise budget. Moving beyond simple spreadsheets can make a planners job much easier and more accurate as well.

2. Markdown Optimization
Divorcing fact from emotion can help merchants maximize profits and minimize inventory. Marking down merchandise too early will erode margins, and waiting too long could leave you with extra stock. Finding the right markdown cadence is crucial for fashion retailers.

3. Replenishment Optimization
There's a balancing point between maintaining adequate stock levels and minimizing carried inventory. Stock too little and customer leave; stock too much and you're stuck running a clearance. Optimizing for for the right compromise has big returns for retailers.

What are some other small projects with big ROI's that retailers should be considering?

Almost any retail store has at least one server running within its four walls. Larger stores often have servers configured for redundancy because the business has become so dependent on them. Store servers are used for POS, labor management, signage, reporting, etc. Of course being in this cramped, often dirty environment can be hazardous to a server's health. They are often relegated to a closet where there is little air flow, and I've seen situations where rats have eaten through network cables.

But in a retail store square footage is very valuable. So servers have been getting smaller, more energy efficient, and have a longer mean time between failures. With the advent of computer-on-a-chip technology, it won't be long before servers shrink to the size of power bricks. Wait a second... that day has arrived.

Introducing Marvell's Plug Computer (see picture). For $99, you get a 1200 MHz CPU with 1 GB of Flash and DRAM, ethernet, and USB connectivity using only 5 watts of electricity. It runs various distributions of Linux, and typically is always on. See TG Daily for a more complete review.

Ok, maybe its not powerful enough to actually run a store, but the industry is making fast progress in the right direction. Recall paying $5000 for that first PC that ran DOS? Now you can get a PC that's 1/25 the size, 1000x more powerful, and costs under $100. They are quickly closing in on a disposable status.