In an economic downturn the first reaction of most companies is to slash costs, often indiscriminately, in an effort to make up for lost revenue. The most recent slowdown, however, saw nascent signs of a new trend: increase revenue by optimizing price and allocation of existing resources.

While still in very early stages, this trend is most evident in the retail industry, where a handful of companies are running pilots or implementing limited projects under a variety of solution names: revenue management, merchandise optimization, price optimization, precision retailing or demand-based management. Whatever it is called, initial results are encouraging, with users reporting revenue lifts of from 3 percent to 7 percent - an increase that can translate into even bigger percentage hikes for profits.

Eric Mitchell, president of the Professional Pricing Society, agrees that retail is an industry where the "precise price optimization of individual SKUs can pay big dividends. Excel spreadsheets, manual calculations and human intuition are the way most retailers set prices today," he says, "[but] retailers are beginning to understand the potential of optimizing pricing strategies using technology solutions that tell them what their customers are willing to pay."

Most early retail projects and software tools focus on promotions pricing and mark-down optimization, though a wide range of solutions is being developed by a variety of vendors, including traditional supply-chain experts Manugistics and i2 Technologies and retail inventory specialists like JDA Software and E3, a recent JDA acquisition.

Revenue and pricing optimization (RPO) is not unfamiliar territory to these vendors, because they already have experience and solutions in demand planning, which at a more granular level is a core part of RPO. Additionally, the science is similar. Both supply-chain and RPO software rely heavily on complex algorithms to construct models with a vast number of variables and constraints, and to solve for the optimal result. And the development of supply-chain and revenue-management software appears to be following a similar line. In RPO, best-of-breed vendors that first produced point solutions now are beginning to broaden their footprint or become part of larger companies, though it will take some time for a fully integrated suite of products to emerge.

Manugistics aggressively is positioning itself to be the first with a solution that not only integrates its revenue products but links them to its proven supply-chain solutions, providing what it calls Enterprise Profit Optimization.

"Supply-chain solutions that traditionally operate in the cost arena are like having a really good defense," says Lori Mitchell-Keller, senior vice president of technology strategy. "When you start adding pricing on top of that, it's like having a really good offense. So if you don't have both features, you are getting either a good defense or offense, but not the whole solution. You need the whole solution to be really competitive."

Manugistics' RPO capability, which the company says can be used in conjunction with supply-chain solutions other than its own, largely was acquired through its 2000 purchase of Talus Solutions. Talus was created from the merger of Aeronomics and Decision Focus, two companies that had helped pioneer revenue or yield management in the airline and hospitality industries, where it has been successfully used for years.

"These two companies merged because they thought the new market was not going to be just about revenue management for the hospitality industry," says Mitchell-Keller. "The new market is about pricing optimization for all industries."

But Why Retail?
Retail is ripe for this technology for a number of reasons. At the strategic level, one driver is the acceleration of pricing complexity caused by the proliferation of products, channels and customized marketing. That, coupled with increased competition, razor-thin margins and a consumer base "trained" to buy only on sale, is forcing companies to make more precise decisions about what products to promote to which customers and at what price.

"It is just boiling down to where people are saying, we can't do this anymore without system support and the system support can't be stovepiped where the buyer does his thing and the marketing guy does his thing. It all has to tie together," says Peter Charness, senior vice president of global marketing and chief product officer at JDA.

At the tactical level the driver is the availability of powerful computers and of better data, both from point-of-sale systems and from data warehouses that aggregate and analyze consumer behavior.

"Until now this data was not being well used," says Augustin Manchon, Deloitte Consulting's head of precision retailing and price management. "Now all of a sudden this new thing is on the horizon that adds science to the art of merchandising and that gives you the ability to make fact-based decisions."

Getting the good, clean data these systems need is still an issue, however. "They are data hogs," says Steve Banker, director of supply-chain research at ARC Advisory Group. But, on the plus side, revenue management systems also are "living models," which allows them to get better with use. "Most folks are going to find that they have to begin with fairly simple models, but over time the models will become more complex and their predictive value will increase as the data quality improves and the data history they produce is incorporated," Banker says. "It's going to take them longer than they think to get up to speed on this and it will take them a while to grow the model fidelity, but over time there will be a huge payback."

"When people install our software they usually find there is a need for data that they have not historically been gathering," says Mitchell-Keller. "One of the things we do is help companies understand what process changes they need to undertake to capture that data."

Some needed data, for example, is outside a company's four walls, such as competitors' prices, and some involves the movement of products in relationship to one another - how a promotion on one product may cannibalize sales of another and boost sales of a third. Promoting one brand of hot dogs, for instance, may reduce the sale of other brands but increase the sale of hot dog buns.

"You really have to look at two products in relation to each other - elasticity not just by product but across categories," says Manchon. "Mathematically that is very hard to do. You had better have some Ph.D.s around to do that."

There clearly is no dearth of them working on revenue management at the companies interviewed for this article and at such demand-focused vendors as DemandTec, KhiMetrics and KSS Group. "When the internet bubble burst a lot of those great resources were reallocated to the retail niche," says Manchon. DemandTec, for example, has a patent pending on its parallel-computer platform that analyzes POS data to identify or infer buying patterns and to help retailers understand price and promotion sensitivities at the SKU and store level. Item and store level detail is more granular information than that provided by traditional demand forecasting systems, which were designed to aggregate demand for manufacturers.

Deloitte estimates that between 20 and 30 retailers have gone through some type of price optimization pilot while a new survey by the Professional Pricing Society reveals that 7 percent of 260 retailers polled are using price optimization tools. "That's a big number, higher than we expected," says Deborah Vollmer Dahlke, PPS board member and director of analyst and media relations. "Retailers are turning to these solutions because they have to. There is no way they can manually do the kind of 'what if' analysis needed in today's marketplace to understand what their prices should be."

"One of best selling points of this technology is that the results are very tangible," says Evan Gold, a consultant with LakeWest Group, with specializes in retail and consumer products businesses. "Top-level executives can see the increases to margin and profit lines. The big companies will jump on first and then there will be a swell of other retailers trying to do the same thing."

In a white paper on the subject, Gold predicts that revenue management will someday be applied to almost everything that is sold, "and will prove to be such a powerful competitive weapon that major firms will be living, and in many cases dying, according to RM algorithms."

Focus on Promotions
Promotions management is one of the first focus areas for RPO solutions. A specialization within that is circular optimization - determining which products to promote in which regions and at what price. Staples is implementing the Manugistics tool to help it decide what promotions it should run in its retail stores, while Best Buy is implementing a promotions solution it co-developed with i2.

This is another highly complex area because the goal is to enable stores to promote different products and/or prices in different regions. And promotion decisions depend on many factors. Charness of JDA notes that promotions management software needs to consider such elements as whether the vendor will collaboratively participate in the promotion and if so what return he expects. "Do I have product available to support what I expect to sell and do I have space in the store? If I can't display a product properly and reserve the space, will it all work? These are questions that need to be answered," he says. "When we tied the decision model together for what we are building as revenue management, we identified about eight to 10 facets that really have to line up or it is not worth running this product and this ad."

Manugistics emphasizes that its promotion management tools focus on profit lift as opposed to revenue lift. "Just because you sell 1,000 more of a particular stapler, for instance, doesn't mean you made any more money, because you may have cannibalized a more profitable stapler," says Mitchell-Keller. "When you run promotions, you need to understand different cannibalization effects as well as the effects of buying-forward dilution."

Buying-forward dilution was a problem for Ford, which realized that its new-car incentives merely shifted the timing of a purchase for many consumers, decreasing profits without generating new demand, Mitchell-Keller explains. Using the Manugistics solution, the car company was better able to target incentive offers to consumers that would not otherwise have been buyers. "What our software does is to model market segments and consumer buying behavior and the price elasticity associated with those to help Ford determine how best to promote its products," she says. Ford reports that the Manugistics promotion optimization solution added $800m annually to its bottom line, a result that includes savings in advertising costs.

In a different type of application, UPS uses the Manugistics product to help it understand what prices it should quote to particular clients during contract negotiations.

Markdowns a Concern
For retailers with a high proportion of short life cycle products that need to be cleared from the store by a certain date, markdowns are a key concern. ProfitLogic and Spotlight, leading vendors in markdown optimization, help retailers maximize margins by calculating how soon and how deeply to cut prices.

"The system creates a natural product lifecycle and on top of that it looks at seasonality and price elasticity and other variables," says Greg Girard, supply-chain and retail specialist at AMR Research. Such systems also should be able to consider constraints, such as the store's ability to execute markdowns, he adds. If a lot of labels have to be manually changed, then the number of markdowns may need to be limited. Also they must take into account individual business rules, such as taking no markdowns until a product has been on the floor for 30 days.

ProfitLogic is developing a full suite of what it calls Merchandising Optimization, but markdown software is where a lot of customers get started, says Julie Driscoll, director of marketing. "Retailers really need fact-based help in making what is often a psychological decision," she says. "If a merchant buys something, it is because he likes it and thinks it will sell. As you get into the season, you often find that buyers hold on, thinking a product that didn't sell well this week will pick up next week. Then they end up taking drastic decreases too late in the cycle."

The great thing about technology, she says, "is that it is not emotional or biased. It just takes an analytical view of predicted demand, it refines that view with each passing week and then it says, here is the best action to take."

"It may be better to take a deeper initial markdown and not have to come back as many times to take a further 10 percent off," says Greg Cudahy, executive vice president of pricing and revenue optimization at Manugistics. "Understanding that is one of the advantages of the science. Anytime we see someone using a rule of thumb or set percentage markdowns we know there is opportunity because it means they are not dealing with different segments of buyer behavior."

Driscoll stresses, however, that no system will ever replace a merchant's judgment. "This is important," she says. "What the technology does is allow merchants to focus on the exceptions where their judgment and experience is critical." It also frees them to spend more time managing category and market strategies to achieve enterprise goals.

Casual Male Corp., a leading retailer of clothing for big and tall men, recently decided to roll out ProfitLogic's markdown optimizer to all of its stores, after a pilot implementation improved both sell-through and GMROI (gross margin return on investment, a financial measure that captures both margin improvement and inventory productivity). Using the system, corporate buyers are able to centrally make local-level pricing decisions and Casual Male reports that they are more confident in their decision-making because of the analytic rigor associated with ProfitLogic's markdown recommendations.

"Our initial deployment validated the markdown optimization opportunity for Casual Male, and by expanding our use of the system we anticipate further improvements in gross margin, reductions in markdown budget, and more efficient clearance of merchandise," says Jay Scheiner, executive vice president and CIO.

The ProfitLogic application also has been successfully deployed at department store, mass merchant, chain drug store, and specialty segments of retail.

Better Allocation
Of course the best situation would be to have demand and supply so closely aligned there would be no need for markdowns. One step in that direction is to better allocate goods at the store level, which can be done with more granular forecasts.

"Retailers have typically anywhere from 20,000 to 50,000 products and may have 400 to 1,000-plus stores, and so understanding how their merchandise will behave at each location is a humanly impossible task," says Driscoll. "But if you apply these same analytical techniques to the planning and buying process and to the allocation, you can get close to the optimal distribution per store."

KB Toys turned to ProfitLogic's allocation tool in early 2000 for help in evaluating its method of distributing inventory across its 1,300 stores in four divisions. Bob Muller, vice president of inventory management, says the company had been categorizing stores into five ranges by revenue volume. A store's volume category, coupled with sales history and forecasts, determined how much of each product it received. "We provided ProfitLogic a significant amount of inventory and sales data for a one-year period and then asked them to crunch that data to give us a statistical perspective to either support the five volume codes we were using or give us a new recommendation," says Muller.

ProfitLogic recommended that KB expand its volume codes from five to nine and that it treat each division separately so that the volume ranges for KB's mall stores would be different than the volume ranges for its Toy Works stores, outlet stores and express stores.

"This was a very big change from using five codes across all our stores," says Muller, "but we could see very clearly from the data they presented that the recommendation would give us significant opportunities to reduce our overall inventory base in the field, and at the same time create a higher probability that a customer going into a store would find the product he was looking for."

KB worked on ProfitLogic's recommendations in the fourth quarter of 2000 and began implementation in the first quarter of 2001, completing it in Q2. "As part of this process, ProfitLogic also showed us that the volume code of stores could change from quarter to quarter," says Muller. "In the past if a store was coded as high volume, we treated it as high volume all the time."

KB now runs this analysis twice a year. "It has been really effective at choosing the right assortment to put into each location, which is especially important since we are a small-box retailer in terms of space," says Muller. "We don't have big Wal-Mart-sized stores, so we have to be very specific about what we put in each one. ProfitLogic has been instrumental in helping us do that."

Improving Shelf Pricing
A third pricing aspect of RPO is optimization of everyday shelf pricing in relation to local market conditions.

The first decision a retailer has to make is how it is going to price the highly visible, high demand items, those that are most susceptible to comparison shopping and, consequently, to promotions, says Deloitte's Manchon. "But at the other extreme you have the products that, frankly, are neglected by everybody - by the retailer, who doesn't change prices very often, and by the consumer, who doesn't buy them that often or who doesn't pay much attention because the price is relatively low." A good example, he says, is shoe polish, a product where price changes have little impact on demand. Understanding what consumers will pay for these blind products, he says, "allows you to build margin on them while building volume with the highly visible items."

Forrester Research cites a grocery chain that achieved a 16.9 percent net margin increase with DemandTec's Price Center software, which recommended an average 3.8 percent price hike on everyday items like canned beans.

Manchon says a store's products are like a pyramid. At the top are the 2 percent to 5 percent of high demand products that drive volume. "Below that you have a lot of SKUs, thousands of SKUs, that are much less elastic and that are not really managed well. Now, with this technology, instead of spending all your time managing that small percentage at the top, you can manage 100 percent of products in a store because a lot of those decisions become automated."

There are execution issues with constantly updating prices on all SKUs, but Manchon foresees a time in the not too distant future when electronic labeling will make it truly possible for stores to change prices, perhaps as often as several times a day, depending on store traffic. Another future scenario: customers swipe their loyalty cards as they enter a store and receive personalized sale coupons before they shop.

The real power from these revenue management and pricing optimization systems will come when they are fully integrated with supply-chain management, a path Manugistics is forging. Forrester predicts that widespread integration will not occur until 2006. In the meantime, solutions will add functionality and gain acceptance throughout the retail industry. Smaller vendors, says Forrester, will have one option: merge or die. Look for more acquisitions that combine supply-chain, revenue management and even activity-based costing companies.