At many large businesses, marketing and revenue management can work at cross-purposes. Both are focused on generating greater profits at a strategic level, but at a tactical level the definitions of success can be very different. I have seen marketers celebrate successfully selling out a loss leader while revenue managers watched their revenue per inventory unit goals slip out of reach. Alternatively, revenue managers can react to the demand surge due to a marketing campaign by making it impossible for anyone targeted by that campaign to buy at a competitive price. I got into trouble at a revenue management department party once for remarking that the fully loaded cost of the cake should include all the marketing dollars we had wasted.

Figure 1.

The conflict, and potential synergy, between digital marketing and revenue management arises because they control two different components of the marketing funnel. Digital marketing generates demand by creating awareness in a prospect population, arguing for consideration and delivering the customer with an effective call-to-action. The customer experience of revenue management comes after all that. Revenue management stimulates or restricts demand at the point of conversion by moving price and availability. It creates optimal pricing strategies that target conversion rates to available inventory.

This article investigates the synergy between revenue management, which maximizes profit generation by moving prices to moderate conversion rates, and digital marketing, which can drive demand to available capacity with a precision and responsiveness that is not achievable in the offline marketing world. It presents a couple of case studies from clients that have begun to capitalize on these synergies.

Figure 2.


What is Revenue Management?


Revenue management is an established business process that has significantly altered the travel and hospitality industry since its inception in the early 1980s. It requires analysts with detailed market knowledge and advanced computing systems that implement sophisticated mathematical techniques to thoroughly analyze market behavior and capture revenue opportunities.

Originally developed by the airlines as a response to deregulation, it has quickly spread to hotels, car rental firms, cruise lines, media and energy. Its effectiveness in generating incremental revenues from an existing operation and customer base has made it particularly attractive to business leaders who prefer to generate return from revenue growth and enhanced capability rather than downsizing and cost cutting.

What is Digital Marketing?


Digital marketing is a broad term encompassing any technique applied on the Internet to deliver a marketing message. Probably the most appropriate source for a definition is Wikipedia, which states: "Digital marketing is the practice of promoting products and services using digital distribution channels to reach consumers in a timely, relevant, personal and cost-effective manner."

This definition incorporates a number potential contact mechanisms including natural and sponsored links on search engine results, display advertisements, e-mail, and social media such as widgets and blogs. With the advent of Universal Search, the response to a search query can include video, images, news and local maps, as well as text listings. Search remains a remarkable channel for capturing customers that have raised their hand and demonstrated interest in your product.

Figure 3: A search results page (SERP) for a search on the keyword "Socks." The results along the right hand side and in the tan area across the top are paid search ads, sometimes called sponsored links. Advertisers bid to appear in these results on a cost per click (CPC) basis. The natural or algorithmic results in the white area to the left are based purely on Google's assessment of what Web content is most relevant to the search query.

When you type a search query into a search engine and hit enter you will be presented with a search engine results page (SERP). In general, you will see a combination of sponsored links, from which the Search Engine makes its money and natural results based on the engines assessment of the best match to your query. Management of sponsored links in search engines is referred to as paid search. Taking steps to make your Web site attractive to search engines is called search engine optimization or SEO.

Display advertising tends to be regarded as a higher funnel activity than search, although behavioral marketing techniques often target consumers with demonstrated interest as well. Online display can facilitate immediate gratification because, unlike television or billboard displays, consumers can often click on the display ad and purchase during the same browser session. In general display does not justify its value solely on these click-throughs. It also creates view-throughs — consumers that see a display ad and later navigate to a site to purchase.

E-mail is a highly controllable marketing tool and can be used to great effect to spike demand for promotions either in store or online. Because of prevailing attitudes and laws regarding spam, e-mail requires a more circumspect approach than other marketing tactics, and many firms limit its use to communication with existing customers.

Figure 4: Display ads, delivered across networks of sites, can have broad impact but are generally considered higher funnel activities.

A widget in a social media forum such as Facebook can tap into naturally occurring social contact to deliver a marketing message. Widgets have been very successful marketing entertainment-oriented products such as movies because clips can be shared in the widget.

The most effective digital marketing technique for integration with revenue management is one that allows demand creation in a precise and focused way. Since paid search campaigns can be switched on and off with a click and no ramp-up time, it is a natural first choice for revenue managers.

What is Paid Search?


Paid search or PPC (pay per click) marketing is an online demand creation and capture process that has a lot of features that are complementary to a revenue management program. Both are highly analytic endeavors requiring detailed forecasts of consumer behavior and, when sophisticated practitioners are involved, advanced mathematical optimization techniques. In both cases automated systems have been developed to manage the large number of tactical business decisions required to maximize profits. The accuracy and velocity of automated decision support is key to the success of both processes.

Paid search and revenue management are complementary because they occur at two different points of the marketing funnel. While paid search is generally used to capture already highly qualified leads, i.e. customers who are searching on relevant keywords, it can be used throughout the advertising process from awareness and consideration through to conversion. Different keywords are important for different stages of demand capture. By contrast, revenue management stimulates or restricts demand at the point of conversion by moving price and availability. It matches inventory levels to demand forecasts to develop optimal pricing strategies that target conversion rates to available inventory. In cases where inventory is flexible, the cost of substitution can also be incorporated.

Figure 5: Widgets are fun and can be extremely successful if they go viral. This widget for The Colbert Report packs in video content, membership, voting and e-commerce capability.

Revenue management is very effective in extracting revenues from strong markets. This often makes up for a key weakness: over-reaction and lack of precision when demand is soft. The response to soft market conditions can be to make price cuts across the board to stimulate demand. In some cases this is appropriate, but in many cases it is extremely expensive. Unlike marketing spend, the impact of price cuts do not show up as explicit expenses. However, price cuts can have a devastating impact on profitability.

When marketing and pricing are integrated, the cost of a price cut is weighed against the cost of driving extra business. In many areas of marketing practice, the level of granularity available to marketers is insufficient to create a clear understanding of the impact a specific investment on tactical pricing and conversion. Direct marketing strategies such as paid search do have the granularity to target specific timeframes and geographical locations to offset the need for price cuts. When this is combined with competitive monitoring, a clear picture emerges of where and when to deploy paid search spend and which products to discount. By using paid search in tandem with yield management, substantial revenue gains and marketing efficiencies can be realized.

Implementing an integrated program is a multi-stage process. Initial data integration for scheduled reporting can be used to bootstrap the business process changes needed. Once a business process is in place, it needs to be informed by analytic models that can calculate displacement costs, expected paid search return on investment and the intersection of the diminishing return curves of both practices. Finally, a decision-support system is required to fully incorporate the business expertise of revenue controllers and paid search marketers into the automated system recommendations.

Figure 6: This eye-tracking map from Enquiro [1] shows where searchers eyes most naturally gravitate. It matches very closely with our experience of click volumes by SERP location.

Financial Services Example


Dynamic ratemaking presents a huge opportunity and challenge for insurance companies. One unique feature of the insurance industry is the cost of goods sold, i.e. claims, is not incurred until long after the sale takes place. This has led to tight margins that can be doubled or halved by small percentage changes in the price of a policy. Another feature is how slowly rates are adjusted in the marketplace. Ratemaking involves evaluation of risk by the actuarial team, assessment of the competitive environment by a product manager and presentation of a compelling argument to the insurance commissioner. The result is a rate for any one customer that may be adjusted only once or twice a year. If you have five competitors in the market doing the same thing, then for five out of six months your price will be either too high to convert anyone or too low to capture the opportunity in the marketplace. Without adequate rate velocity, you spend 10 months out of 12 either losing money or losing opportunity.

In the early 2000s, I worked with GMAC Insurance to launch their e-commerce Web site. Since Google was not a dominant force at that time, most online advertising was focused on banner purchases. An early PPC (pay per click) model was available from Overture, which later became Yahoo Search Marketing. Performance of the marketing program was measured in cost per application (CPA), which incented the marketers to drive traffic from the highest converting sources. This had the effect of finding holes in our pricing, i.e. segments that had been underpriced. Fortunately we had the ability to cope with adverse selection because the site had been developed with built in revenue management capability.

Insurance is priced in a two-tiered way: underwriting and ratemaking. Underwriting splits prospects into risk tiers that are associated with a base rate. Ratemaking applies additional factors that are specific to the individual applying for insurance. We introduced sub-tiers into the underwriting process that adjusted the base rates in response to changes in the conversion rates in that tier. It functioned much the same as airline revenue management does, closing lower class availability when the booking pace gets too high. Marketing spend on ads was adjusted based on the new expected conversion rates from the revenue management system.

The result was a successful Web implementation that went from zero to $2 million per month in written premium in the first six months. As well as being the first implementation of online revenue management in the insurance industry it was the first implementation of online real-time underwriting. In the early 2000s, when you applied for a policy your quote was contingent on the company verifying your information and you could get a letter two weeks later with a modified premium. Because our application process verified much of the data in real time and because revenue management provided a backstop against adverse selection, we could be confident that we were making money on the business we were driving.

Figure 7: E-commerce sites have improved in leaps and bounds since the 1990s. The new Borders site is designed, according to Kevin Ertell, Borders vice president of e-business, "to make customers feel they're browsing one of the front tables at one of our stores" [2].

Retail Example


Retail price optimization has been given a boost by rich data capture available on e-commerce sites. A much clearer picture of shopping behavior and price sensitivity emerges from a retail Weblog than from monitoring in-store traffic. It is also much clearer to customers that prices are moving as Amazon found out with their DVD pricing test [3]. Shoppers tend to be more tolerant of markdown management, promotions and other pricing actions that are familiar to them from the offline world.

However, our initial foray into online retail revenue management focused on capacity management. Our client is a multi-channel BTC (business to consumer) and BTB (business to business) retailer who runs a top-five worldwide revenue-generating site. With more than 30,000 SKUs (stock keeping units) to manage, they were having trouble matching their search marketing campaign to stock availability. Instead of developing an internal system to monitor inventory, we set up a "spider" to check the status of each item. A spider is a computer program that reads Web pages automatically. Our spider was called the URL validator because it was originally intended to check that clicking on a paid search ad did actually bring you to a functioning page. It was a natural next step to ad functionality to check if the item was in stock. Regular reports were generated, and the retailer began to make adjustments to availability levels.

But it didn't stop there. The normal revenue management reaction to items that stock out due to high conversion rates is to raise the price. Instead of changing the price on already profitable goods, the retailer chose a different strategy. One way to drive increased traffic in paid search is to increase your bid on the search engines. This will lead to your ad being displayed more prominently in the search results. Another way is to alter your creative, i.e. the words used in the ad to entice the customer. We decided to change the creative for items that were converting well to include the price. The result was a 100-percent increase in conversion rate with no noticeable loss in traffic volume. Instead of trying to extract the most revenue out of limited capacity, we increased stock levels and generated far more value from effectively communicating a great value proposition for the consumer.

Hospitality Example


Revenue management has been central to profitability in the hospitality industry for years. Travel and hospitality providers tend to be less effective at search marketing and have ceded much of their customer base to online travel agencies. We initially were asked to address an online marketing problem for a large hospitality provider that depended almost exclusively on its call center to process reservations. The job of online marketing was to drive people to the phone or else collect leads for outbound calls. The challenge to managing a campaign like this is attribution of value to the online activity.

Simply setting up a unique tracking code that connected a phone call back to the online activity allowed for significant gains in search marketing performance. It was complemented with a reorganization of the search campaign along geographic lines. Areas that prove sensitive to online demand stimulation get extra funding. As we refine our understanding of click-through and conversion rates by geography, we are able to adjust our bid amount, and hence our traffic levels, to target vacancies that would have otherwise required price cuts to stimulate demand.

Summary


I have been very lucky to land in an industry that is booming. Although this article ostensibly touts the opportunity for integration of business processes, I hope it also is clear that operations research professionals are the right people to make it happen. Digital marketing has a bright future, and I have found it a very welcoming place for bearers of analytic gifts.



Kevin Geraghty (kgeraghty@360i.com) is vice president of research and analytics for 360i, an independent digital marketing agency based in Atlanta. A graduate of the O.R. program at University College, Dublin, Ireland, in 1985, he was a co-author of the Edelman Award finalist paper, "Revenue Management Saves National Car Rental," in 1996 and subsequently founded Revenue Research, Inc., a revenue management consulting firm.

1. Eye Tracking Report, Enquiro, November 2006.
2. Kevin Ertell, Borders vice president of e-business, Publishers Weekly, May 27, 2008.
3. "Varies Prices of Identical Items for Test," Wall Street Journal, Sept. 7, 2000.

Table of Contents

OR/MS Today Home Page


OR/MS Today copyright © 2008 by the Institute for Operations Research and the Management Sciences. All rights reserved.


Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com


Web Site © Copyright 2008 by Lionheart Publishing, Inc. All rights reserved.