Chapter 7
Scientific Bundling and Tiering
A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.
—MAX PLANCK, C. 1949
Few pricing and product tasks are more important than combining multiple products and services into a single offer—a practice known as bundling. Technology and cost-cutting are creating a world in which many customers now prefer such integrated solutions, and this demand must be met in order for companies to stay competitive. Also, bundling is a way to rapidly create a unique offer and capture the benefits of differentiation.
Few pricing exercises are more context-sensitive than bundling. This is because bundling in fact poses a direct question to the potential buyer: “Do you want to buy these products in combination?” Then buyers ask themselves whether they really need all the bundle elements now; if the answer is only “maybe,” what reward would they need for expanding their purchase?
What Constitutes a Bundle?
While many managers think of their market offer as a single product or service, often it is really a combination of products and services. Every market offer is a bundle. Sometimes a rigorous listing of components may surprise product management. For instance, the product management of a leading implantable contraceptive device said with assurance that no bundling was possible—or legal—because of FDA regulations. Yet, in fact, the offer on the market was a bundle of many services, each of which could be tweaked to optimize the bundle for differences among ob-gyn buyer segments.
When considering a product, note that very rarely is the product available without support, without financing or payment terms, without delivery, without packaging, without guarantees, etc. Thus, the birth control device actually comprised many components, both tangible and intangible, as seen in the decomposition of a single “product” shown in Figure 7-1.
Figure 7-1 Contraceptive device: almost all products and services are bundles and offer management choice.
The primary advantage in understanding the components is that it expands the discussion and the strategic choices. If you regard a bundle as inseparable, then it is much more difficult to adapt price to context or segment. So thinking carefully about bundling often moves the discussion from merely a price level question to one which provides sellers with more flexible market tools.
For instance, in the case of the contraceptive there could be many economically important variations. While the device itself could not be varied for different buyer segments, other parts of the bundle could be—like the box. To discriminate between low-volume buyers and high-volume buyers, the manufacturer could make the box very big or small. Big is bad because it fills up storage closets. For high-volume buyers, only the first boxes might be large and the remaining ones could be small—a rationale for the box sizes would be that the first few boxes need voluminous product documentation, while later boxes might have less. It’s a perfect excuse to volume price-discriminate. The same approach applies to shipping, samples, benefit verification, copay levels, and other elements of the “bundle.” While marketing may think it is selling a unitary good, it’s actually a bundle—and that allows much better price targeting.
Almost every offer is a bundle when you consider it carefully.
Pricing itself can also lump together previously stand-alone sales into bundles. For instance, a leading supplier of floorboards for residential housing understood it had a range of ways to sell the product to developers. Traditionally, boards were sold individually or by the palette, FOB.1
They could also be sold by the truckload, by the development (an “all-you-can-eat” contract specific to a project), or even through a requirements contract for a developer for a period of time.
The pricing dynamics change with progressively larger bundles. When buying by the palette, the seller found that specials at the local Home Depot would eat into his sales. Worse yet, the company’s reputation was at risk when inferior substitute boards underperformed. When sold on a requirements basis, there was no builder benefit to buying cheap boards anywhere—the incremental cost of boards from the manufacturer was zero!
Bundle Components Should Be Complementary
Bundles are held together by what we call glue.2 The glue can take the form of price, product integration, promotion, or channel. For instance, travel involving airplanes and hotels can be bundled along all four dimensions through travel agencies: one price, one trip, promoted jointly and sold jointly (a “hard” bundle). On the other hand, the bundle can be much more limited: after you have made your (stand-alone) flight reservation with an airline, you might be asked “Will you need a rental car with that? May we transfer you to a rental car company?” In that case, the only link is channel, so less glue (a “soft” bundle).
In general, bundles using less glue are lower risk. Success rates for hard bundles are surprisingly low. This is because buyers will need to see the benefits to a bundle buy, the bundle needs to be priced right, and the bundle actually needs to work. The latter issue is sometimes underestimated, but remember that the combination of car + car radio was attempted for many years with no success until a small company called Motorola developed a filter for screening out the electronic noise of the car engine from the radio reception.
In that case, the core (primary driver) of the car + radio bundle was certainly the car. Experience has shown that the vast majority of bundles have a structure, or taxonomy, which includes a core (the primary component motivating the purchase) and add-on components. Because buyers know they would have bought the core element anyhow, they tend to look for discounts on the add-on elements.3 If those elements do not appear to have an adequate discount, then the bundle will fail.
For practical bundling purposes, most bundles can be analyzed based on a core + add-on taxonomy. Figure 7-2 shows such a structure, in this case looking at a cable TV package.
Figure 7-2 Taxonomy of a video + broadband + telephony bundle.
In this illustration, the core is plain old (“linear”) video programming. Most of the add-ons, such as premium channels, are commonly seen as complementary to the core, and so they’re favored by buyers with only modest discounts.
For pricers, frequently it’s the non-complementary bundle elements that are important. In the illustrated figure, a possible non-complement is landline cable telephony. Cable telephony (VoIP) is not seen as a good add-in by some (high-end) buyer segments because they don’t trust telephone service from a cable company—for them it may be an unwanted component. High-speed data today is seen as a complement to video, but that may change if television over the Internet (IPTV), like Netfl ix or Hulu, becomes a leading choice. Then video and high-speed data may morph from complements to substitutes, as many customers may conclude they need only the data service.
Almost all bundles have a core component that drives a segment’s purchase. To design and price bundles, you must identify that core.
How can prospective bundlers estimate how a buyer will evaluate a bundle? One approach is to examine the stand-alone purchase patterns of potential bundle component candidates for grouping into a bundle. The first question is: are these candidates complements or substitutes?
Bundles should contain only complements. You need to ensure that only bundle elements that go together are offered together. If substitutes are combined, rational buyers will see that they are being asked to pay for redundant components. That kind of offer will either be rejected by the market or severely penalized with a higher bundle discount. This is why successful bundles include pickup trucks + trailer hitches (complements) but not rail tickets + airline tickets (usually substitutes.)
Surprisingly, intuition is not always adequate to evaluate whether two potential bundle components are complements or substitutes. Verizon has offered bundles of optional calling features that have included both the caller identification (CID) and the return call (*69) features. While they might, on the surface, appear to be complements (one lets you know who is calling, the other lets you dial earlier callers by hitting *69), in fact they are used in very similar ways and are viewed as substitutes by potential buyers. This became apparent to Verizon when bundled offers containing both products tended to achieve much lower marketing yields.
How to bundle “like with like” operates from the customer point of view. For instance, usually department stores promote like with like clothing by economic tier. A mannequin will typically have a dress, blouse, and handbag from the same strata of fashion. When a store foolishly put a $500 hat in the same display as $100 hats, the expensive hat rapidly became shop-worn as visitors handled it, but did not buy it. That promotional display mistakenly bundled substitutes, not complements.4
The measure that indicates whether different components are complements or substitutes is correlation. Correlation will show whether different bundle candidates tend to sell together, or if one tends to cannibalize the other. When increases of sales of product A are associated with increases of sales of product B, it’s a complement. This will manifest itself as a positive correlation.5
Note that correlation of sales, not percentage cross-sell, is the more reliable measure of bundle complements: you can have positive cross-sell but no real complements if you have a strong sales program. In other words, at the cost of goodwill and price pressure, a powerful seller can force suboptimum bundles on its customers. This is why frequently the effectiveness of bundles needs to be measured over time. If a bundle is associated with a decrease in market share or an increase in discounting, it’s a bad bundle.
Bundles should be comprised only of complements, not substitutes.
What Is the Value of a Bundle?
It’s widely known that bundles tend to sell at a discount off the sum of the stand-alone bundle component prices. If you sell a software program + training as a bundle, the combined price is likely to be less than the sum of each sold separately. One reason for this is because not everyone wants a bundle, and in some cases many of the buyers are unsure that they will take advantage of both components. Therefore, buyers must be offered a discount to buy a bundle. This means that through a form of alchemy, the market value of some bundle components instantly change when they go from being stand-alone sales to being part of a bundle—the context has changed.
That discount may be a reasonable cost for sellers. Your company, may welcome exchanging a 20 percent discount in return for channel efficiencies or reach. How do you calculate the bundle discount? In certain circumstances, use of correlation will also allow a measure of the bundled value of a bundle component to the market. Generally, the value of a given bundle component can be calculated based on the formula:
Value of the bundled component = Stand-alone value of the component × Correlation of component to core component of bundle
Thus, for instance, if the correlation between regular cola and diet cola is 0.9, each is priced at $9 per gallon when sold stand-alone, and regular cola is the “core” of this bundle, then a bundle of the two should be:
Baseline price of the bundle = Regular cola @ $9 per gallon ($9 × 1.0) + Diet cola @ $9 × 0.9 = $9 + ($9× 0.9) = $17.10 per bundle (2 gallons)
This logic is the basis for scientific bundle pricing. It will weed out negative contribution additions to a bundle. Adding together all the component values gives you a first cut at a market price for the bundle.
The market value of some bundle components instantly changes when they go from being stand-alone sales to being part of a bundle. There are scientific ways to estimate the new value of a component after it is included in a bundle offer.
Bundling Mismatches and Mistakes
Finding new ways to bundle your products and services is very worthwhile. Bundles can increase share, reduce loss, and improve efficiency. Yet pitfalls lurk because the more novel the bundle, the greater the danger of bundling missteps. There are two dangers in particular: (1) mismatching the various elements of a bundle and (2) mispricing the offer. If a bundle includes elements no one really wants, customers won’t buy the bundle without a deep discount. Similarly, if a company combines products or services without enough thought as to price, the result is often gross overpricing or underpricing.
In helping clients with their bundles, we find there are four marketing mistakes which typically lead to mismatches and mispricing:
Failing to create bundles for special purposes
Making bundles too big
Using tiering instead of bundles
Failing to innovate on bundle definitions
Sometimes mistakes are obvious, but often, the more successful the core product, the more its supporting bundle components can fall short of the optimum while appearing to be successful (i.e., “a rising tide floats all boats”—even unseaworthy ones).6 We’ll examine each category of mistake, in turn, showing how things can go wrong and what to do about it.
Failing to Create Bundles for Special Purposes
Bundles can have very different objectives. Bundle missions can include reducing customer churn, displacing competitors, rendering competitor pricing irrelevant, increasing channel efficiency, or deterring aggressive price behavior by upstarts. Such differing objectives typically demand different bundle structures; a company may even wish to have a suite of bundles available for specific competitive situations. We’ll look at two special purposes in particular: bundles intended for attack and bundles intended for defense.
Attack Bundles. An attack bundle seeks to displace a competitor; to achieve success the bundle must better the incumbent vendor’s price/ value proposition, typically via a lower price. The price must naturally account for switching costs as well, but still preserve long-term profitability. This can be done by crafting better and leaner bundles for core users, while often the incumbent is distracted trying to expand bundles to attract marginal (additional) users.
For instance, health insurers have added on a number of features to make their coverage satisfy all requirements. An interesting example is “gastric bypass” surgery for insured people who wish to lose weight and find that they cannot stem their eating. This expensive surgery shortens the intestines and so cuts down on food absorption. While a few of the insured will find this feature useful, to many it will be quite unnecessary. This means that for most market segments, a new insurance offer could drop gastric bypass without much notice. Similarly, a new auto policy might cut towing services or legal support without much notice or a notebook computer system might offer only minimal battery capacity. Each of these measures would lower costs and allow a lower price. A lower price is the opening wedge for competitive entry in many accounts. Giving core users what they want, while the incumbent is addressing marginal (but perhaps vocal) concerns often is a successful entry strategy.
Note that this lower price is not a discount; it is merely offering less. The rationale for this offer approach is twofold:
Many users may never need the added features, but will appreciate the lower price.
Those who will need these services can add them back at a later date.
An attack bundle is usually focused on the core component and may strip out low-value “add-on” bundle components.
Defense Bundles. Another type of bundle is defensive in nature and typically employed by market leaders and incumbents. Very often, market leaders must fend off upstarts; in doing so, they have a wider range of options than they typically realize. One such option—which we call the Bear Hug—is especially useful for discouraging upstart entrants from continuing to offer lowball prices to customers.
The objective of the Bear Hug is to devalue competitor offers. For instance, in a corporate governance software infrastructure market where companies typically purchased a dozen or so core services, the incumbent offered all the services. The upstart’s strategy involved offering fewer services—only three core offers—at lowball prices.
Where the upstart had made inroads, the incumbent successfully defended itself by offering customers an “all-you-can-eat” bundle. This customized bundle consisted of all services currently being purchased by that account—but at a price equal to the full market rate for the relevant products not overlapping with the competitor’s product set, plus a bargain rate on the overlapping products offered by the upstart. In other words, if a major customer buys products A, B, C, and D from your company, and purchases product E from a competitor, the new price would include all five (A, B, C, D and E) for the old price of A, B, C and D.
A defense bundle invoice to customers showed below-competitive pricing for overlapping services. That bundle was hard to resist, and revenues steadily increased. Because the offer was sold selectively to complex larger accounts that were already splitting their purchases between the incumbent and the upstart, it was immune to comparisons by other customers; there was no danger of a general price meltdown. The pricing sent a strong message to the upstart challenger not to poach on existing customer relationships. Most important, the upstart could not retaliate because the bundle product scope was wider than its own offers.
Many companies react to low-price competitive offers by tieing a weaker product with one stronger product. This is bad practice. Unlike the comprehensive Bear Hugs, such bundles tend to be mispriced because of poor component fit and/or the ratio of weak to strong products is often too high. If the bundle price does not accurately reflect the value of bundle components, then—like a poorly fitting shoe—it will start to chafe. Customers will tire of the forced marriage between strong and weak bundle components, and market share will begin to erode.
One example comes from the high-technology arena, where a producer of optimization software for computers, Server Farms, bundled in high-priced support and maintenance charges as well. This plan worked well for a few years, but as soon as rivals matched the software, cancellations rose because the maintenance was overpriced and customers resented having to buy both. Ironically, the software company knew of the growing problem but held on to the lucrative maintenance in order to reach initial public offering (IPO) revenue targets. Today the company is an also-ran in that market space.
Ad hoc bundling to defend a troubled product usually fails.
Making Bundles Too Big
An example of overbundling comes from an information service that garnered a premium position in the early days of its market, pioneering the conversion of forms and print to electronic. The service constructed its bundles around federal materials and grew by adding such further legal jurisdictional categories as states, international, and arbitration.
The jurisdiction-based growth of the bundle made sense in the short term—each additional element added incremental revenues—yet the increments became smaller and smaller each time, particularly as niche competitors established offers in more categories. Soon the sales force began discounting heavily, unable to sell the value of the incremental components.
The solution to this situation was to break up the legacy bundle to match component values to user needs and reprice accordingly. In fact, when this information service provider split its comprehensive bundle into functional sections, it obtained price uplifts for virtually all customers.
Why did this happen? Because although the individual smaller bundles had lower list prices than the larger bundle, most customers needed multiple new bundles—and the smaller, more relevant bundles carried much higher value. The internal consistency of each specialized bundle was much higher than that of the old, fat, flabby bundle. It turned out that almost no customer had needed most of the big bundle, which is why they’d negotiated so fiercely to reduce its price.
Another compelling example of overbundling comes from the cable television industry. Although subscribers welcomed the first few waves of channel expansions, enthusiasm dimmed as packages went far beyond the number of channels they would ever watch. A much better approach would be to sell narrower, specialized packages, each extracting value from particular audiences. However, management is often reluctant to create specialized bundles, citing operational and cost constraints. Also, without the ability to actually calculate component added value, many managers ignore market evidence that “less is more.”
Packing too many components into a bundle—overbundling—is the most common mistake made by product management at incumbents.
Sometimes operational bundle development costs can be minimized. In many cases when seeking to narrow a bundle, there’s no need to touch the product. Often limiting the breadth of the bundle is less important than narrowing the product’s value message and price plans. In many cases, customers won’t notice they’re getting “more” than advertised—and even if they do, complaints are generally few. In other words, actual changes in the product can be minimal: the work is in the message and price, not the actual product. In B2C software, this is particularly easy—just leave components off the menu; you need not delete actual code.
Using Tiering Instead of Bundles
Hierarchical tiering offers is a popular pricing technique. It’s actually a form of bundling, but it differs from most bundling initiatives in this respect: rather than build from the ground up, tiering begins with an existing offer, and then adds or strips out features to achieve the end result. Sadly, such additions or subtractions tend to be done blindly.
The value principles are the same as any bundle: try to build a compelling value price proposition for a particular market segment. The purpose is usually to sell to a new tier of customer, but there is another purpose to tiering. Sometimes the purpose of tiering is not to sell all the tiers but to show consumers that one of the tiers is a good deal through a tiered context reinforcing the message that they are getting a good deal.
Tiering works quite well when consumers feel they have discretionary spending ability—that is, they can spend more on the features they find appealing (up-tiering). Tiering works less well with buyers faced with highly constrained budgets, who want the lowest price that meets their essential needs.
When customers are reluctant to spend more, a second contextual technique is worth investigating: linking bundles so that the purchase of one bundle must lead to the purchase of a second bundle. Up-tiering in this manner is a good way to increase revenues either at the time of the initial purchase or later on. The underlying insights are twofold: first, customers will likely experience moments after the initial purchase when added features or bundle elements seem more appealing; second, customers’ price sensitivities will vary over time. What seemed unaffordable one day will not be on another. Linked bundles directly help sellers profit from the happy moment when needs are high or price sensitivities are low.
An example of such contextual linking over time comes from the dental supply industry. Here suppliers have found that entry is easy with disposables, such as sponges and tongue depressors. Once that relationship is established, hygienist handpieces can be more easily sold. From there surgical handpieces, cosmetics, and cameras are logical next steps in the growth of purchases from dental offices.
Hooks. For this technique to work, structures that we call “hooks” are necessary to encourage customers to voluntarily upgrade. We have found this to be especially true in tough, mature markets. Here are some examples of hooks:
In the construction industry, many building contractors price their initial bids low but charge high markups for change orders in the midst of construction—when owners have little choice but to pay higher prices.
LexisNexis sells legal libraries in packages by tribunal and/or geographic region—yet its search mechanism highlights relevant cases in the total legal database, not just by region. Therefore, even if an attorney steadfastly refuses to buy more than the library of his own state’s decisions, he will keep seeing searches that show all cases, from any tribunal. On the eve of an important trial, if the search engine now shows a case outside his library that appears highly relevant, even a very stubborn buyer may decide at that moment to expand his purchase. This factor has led many users of LexisNexis to buy new libraries or to buy specific cases, last minute and undis-counted, at several times the normal subscription price.
A great consumer example of a hooking mechanism comes from the car rental business. Recently at Los Angeles International Airport (LAX), a sign at the Hertz rental counter explained the following provisions for gasoline:
Gasoline in the vicinity of LAX: $4.29/gallon
Gasoline from Hertz prepaid: $4.04/gallon
Gasoline from Hertz if you return the car less than full: $9.29/gallon
So, the hook is very clear: if you don’t care about expense or are in a hurry when you return the car, Hertz will collect a premium. If you think that is unfair, you have a chance to avoid the charge since Hertz is offering you a bargain rate to begin with. If you turn the bargain rate down, don’t complain later!
Firestone Tire dealers offer tire warranties that exceed the likely usable mileage of its tires. Why? Because the pro rata warranty rebate is well worth it to get former customers back at the dealership at the moment they need new tires.
Migration paths prompted by customers’ key needs—the hooks—can be far more effective than price structures founded only on a hierarchy of good, better, and best.
Where buyers are reluctant to spend, hooking works better than hierarchical tiering.
Failing to Innovate on Bundle Definitions
Back to incumbents: another common problem in the bundles of established players is that the bundle definitions no longer effectively communicate value, especially when compared with competitors’ offerings: everyone is now claiming the same benefits. Worse yet, the market leader may even find that its definitions are being used against it.
For example, a company in the medical-testing market initially grew to prominence by offering services in more cities than did its competitors. This claim was backed by well-equipped facilities in more than 40 states. When competitors arrived on the scene, however, through reciprocal agreements, they laid claim to all 50 states! Even though the competitors’ capabilities and quality were inconsistent, customers found the pitch convincing—after all, they had been trained by the incumbent itself to believe that geographic coverage was an important part of the buying decision.
The solution for the incumbent was to redefine its product and message. Promoting depth of equipment, not merely geography, plus using service bundles and pricing to highlight its superior capabilities began to address the problem. Message underscored the low value of mere geographic presence without appropriate equipment and quality.
Depth versus Breadth. One way of describing this idea is that it focuses on depth rather than breadth. Breadth is how customers assess the initial value of an offer; depth comprises the ineffable attributes that customers discover only after experience with the product. This notion applies to any industry: there are always attributes that can’t be easily proven or sold up front but must instead be experienced to be believed.
As an example, one transportation concern buys only high-quality Peterbilt trucks, after having discovered that their drivers treat Peterbilts better, with subsequent lower operating costs. This would be tough to prove and quantify ahead of time, but it is now factored into financial decisions. Similarly, television advertisers generally demand evidence or direct experience to assess the effectiveness of media with above-average advertising rates per thousand (CPM). It is the actual propensity of different audiences to buy that is the best evidence. After trial or observing other’s results, advertisers will pay premium CPM rates if warranted.
Depth versus breadth is often key in fights between incumbents and challengers. Challengers are typically more innovative—partly because they need to be creative to succeed and partly because they are unconstrained by contracts, tradition, or older computer systems. The classic example of depth versus breadth was the fight between upstart long-distance telephony carrier MCI and incumbent AT&T during the 1980s. AT&T defined a minute of conversation as 60 seconds of talk time; MCI defined it as setup time (ringing, etc.) and talk time. Effectively, this meant that MCI had a 57-second minute—and this helped it quote perminute rates lower than AT&T’s.
For challengers, the joy of redefining depth includes the joy of causing competitor managements to throw temper tantrums—but most important, it places a big burden of explanation on the incumbent. Often this burden is insurmountable: for one thing, customers don’t care about long technical explanations. It may be better for the incumbent not to acknowledge the upstart at all.
Summary
Bundles are multipurpose instruments that offer sellers many advantages, including the following benefits:
A bundle may be “stickier” to buyers, and so it may reduce customer defections and churn. Telephony vendors found that when they tied local service with long distance, their churn rates fell: it was more daunting for subscribers to switch their local service—concerns about repairs and other service issues reduced overall churn.
There may be economies of product scope, either in selling (channel synergies) or the product. For instance, AOL found that including diagnostic applications with its basic service appealed to subscribers and actually reduced overall help-desk service costs.
Bundling may allow the sale of incremental products to groups with diverse product preferences. Thus, if one group places a high value on pizza and the other places a high value on beer, a beer + pizza bundle may sell incremental pizza and beer. Effectively, the bundle combines the value of each and can extract extra value.7
Thus, there are many fundamental reasons for companies to bundle, and it is worth considerable managerial effort to get it right. Actually, your company may have little choice to bundle and to get bundling right. Context is the pillar of any bundle pricing.
But bundles are high-risk undertakings. Particularly when scientific bundling analysis is not used, the failure rates are high. For this reason, try soft bundles with relatively less glue (linkages among components) and use the right bundle design techniques—use of correlation and clear bundle missions.
Notes
1. FOB stands for free off board. In other words, the buyer pays the freight.
2. Not a reference to the binding material in the floorboards! Rather a reference to bundle linkage approaches.
3. This can be demonstrated by the sales success rates of different bundles to different segments at different prices. Where there are segments with different cores and there are big differences in stand-alone bundle component prices, then mathematically you can isolate the discount implied by the market that is on the addon components. For example, a bundle of video (at $60 per month) + phone (at $30 per month) will sell better to the video segment than to the phone-oriented segment.
4. Comments by Ursula H. Moran, top-rated specialty retail analyst, formerly with Sanford C. Bernstein.
5. J. Adams and J. Yellen, “Commodity Bundling and the Burden of Monopoly,” Quarterly Journal of Economics, vol. 90, No. 3, August 1976, p. 486. This is a seminal work in this area, but full of numbers. To avoid a long discussion of correlation, we note only that the nice thing about correlation is that it has no preconceptions: it just reports relationships. For instance, it turns out that a slumping economy is correlated to fewer auto fatalities. Who would have expected that, a priori? As the economy slumped in 2005 to 2009, so did car-crash counts. Two researchers at the University of Michigan suggest that during a slump, drivers slow down, and that leads to fewer accidents. “New Puzzle: Why Fewer Are Killed in Car Crashes,” The Wall Street Journal, December 15, 2010, p. D1 and p. D3. Similarly, the “open-minded” nature of correlation is relevant to bundles because it allows management to test creative new bundles before they get to the market.
6. The penalty for a bad structure is best measured relative to other structures. Similarly, bad bundles will underindex compared with good ones, and if management has been giving discounts, larger discounts.
7. This concept can be shown to generally increase overall revenues, provided that bundle discounts do not offset the gains to the combination. In simple terms, combining two goods A and B (in a “hard bundle”) means that money left on the table because some consumers would have paid more for A can be used to pay for B, and vice versa. See http://en.wikipedia.org/wiki/Pricing_of _Bundles_and_Packages.