CI 7 Business Models

A business model is a compact account of how a company actually works: whom it serves, what it sells, how buyers are reached, what it charges, and which assets and partners keep the whole thing running. Reading it well is the starting point for judging a firm’s strategy, its prospects, and the risks built into how it earns money. Many issuers run familiar models that have existed for generations, such as manufacturers, wholesalers, retailers, banks, and insurers; for these, success comes less from novelty than from execution, skill, scale, brand strength, proprietary technology, or economies of scope. Other issuers combine models, or build something genuinely new, often on the back of digital innovation. Because management can reshape a model or add new lines of business with distinct models inside the same company, an analyst should treat this as an ongoing assessment and should not simply accept the firm’s own account at face value.

There is no single agreed definition, but a workable business model answers a short list of questions:

• Who are the target customers, and what market do they form?
• What product or service does the firm offer, and often why do buyers want it?
• Where does the firm sell, and how do its goods and services reach buyers?
• How much does it charge relative to rivals?
• How is the firm organized, and which key assets, partners, and suppliers does it depend on?

The description should carry enough detail to make the moving parts and their relationships clear, without swelling into a full business plan. A complete picture also links these choices to the firm’s revenue model, cost structure, asset profile, financing, and ultimately its profitability.

Reading the model from disclosures

Public issuers usually reveal enough in annual reports and investor presentations to reconstruct their model, and management may state it outright. Tesla’s annual report opens with such a description, from which an analyst can read off several elements directly.

The customers and the market (“who”)

A model should name the customers it aims at. Are there specific segments? Are the buyers other firms (a business-to-business, or B2B, setup) or individuals (a business-to-consumer, or B2C, setup)? Which geographies are served? In consumer markets the reflex is to think in demographic terms, for instance well-off suburban households, but useful segmentation is often far narrower and specific to the category, for example affluent early adopters of technology whose homes support plug-in charging in countries that subsidize electric vehicles. Sophisticated analytics let firms drill down to the postal-code level and group buyers by social group, life stage, and wealth, then recombine those cohorts across regions into larger, similar markets that targeted digital advertising can reach efficiently. Opportunities frequently appear precisely because incumbents fail to serve, or even notice, a segment; at the same time, the choice of target customers can bring its own risks, such as high entry barriers, shifting segments, or intensifying competition.

The product or service (“what” and often “why”)

The model should state precisely what the firm provides and where, if anywhere, it diverges from rival offerings, framed by the needs of the target customer. Doing so helps size the addressable market and surface risks such as imitation or substitution. A sharper way in is to ask what job the customer is hiring the product to do. Trucks and railroads look nothing alike, yet both meet one need, moving goods; the interesting question is which customers, in which situations, pick road over rail (or air, or sea). Business buyers tend to have clear, quantifiable needs, buy in a sophisticated way, and are driven by profit. Consumer needs are more varied and sometimes impossible to pin down: some buyers want the lowest cost per mile, others want styling, performance, or low emissions, and advertising and placement can create such wants in the first place. Firms should describe offerings precisely; vague, sweeping claims about addressable markets, overstated differentiation, or thinly developed “platforms” are often attempts to persuade investors that everyone is a customer, and they deserve scrutiny.

Channels (“where”)

Channel strategy is about where the firm sells and how its offering reaches buyers. It usually splits into two jobs: selling and marketing, and delivering the goods (distribution or logistics). A key question is which functions the firm should keep in-house and which are better left to partners and suppliers; many software vendors, for instance, leave installation to third-party consultants who then provide ongoing services. For physical goods, the traditional channel is a chain of intermediaries, each with its own facilities, buying and reselling the product until it reaches the final buyer. A direct sales strategy skips those intermediaries and sells straight to the end customer; it is long established for complex or high-margin items: medical devices, heavy industrial equipment, and luxury goods; and it suits B2B settings where the pool of buyers is small and easy to reach. Direct selling once meant a large sales force, but digital selling increasingly supplements or replaces that. Channels are often mixed: an omnichannel approach, common in apparel and packaged goods, blends digital and physical steps, so a buyer might order online and collect in store, or choose in store and have the item delivered.

Adidas, the world’s second largest sportswear brand, illustrates a deliberately blended mix. Its 2019 annual report counts over 2,500 of its own retail stores, over 15,000 mono-branded franchise outlets, and over 150,000 wholesale doors, plus an e-commerce channel it calls its single biggest store, live in more than 40 countries. Channels are not a detail: they drive revenue, cost behavior, profitability, and risk exposure. Direct sales bring a close customer relationship and keep the product off a shelf next to rivals, but they carry heavy fixed costs in salaries and benefits. How a firm’s channels differ from competitors’ matters too: Tesla sells direct, whereas US automakers sell through franchised dealers, which complicates any comparison of their financial results.

Pricing (“how much”)

The model should make the pricing logic plain. Does the firm sit above, level with, or below rivals on price; how do prices compare with costs; and what justifies the level chosen? Where competitors are many and products barely differ, sellers are price takers who accept whatever the market sets; these are commodity producers, and examples include oil and gas, off-patent drugs, and many loans. Where competitors are few or products are highly differentiated, a firm has pricing power: the ability to hold prices above cost, or raise them, without losing demand. A firm without pricing power usually chases cost leadership, aiming to be the lowest-cost producer, as with discount retailers Costco and Lidl or the low-cost airline Ryanair. Premium prices can be defended by pointing to savings elsewhere: Tesla prices at a premium but stresses a lower total cost of ownership (the full direct and indirect cost of owning an asset across its life), which, after subsidies and cheaper charging, narrows the gap, though subsidies could shrink or vanish. Often, though, premium prices rest not on such arithmetic but on brand, aesthetics, and intangibles that customers simply value, as with most luxury goods.

A pricing model sets out how much a customer is billed for a unit of product or service. Because a unit can be defined in many ways, and prices can vary by grade, volume, channel, customer type, or unit cost, there are many models. Charging different prices to buyers with different willingness or ability to pay is price discrimination, and its aim is to lift profit by capturing more of what each segment will bear; serving some segments, such as high-volume buyers, may also cost less. Rather than move published list prices, firms often adjust the real, or net, price through selective discounts, promotions, and bundles.

Pricing for multiple or complex products

• Bundling combines products or services so buyers are encouraged or required to take them together; it works best for complements with high incremental margins and high marketing cost relative to product cost, such as phone, cable, and internet packages, or software with cloud storage.
• Razor, razorblade pricing pairs a cheap durable good (razor, printer, gaming console, diagnostic instrument) with high-margin repeat consumables (blades, ink, games, reagents), often engineered to work only with the maker’s proprietary supplies; rival generic consumables can break the model. It is named for Gillette; Nespresso and Keurig machines are similar cases.
• Add-on pricing charges for optional extras at or after purchase, aiming for profitable margins once the buyer is captive, as with in-app or in-game content; pushed too far, it can damage goodwill.

Pricing to build scale

Firms racing to grow may use penetration pricing, discounting deliberately to win share at the expense of margin, effectively a marketing cost to acquire customers; examples include subsidized smartphones and Amazon discounting its Alexa-enabled devices. Held too long, it can draw regulatory scrutiny as unfair or anti-competitive and can make investors question when profits will arrive. For digital businesses, where the cost of one more subscriber is often tiny and the upside (word of mouth, network effects) can be large, growth is frequently the priority, encouraged by:

• Freemium, giving a level of usage or functionality free or ad-supported (news, an application, a game), widely used in digital services with network effects during a push to scale.
• Hidden revenue, serving users free and earning elsewhere: “free” media content paid for by advertising, marketplaces where the seller pays while the buyer does not, and financial services that look free.

Recurring-revenue models

Some models sell use rather than ownership, and the appeal of predictable, recurring revenue (helped by easy electronic invoicing and payment) has spread subscription pricing into new areas.

• Subscriptions deliver a product, service, or access to it for recurring fees. Long linked to media and utilities, they now reach application software and cloud infrastructure; fees can be fixed, promotional, tiered by feature, or usage-based.
• Leasing, licensing, and franchising resemble subscriptions but hand over property, or the right to use it, instead of a product: leasing suits real estate and physical assets, licensing covers intellectual property, and franchising is a fuller form of licensing, granting the right to sell in a defined territory with marketing and other support.

A value proposition is the set of attributes that make customers buy from one firm rather than another, given relative prices. These attributes range widely and include the product itself (capability, performance, features, style); the service and support around it (high-touch or low-touch service, access to repairs and spare parts); the buying process (convenience, easy returns); and price relative to rivals. Tesla’s proposition, for instance, reaches past transportation to electric propulsion, zero emissions, strong and quiet acceleration, software-upgradable self-driving features, a proprietary fast-charging network that is hard for entrants to match, and considerable brand loyalty built on design, founder personality, and word of mouth.

Organization and capabilities (“how”)

Judging a model means asking how the firm is built to deliver: which labor, capital, relationships, intellectual property, and other capabilities it needs, and whether it must own them or can contract for them. Dependence on outside firms for critical inputs makes those supplier relationships part of the strategy and a source of risk. The auto industry traditionally pairs assemblers with networks of parts makers supplying components (tires) or whole assemblies (engines), often custom-designed for a given vehicle; involving suppliers early improves quality and fit but raises reliance on single sources, as the COVID-19 semiconductor shortage showed when it cut output and sales. Some makers outsource assembly entirely, as with Porsche using Valmet and Magna Steyr. Tesla runs the opposite way, emphasizing vertical integration with in-house batteries and software and electronics, a capital-heavy, unusual choice for a young company but consistent with its proprietary-technology strategy.

The “how” is also called the firm’s value chain: the systems and processes inside a single firm that create value for its customers, including value-adding functions that do not physically transform or handle the product. This differs from the supply chain, which is the full sequence of steps that produce and deliver a physical product to the end customer, inside and outside the firm, regardless of who performs them. Value chain analysis, framed by Porter in 1985, links the value proposition to profitability. It groups a firm’s work into primary and support activities.

Each component is assessed by identifying the specific activities the firm performs, estimating the value added and cost of each, and locating where competitive advantage might sit; digital technology has reshaped how many of these functions are carried out. Finally, a model should show how profit is meant to arise: how prices and volumes cover, then exceed, variable and fixed costs now or at greater scale. Analysts look at margins, at break-even points, and at unit economics (revenue and cost expressed for each unit). Amazon Web Services illustrates the payoff of scale: as volumes rise and technology improves, both unit costs and unit prices fall, but because most costs are fixed, unit costs fall faster than prices, so lower prices win more usage in a virtuous cycle and profits climb, a formidable barrier since a new entrant matching the price would lack the scale to avoid losses.

Some models are common and long established, and on their own or combined they describe most firms in practice, so analysts need them cold. Each industry tends to run its own version, for example an “athletic apparel manufacturer.” Eight conventional models are summarized below.

Variations on the conventional models

Many firms are simply combinations or industry-specific twists of these models:

• Private label or contract manufacturers make goods that others market, common in offshore production; Apple and NVIDIA design and sell but contract production to specialists in Asia while managing complex supplier webs and focusing on R&D, product design, and marketing.
• Value added resellers both distribute a product and handle the harder tasks of installation, customization, service, or support, common for construction machinery, for IT hardware, and for enterprise-software systems.
• Licensing arrangements let a firm make a product under someone else’s brand for a royalty, common in toys and apparel using film characters, sports teams, or popular brands.
• Franchise models give franchisees a tightly defined, exclusive right to operate under one brand using proprietary products and methods; the franchisor collects a royalty on their sales and typically funds product development and advertising, sometimes for an added fee. Restaurants, retailers, and auto dealerships are common examples.

Business model innovation

Much of the interest in business models is in innovation: introducing or adapting a model in an existing market. It is often paired with new technology and led by entrants rather than incumbents, and over time an innovation hardens into convention. Historical cases include low-cost and ultra-low-cost airlines (leisure mass market, point-to-point, no frills, direct digital sales), software as a service (monthly fees instead of an upfront license, widening the customer base), discount and e-commerce retailing and digital marketplaces (price, selection, channel), and discount brokers (individual customers, low or free commissions, no extra research, direct digital sales). Digital technology has both created new products (streaming video, digital advertising, social media, dating apps) and enabled new models (digital marketplaces), chiefly by slashing the cost of communicating, sharing information, and transacting. The direct implications are large: location matters less, since customers can buy from firms with no local presence; outsourcing is easier; digital marketing reaches very specific groups cheaply and engages them more deeply; and network effects become more powerful and available to more firms.

A network effect is the way a network grows more valuable to each user as additional users sign on. A great many internet businesses rest on this: China’s WeChat, a combined messaging and payment platform, is useful mainly because such a large share of people already rely on it, and once its base passed a threshold it reached marginal users too, fuelling more growth. Online listings, social platforms, and ride-hailing are further cases, and the same effect powered older businesses from before the internet, including telephone service, card payment systems, and stock exchanges. Such networks draw on economies of both scale and scope, and the metric that matters is the count of users and the pace at which it grows. Because they are hard to dislodge once established, network effects form strong barriers to entry and a real competitive advantage.

One-sided and multi-sided networks

In a one-sided network, only one kind of user creates value for the others, as with telephone service or peer-to-peer payment apps such as Venmo, owned by PayPal; connections among users are what matter, but there is a single user type, and beyond a certain size the cost of adding a user is below the revenue that user brings. A two-sided or multi-sided network has two or more user types, such as buyers and sellers in a marketplace: card networks like Visa and China UnionPay join merchants and cardholders, and Airbnb joins hosts and guests. These can grow exponentially, because more users on one side attract more on another, which attracts still more. Food delivery is a three-sided case linking restaurants, drivers, and customers, where growth in any group enlarges the network for the others.

Crowdsourcing

Network models often rely on crowdsourcing, where users contribute directly to the product or service: social platforms like TikTok/Douyin and Reddit; open-source software projects; reference wikis like Wikipedia; and customer reviews on sites like Yelp, Amazon, and TripAdvisor. These firms host user communities that collaborate voluntarily, with only light moderation and oversight from the operator.

How models evolve: hotels and travel

The lodging industry shows models evolving and multiplying, each with different financial characteristics. Change was slow until the 20th century, when a larger, more mobile, and increasingly corporate customer base pushed operators to add scale, footprint, consistency, and efficiency, even as unique single-property hotels remained viable.

Digital technology reshaped the sector too. Online travel agencies (OTAs) automated research, price comparison, booking, and logistics and delivered them as web services, transforming the model and shifting power away from hotels; effectively, network effects let OTAs disintermediate traditional agencies and forced survivors to specialize. Both traditional and online travel agencies are two-sided networks, serving travelers on one side and hotels on the other.

Hotels have fought back by investing in their own sites, customer data, and direct booking, launching soft brands (hotels under their own name but with more local operator autonomy), and creating hybrids such as condo-hotels (owner-occupied condos that the hotel rents out otherwise).