Redinel Korfuzi, analyst on the Pan European Equity Strategy managed by Jamie Ross, explores the power of platform businesses and how the trajectory of their corporate models can often be misunderstood.

   Key takeaways:

• Platform businesses facilitate the exchange of value between producers and consumers for something that is not produced by the platform itself. Rather than creating value by making and selling things, platforms provide a mechanism for exchange which is much less capital and labour intensive.
• These types of firms often undergo network effects which make them more valuable the more people use the product.
• Most investors continue to apply linear business prospects to platform businesses which does not consider the flywheel effect that is often present. The market can sometimes struggle to digest this concept, which can lead to inefficient pricing.

 

In our previous article “Knowing where to fish” we outlined how internet-enabled business models offer immense opportunity to the patient investor, if they pick the right companies. Within internet-enabled businesses, we focus particularly on platform business models.

Since the industrial revolution, when the invention of machine tools and the development of steam power marked the transition from hand production to mechanised manufacturing processes, the industrial landscape has been characterised by ‘linear’ business models. Linear businesses take inputs such as physical production assets and labour to make goods and services that are sold to customers, according to a linear, unidirectional value chain. Think car manufacturers, steel makers and consumer goods companies. In many cases technological advancement has not changed the linear aspect of these companies’ value chains.

In several industries, the internet has dramatically lowered distribution costs, permanently impairing capital in those industries that previously enjoyed economies of scale in physical distribution. Examples include travel agents, newspapers, bookstores, video rental services and shopping malls. Businesses with high fixed costs of incremental production faced competition from companies enjoying negligible marginal costs of internet distribution. The transition that many software companies have made to recurring revenue models has been driven by the desire to lower the marginal cost of distribution and increase incremental returns on capital. But these software companies still operate a linear supply chain. Linear businesses create products and services using their internal resources and means of production.

 We tend to find platform businesses far more interesting from an investment perspective. A platform business facilitates interactions and the exchange of value between producers and consumers of something that is not produced by the platform itself. Rather than creating value by making and selling things, platforms provide a mechanism for exchange. As such, platforms are much less capital and labour intensive than linear businesses. Platforms generate revenues by capturing a share of the transaction value their network enables. Producers that use a platform accept an additional cost of unit production in return for the larger volume opportunity that the platform affords.

While the internet has lowered the cost of distribution for many businesses, platforms tend to enable even lower marginal costs of distribution, driving superior unit economics relative to linear businesses. Building a successful platform is not straightforward and necessitates solving the ‘chicken-and-egg’ problem of needing to have enough producers to attract consumers whilst simultaneously needing enough consumers to attract producers. Without adequate scale and liquidity, the value to users will be lower than the cost of participation. It is this dynamic that makes platform businesses more difficult to build from a standing start when compared to their linear counterparts. However, once you are past the tipping point at which user value is greater than the cost to participate, powerful network effects take hold which can rapidly increase the per-user value of the network. Therein lies the real beauty of this type of business model.

Network effects: value grows as the square of the customers

This idea of network effects comes from computer networking. Bob Metcalfe, who created Ethernet, famously coined Metcalfe’s Law, which is: the value of a network is proportional to the square of the number of nodes in the network. If a network with 10 nodes has a value of 100, a network with 100 nodes would have a value of 10,000. The larger network is not just 10 times as valuable as the smaller network; it is 100 times more valuable. Looked at another way, we are talking about the nature of compounding.

A network effect exists when each additional user adds value to the existing user base. The new users are themselves creating some additional value for the existing users. The classic example to illustrate this is language. Imagine 100 people living in the community that speak 10 different languages, and each person speaks just one of those 10 languages. In that case one would have to translate all the time, which is work intensive and time consuming. If one extra person learns English, taking the total number of English speakers to 11, the next person that decides to learn a new language is probably going to choose English. At some point, if there are 20 or 30 English speakers, English has won the game. English will dominate the entire language marketplace, leaving other languages at a competitive disadvantage.

All of us can relate to this example. Nowadays, English is indeed the mother tongue of all technical education. If a scientist, for example, does not read or speak English he is likely to be at a severe disadvantage to his peers.

Languages are probably the oldest example of a network effect.

This idea of network effects comes from computer networking. Bob Metcalfe, who created Ethernet, famously coined Metcalfe’s Law, which is: the value of a network is proportional to the square of the number of nodes in the network. If a network with 10 nodes has a value of 100, a network with 100 nodes would have a value of 10,000. The larger network is not just 10 times as valuable as the smaller network; it is 100 times more valuable. Looked at another way, we are talking about the nature of compounding.

A network effect exists when each additional user adds value to the existing user base. The new users are themselves creating some additional value for the existing users. The classic example to illustrate this is language. Imagine 100 people living in the community that speak 10 different languages, and each person speaks just one of those 10 languages. In that case one would have to translate all the time, which is work intensive and time consuming. If one extra person learns English, taking the total number of English speakers to 11, the next person that decides to learn a new language is probably going to choose English. At some point, if there are 20 or 30 English speakers, English has won the game. English will dominate the entire language marketplace, leaving other languages at a competitive disadvantage.

All of us can relate to this example. Nowadays, English is indeed the mother tongue of all technical education. If a scientist, for example, does not read or speak English he is likely to be at a severe disadvantage to his peers.

Languages are probably the oldest example of a network effect.

 

In addition to network effects, successful platforms also benefit from the impacts of producer switching costs. On a platform, consumers often provide non-monetary benefits to producers in the form of reviews, ratings and other feedback such as ‘likes’ and ‘shares’. This is an important mechanism encouraging good behaviour and preserving required levels of customer service. As producers build up this goodwill asset, the costs of leaving the network and establishing a reputation from scratch increase.

The platform business model is also less susceptible to antitrust than linear business models. The objective of antitrust is to promote fair competition for the benefit of consumers. Traditional monopolists control the means of production. They earn supernormal profits by restricting output, charging anticompetitive prices and minimising consumer surplus. Platform businesses do not own the means of production, but simply the means of connection. Platforms do not control the prices and outputs of producers. But by facilitating more touchpoints between consumers and producers they can increase competition, to the ultimate benefit of the consumer. The goal to increase platform scale benefits both its owners and users. Platforms are therefore characterised by an alignment of interest between shareholders and customers that is not the case in traditional linear monopolists. A regulatory intention to curb a platform’s dominance could have the undesirable effect of harming consumers’ interests.

 Platforms make fragmented supply available to consumers in one place. Consider consumers wishing to find the best price for a house or a car. There is an opportunity cost associated with the time spent manually finding the owners of that inventory and comparing availability and pricing. Conducting that price comparison exercise for your property on a website like Immoscout (owned by the German listed business Scout 24, in which we have a shareholding) is a much more efficient process.

In addition, price transparency leads to a more competitive outcome for the consumer. The platform increases choice and saves the consumer time and money. And the more scale the platform has, the greater the choice and the more competitive the price. In this way platforms grow by delivering more value to their users rather than by controlling the supply chain. However, while these considerations may be logical, until the regulatory regime evolves to be suitable for 21st century business models regulatory bodies are likely to use arguably unfit-for-purpose antitrust frameworks to make life difficult for some of the dominant platforms (such as Airbnb and Uber).

 One of the main challenges for regulators as well as investors in analysing such businesses is identifying the total addressable market (TAM), especially at the beginning of a platform’s life. For example, the valuation exercises for Uber’s equity value at its origins started with the global taxi industry as the addressable market. With the benefit of hindsight, we can say that was plainly wrong because the potential market was much greater than the taxi market due to the superior customer experience when compared with traditional taxis (cheaper, shorter pick up times, more convenient payment methods, tracking technology, greater safety and trust). As such, the relevant addressable market was not just the taxi industry, but also public transport, walking, rental cars and driving owned vehicles.

The outsized return on investment for Uber’s early investors was made available in part because of the equity market’s failure to look beyond the incumbent industry in determining the prospects of those disrupting it. Platform business can thus be underappreciated when there exists the potential to leverage the strength of the network to invest into adjacencies.

In his 2011 annual letter, Amazon’s Jeff Bezos explained at length how a platform nourishes creativity, fosters innovations and as a result, creates new TAMs.

Jeff Bezos, 2011 annual letter to Amazon shareholders

Even today, the market has a tough time to digest this concept and remains quite inefficient in pricing in this flywheel effect. Most investors continue to apply linear business prospects to a set of outcomes which could include exponential value growth.

Linear businesses’ customer acquisition efforts result in singular relationships per customer added. Recurring revenue business models may lower the cost of customer acquisition and retention, but they still result in a singular profit stream associated with each customer. When platforms acquire users, whether they are customers or producers, those users form relationships and transact with the existing users in the platform. As the number of users grows arithmetically, the number of potential relationships and interactions grows exponentially. And due to the very low marginal costs of distribution and production, profit margins can expand materially. The application of linear growth expectations to platform businesses therefore implies a failure of the platform to facilitate, nurture and monetise these potential interactions.  Once we understand the dramatically lower cost of growth for these businesses, we should question the relevance of previously accepted valuation methods based on outdated accounting metrics.

Mastercard is the perfect example here. Over the last decade Mastercard has compounded its earnings per share at 20% per annum (pa), and its share price has grown 30% pa. Ten years ago, Mastercard was valued at a mid-teen multiple of its per share GAAP* earnings. An investor could have paid the equivalent of 80 times the company’s after-tax accounting profits and still achieved an opportunity cost on investment of 10% per year for the next decade. Clearly, these metrics didn’t help the investors at the time to grasp the full value creation potential of Mastercard.

The availability of this opportunity to public equity investors implies a dramatic underappreciation of the exponential growth properties of successful platform companies. We will remain vigilant and seek to identify these mispriced opportunities.