Vertical Integration in Plant-Based and Cell-Based Meat
It’s clear that alternative proteins are at the very beginning of an upward trajectory. What’s not as clear is where exactly the industry will end up and who will capture most of the new value created. One common question facing entrepreneurs is whether to try to create new products from scratch, or to act as a supplier for other alternative protein companies. For example, is it better to create a better plant-based chicken, or to develop generalizable texturizing technology that can be incorporated into many other companies’ formulations?
One way to look at this question is through the theories of Clay Christensen. Christensen is primarily known for his theory of new market disruption as detailed in the “Silicon Valley Bible,” The Innovator’s Dilemma. However, here I want to focus on his lesser-known theory of disruption, called “low-end disruption,” articulated in The Innovator’s Solution. I’ll first explain the theory, then look at what it might tell us about the plant-based and cell-based meat industries. In particular, I think Christensen’s model clearly predicts that vertically integrated companies  will have the advantage in this early stage of the alternative protein industry.
Many industries throughout history have had the following pattern: at the beginning of the industry, a few companies dominate with highly integrated and proprietary product architectures. Each component is specially designed to work together with the other components. These companies go through a period of prosperity, but are eventually disrupted by a collection of companies, each of which specializes in an individual part of the value chain. Standardization and protocols are put in place to make it easier for firms at different stages of the supply chain to work together.
Christensen often uses the microprocessor-based computer industry as an example. At the very early stages of this industry, the dominant companies were highly integrating, making both the OS and application software, and also designing and selling the product. These companies remained dominant for a few years, and competition pushed product quality ever higher. However, eventually horizontal suppliers began making superior products focused on individual parts of the value chain. Microsoft created the Windows operating system, Dell offered superior product design and owned customer relationships. This trend was also self feeding–as modularization lead to industry-wide standardization, specialization in new parts of the value chain became relatively easier.
Why do we seem to see this pattern repeated in different industries? The theory of low-end disruption explains exactly this, illustrated in the chart below.
To understand what’s going on here, consider the costs and benefits of a single company with a vertically integrated infrastructure vs. a collection of companies working together across a value chain. The biggest advantage that vertical integration provides is that coordination is simplified between different components. E.g. The operating system can be custom built to work well with the application software. This allows the end product to be higher quality.
In more modular value chains, transaction costs between firms can be high. To combat this, industries will often develop standardized interfaces between components. For a modern example, standardization of the Java programming languages allowed third-party smartphone app developers to easily develop new apps for Android. Standardization allows components to work together without explicit coordination, but imposes restrictions on engineers at each level, lowering the quality of the end product. However, once products are assemblies of many different parts, mixing and matching allows products to be specialized towards particular consumer segments, or even individual consumers . Competition at each stage of the value chain also drives down prices. Therefore, modular industries generally have lower quality, cheaper products, but better meet consumer demands .
As a result, different product architectures are advantaged at different stages of an industry’s development. At the very beginning of an industry, overall technical capability is very low, meaning that products are struggling to meet consumer demands on quality. Consumers are willing to pay a premium for a higher quality product, which advantages vertically integrated firms . Gradually, the technical capabilities of the entire industry increases, and eventually there is a performance surplus. At this point, consumers stop being willing to pay a price premium for higher quality. In Christensen’s words, this transition occurs when products are “good enough” for consumers. After this point, modular architectures have the advantage since their lower quality is no longer a detriment.
This explains the pattern we see in the microprocessor-based computer industry. In the later 1970s, computers had a quality gap, which favored integrated companies like Silicon Graphics. Later, computers developed a performance surplus, which favored specialized companies like Microsoft.
Understanding Plant-Based Meat
The theory of low-end disruption can help us understand the plant-based meat industry’s current state, and where it it might go in the future.
Let’s start with the value chain for plant-based meat. There are five main parts:
Intellectual property (in the form of patents and trade secrets).
Material inputs like soybeans, coconut oil, beet juice, etc.
Activities related to creating the base material for the product. For example, soybeans may be dried or extruded to make the material that will become the majority of the volume of the meat. This requires machinery and labor.
Processing such as mixing, flavoring, texturizing, cooking, functionalizing, and packaging.
Distribution, which encompasses everything from sales and marketing to shipping logistics.
The plant-based meat industry has a straightforward analysis under the low-end disruption framework since there is a clear metric of “good enough”—whether products stand up to animal meat in blind taste tests . Current products aren’t good enough yet, although companies are making a lot of exciting progress. Given that there is a performance gap in current products, Christensen’s theory would predict that integrated companies are currently advantaged. Once products do stand up in blind tests, then the theory would predict a fragmentation in the industry with individual companies specializing in e.g. protein sourcing, formulations, fat incorporation, structuring, etc. This fragmentation could then facilitate a larger selection of plant-based meat products, accounting for all the cuts and preparations that currently exist for animal meat.
How well do the empirical facts match the theory’s predictions? Beyond Meat and Impossible Foods are the two plant-based meat companies with the most attention right now, but older brands like Tofurky, Quorn, Field Roast, Morningstar, and Boca also have substantial market share. The whole industry is growing quickly, but I’ll focus on Beyond Meat since their recent IPO filings give us a window into their business operations.
Beyond Meat’s major success started with the Beyond Burger, a pea-protein-based hamburger. More recently, they created a sausage and a ground beef product, all using the same “woven” pea-protein base. They developed the technology to make this woven protein, as well as their own product formulations. They also have their own production facility where they create the protein, which they then ship to co-manufacturers. These co-manufacturers do the rest of processing like mixing, flavoring, and packaging. Beyond Meat uses a third-party logistics firm to manage distribution, although they also maintain their brand and sales themselves. In sum, Beyond Meat is integrated across technology, manufacturing, and brand.
This makes sense given the theory of low-end disruption. Because plant-based meats do not yet stand up on blind taste tests, Beyond Meat benefits from the fact that all of their activities can be optimized to work well together. They develop all of the technology needed to make their products, and also manufacture the new woven protein themselves. I believe we see a similar story in the other market leaders for plant-based meat. Companies generally develop new products using proprietary technologies and formulations, then develop a brand alongside that technology. Companies that have more complicated manufacturing technologies like Impossible Foods also tend to have their own facilities.
Interestingly, the parts of the value chain that Beyond Meat operates in are also the parts where they’re doing something new. For example, the woven pea protein that is the base material for their most successful products isn’t something they can source from an existing supplier, so they have a manufacturing facility where they make it themselves. This could be seen as an example of different components in the value chain being custom built to work with one another–their facility is built to work well with their technology. Finally, their brand is built around appealing to meat eaters, which also hadn’t really been done before. The parts of the value chain that they outsource - raw ingredients, basic processing like packaging and mixing, and logistics - are all areas with similar requirements to other food manufacturers.
Understanding Cell-Based Meat
The cell-based meat industry is substantially younger than the plant-based meat industry (in fact, I’d hesitate to even call it an industry at this point). Again, the theory of low-end disruption predicts that integrated companies will be advantaged in the early stages. The performance gap here is not that products don’t stand up to meat on blind taste tests, but rather that costs are very high.
The value chain for this industry is theoretical at this point, but let’s assume there will be five areas, similar to plant-based meat.
Intellectual property—this is what most companies are currently focused on.
Material inputs, which are the chemicals and nutrients that go into the media, any necessary growth factors, and eventually scaffolding materials.
Activities related to growing the cells that will make up the meat. This includes facilities with bioreactors, as will as labor to operate the bioreactors.
Processing, which includes turning the cells into an edible meat material, then adding flavoring, coloring, and texturing agents, and finally packaging.
Distribution, which again encompasses sales, marketing and logistics.
The most important type of integration that we might see is integration between different technologies at the IP level. Instead of companies specializing in one of the sub-areas of intellectual property, we might predict that companies will have proprietary technology at every level (cell lines, media, bioreactors, bioprocess, and potentially scaffolding), each working together in custom ways with technology at other levels. This makes sense given the immense technical challenge that cell-based meat presents. Bringing down costs will be easier without the requirement of e.g. making cell lines compatible with processes and products from other companies.
On this theory, it’s reasonable to think that a cell-based meat company in the future would look similar to how Beyond Meat currently looks. They would have proprietary technology in all of the main areas and they would own their own manufacturing facility where they would create the raw material for their meat products. After a small amount of processing to turn the mixture of cells into an edible material, they would pass this material off to co-manufacturers who would then take care of simple processing like flavoring and packaging. The company would also have a strong brand that would increase consumer trust in the new technology.
It’s Hard to Start a Horizontal Company Right Now
The Good Food Institute used to publish a list of “whitespace” ideas for new alternative meat companies. Many of these company ideas involved focusing on one part of the value chain, then supplying other vertically integrated companies. However, we’ve seen relatively few of these companies being formed—most cell-based meat startups are developing complete food products. There are lots of reasons this could be the case:
Specialized companies need relatively more technical expertise over a narrower area, as opposed to general biotech expertise.
Companies take a long time to get off the ground, so we could just not be seeing the specialized companies yet.
Cell-based meat companies are all very young, and could specialize in the future.
However, another explanation in light of low end disruption is that a horizontal business model is fundamentally difficult at this point in the industry’s development. Consider a company focusing solely on media for cell-based meat. This company won’t know exactly what the technology will eventually look like, making it difficult to position themself as a supplier. For example, it’s not clear whether they should focus on media for induced pluripotent stem cells or immortalized myocytes. They also don’t know exactly what technologies vertical companies like Memphis Meats will develop in-house. If the industry ends up being concentrated to a small number of companies, each with unique media needs, it’s possible that a specialized company wouldn’t be able to sell their generally applicable product.
Essentially, a horizontal company is doubling down on risk in an already risky business. The “bets” that cell-based meat companies are taking right now are 1) the technology is fundamentally possible in the short-to-medium-term, 2) consumers will eat cell-based meat, and 3) governments won’t block the development and sales of cell-based meat. A horizontal company has these risks plus the added risk that their products won’t be compatible with the dominant processes of other companies.
Some of the efforts for specialization are by pre-existing companies that want to be suppliers in the future. This could make more sense for large biotechs, given that they already have substantial capacity. This strategy would work especially well if aspects of the technology end up not being heavily customized. For example, if basal media ends up being the same in cultured meat as it is in biopharmaceuticals, then cell-based meat companies could source basal media from an existing company, just like Beyond Meat sources their dried pea protein blend from existing suppliers. Here, established companies have an advantage over startups given their existing capacity, and ability to de-risk their exploratory activities with an already-stable business.
The question of who will be advantaged at a particular stage of an industry’s development is only partially related to what an entrepreneur should focus on right now. It takes a long time to start a company and develop a technology platform. Roughly speaking, if one thinks that plant-based meat will become “good enough” in 5 years, but it takes 2 years to bring a new company to market readiness, then the correct time to start a specialized company in plant-based meat is in 3 years.
Of course, what we’ve been discussing is a theoretical model, not empirical truth. Models can help retroactively explain facts, or generate hypothesis about unknowns given a structural understanding of the phenomena. However, in the real world, there’s always random noise, and competition in business tends to select for outliers. Therefore, I think it’s a mistake to place too much weight on any particular model. There could be important differences in the alternative protein industry that will cause it to diverge from the model in ways that are hard to predict, or anomalies could take the industry in unexpected directions . While I think it’s clear what the model says about alternative proteins, I think we should use this conclusion as one consideration among many.
Who do you think will be the most successful companies in the early days of the alternative protein industry? Let me know in the comments!
Thanks to Alene Anello and Abhi Kumar for comments on drafts of this piece.
 Vertical integration occurs when a single firm operates in multiple parts of the value chain for a product. For example, a firm that develops technology and commercializes consumer products based on this technology is vertically integrated. If the firm instead developed technology and then licensed that technology to many other firms which in turn made various products, then the whole value chain would be modularized (think of the various buyers as different “modules” for the technology).
 Think about the difference between Marriott and AirBnb. Marriott is integrated across the physical location and service aspects of their value chain: they own hotels and staff the hotels with company employees. AirBnb contracts out these parts of the business to anyone who owns a home. Marriott is able to offer a higher quality experience at a higher price, since they control the design and location of the room and service processes. On average, quality is lower at AirBnb (e.g. there is no room service) and rooms are cheaper, but AirBnb offers customers the flexibility to stay wherever they’d like in a city, and at many different price brackets.
 I was confused by this when first learning about Christensen’s theory–how can lower quality products better meet consumer demand? Not exactly. Consider a luxury resort of the highest quality–if the resort is located in the suburbs where tourists don’t want to go, it’s high quality, but doesn’t meet consumer demand. Additionally, if the quality overshoots what consumers actually care about, then consumers may prefer to have a lower quality option at a lower price.
 Another factor in highly technical industries is that in the beginning, the technological needs of the product aren’t yet fully known, making modular solutions difficult. This seems to be happening in cell-based meat. We don’t yet know whether the preferred cell type for cell-based meat will be embryonic stem cells, induced pluripotency stem cells, muscle satellite cells, or immortalized somatic cells. The cultured media requirements of each cell type are different, so a company just focused on media might not know where to focus their efforts.
 This is somewhat of a simplification for the sake of the model. Taste and culinary preference are partially culturally determined, meaning that consumer preferences could change to be more accommodating of taste difference between plant and animal meat. Additionally, plant-based could become “good enough” if it beats animal meat along some metrics (e.g. nutritional composition) and loses along others.
 For example, Apple, one of the most successful companies of all time, notably avoided low-end disruption. According to Ben Thompson, this is because UX is different than other things customers care about in that it can’t have a quality surplus.