"Are you a network or a platform? Say “both” and you’ll lose a Series A funding from Benchmark." - Garry Tan, Initialized Capital (former founder of Posterous S08 YC)
In the world of tech startups, there are four types of companies: tools, network, platform, and compounder.
The tools
Most tech companies fall into this category. From ETL, Business Intelligence, Open Source Software, to Content Delivery Network and Development Toolkits, these companies typically dominate in a particular use case that is part of an important production goal. For example, a business report requires ETL and BI tools. A video streaming website requires CDN, and a bunch of open sourced packages.
The “middleware”
These tools are used in connection with other tools and platforms. In software, many of these tools are often referred to as “middle software” that glue services together. Rarely do they perform a critical function on their own. For example, Tableau isn’t very useful without a data source.
Important use cases
Tools deliver important use cases. Our internet wouldn’t function well without CDNs and Server Observability tools. By use case, it is meant to deliver a critical step of a production over and over again such as a RPA tool.
Naming conventions
Tools have cool names such as ETL, RPA, BI, COSS, CDN, CRM, IDE… These abbreviations suggest a use case such as Integrated Development Environments or Robotic Process Automation. In each use case however, there could be many tools or vendors in the exact same space with slightly different product offerings. Think of Business Intelligence space, there are hundreds of dashboard builders and data visualization tools on the market.
The network companies
Network companies typically don't hold assets. They build a network of demand and supply as a business model.
Some rely on advertising revenue such as Twitter and Google (search). Some rely on transactions such as AirBnB and Instacart. Some may rely on subscriptions such as Netflix, Instacart.
As network businesses, they facilitate existing demand and supply and serve as the intermediary to both markets. Many would call them marketplaces.
One of their core missions is to grow both demand and supply. The successful ones would become synonymous to a core use case that grows more significant such as riding sharing, website search, and micro-blogging.
Supply and Demand
The main difference between a network business and a tool is that a network doesn’t necessarily need a strong product to compete and survive in the market. Instead, it requires wide adoption and increasing volume in both demand and supply. Sometimes it’s led by a strong product, but often it’s led by “first-mover advantage”, “proprietary technology”, “favorable fundraising environment”, “market/consumer shift”, “geographical/regulatory advantage”....
A great product doesn’t necessarily increase both supply and demand. All of the above conditions will do however if they’re met all at once. This is why it is so hard to build a network business. Imagine Netflix was built in China in the early 2000s (many Chinese companies actually build similar products), these companies would never succeed with all the knockoff DVDs, pirated content, terrible paying habits, and lack of IP protection… Netflix however was born in the aftermath of the Dot.com bubble, founded in California (near Hollywood), and protected by the US copyright law.
Explosive Growth (not necessarily in money terms)
A network company has to experience explosive growth unlike a tool or platform business. Without viral adoption, the network would not stick together. Again, favorable market/regulatory/consumer terms play a huge role. Imagine Stripe was founded by 2 mid-aged banking executives in Florida, it would never become the huge payment network we know it is today. For a network to win, everything has to go right.
Volume = monetization
The way a network monetizes is drastically different than a tool or platform company. A network typically cannot profit from their “customers” or network early on since they need to reach a critical mass of users/audiences to become the default use case provider. However, they’d still have a north star KPI metric such as number of tweets/searches/interactions/bookings/hours streamed/transactions…Even though they might lose money growing these metrics first, they can leverage it to dominate a market and charge their customer later on. Again, this is very different from a tool/platform which charges a customer right away.
The platform companies
Platform businesses in other words are infrastructure companies that support critical production of an application. They are typically very asset heavy.
Since they have to build supply first, demand is often unclear in the early stages. As platform businesses, they build new solutions of supply to fulfill existing or future demand. Their core mission is to support production of a growing use case such as customer engagement, VR gaming, business website, or e-commerce stores.
Unclear demand but strong supply required
Unlike a tool or a network that either fulfills or increases existing demand, the demand is typically very vague for platform business. Not every company foresaw the domination of a mobile OS, E-Commerce store, cloud computing platform, or VR games (not yet?). These things may seem clear now, they were huge bets made with a lot of risks. Many of the tools building companies thought they were only one step away from becoming a platform but they could fail due to not having enough supply of the demand. For example, Heroku never became the platform it promised to be simply because it didn’t stock up supply fast enough.
Platform or tool?
Too many platforms fail to grow and hence become just another tool. There can be several reasons for this phenomenon.
- The “platform” is not technically scalable to support growing use cases
- The product/service produced by the “platform” isn’t valuable enough due to the fact that the “platform” only support one or a few creators (e.g. alteryx, ETL tools, data science tools)
- The demand isn’t strong or profitable enough for the “platform” to grow in use cases (e.g. microblogging, quant trading,
- The “platform” fails to deliver a great variety of valuable production goals
For a tool to become a platform, the above reasons have to be addressed and overcome. Typically, it’s very hard to morph a tool into a platform organically.
Openness & Monetization
Many platforms fail due to their failure to realize developers’ imagination. Many software companies consider themselves as platforms in which they can support critical applications such as business workflow automation, analytic dashboards, internal tools etc. Their “platform” however often fails to support multiple developers in one place for delivery of high production value. This is why the Dev Platform has been so powerful since the invention of Git. Codes can be compiled together by a huge team of developers and shipped into production for millions of users. A self-service BI, ETL, Workflow tool however does not encapituate enough production value due to the productivity of a single user.
Open sourced SDKs are therefore a powerful way to capture production value. Collaboration-first software is another way to go. However, most collaborative software products (figma, slack, jira, trello) fall in the “tool” category. There has not been a collaborative software that delivers high-production-value projects yet. They typically address use cases such as messaging, project management, design prototyping, documentation etc. These are valuable functions of a team but not high-value production goals such as setting up an E-commerce store, SaaS application, or VR game.
Monetization for platforms is complex. Unlike a network that’s transaction based or a tool that’s feature based, a platform can have different business models for its product lineup. Certain products can cost a lot for power users, whereas others may be free to offer.
Developer Incentives & launch of a platform
A tool is easier to launch since it may address a use case with a better solution. Users can identify value rather quickly after a trial. A network like we discussed is a lot harder to build since many objective conditions need to be met. Additionally, early adopters/advocates need to be discovered very early on for viral growth. A platform on the other hand is similar to a tool in regards to delivering a better solution. Though the main difference is that the platform can find not only “users” but also “developers” for supporting the production goal. This means that the platform charges their customers not based on their costs of developing features & tools but the value they can provide to their users. A network in a lot of cases will need to become a platform as well in order to monetize their network more efficiently. When volume/transactions plateau, a good network business typically identifies one party of their network as their customer such as Facebook selling to advertisers, Airbnb selling to house hosts, Stripe selling to SaaS businesses etc. A platform on the other hand will always rely on its developer community.