Quick thread on how to think about building a business from things like dev tools or OSS projects. Particularly OSS projects.

I don't have a fancy @benthompson name for this, I simply call it the

Project -> Product -> Platform Continuum

The basic idea is that not all projects can or should become products and even less of those have the ability to become platforms.

And the projects that can and do become products are likely not in the initial project form. And products -> platforms changes similarly again.

This feels like pretty basic thinking though I'm seeing many open source projects attempt funding on the premise that they can create an enterprise around the community or project.

Or folks who have an interesting project inside an enterprise and want to spin it out as-is.

This almost never works. So let's explore what you can do.

Fundamentally, *most* projects won't make interesting enterprises unless you want to create a niche consultancy around them. Think language, framework, etc etc

When projects do make interesting products is typically when the project is scaled (think IaaS), adding collaboration (enterprise/community), rooted around a specific problem that has nothing to do with the project (project solves *another* problem).

The first (IaaS) is the most basic and IMV, the hardest to make work. It's becoming harder and harder to have a business around "hosted open source project" for a variety of reasons. It's simply become too easy for most organizations to do this on their own with their own.

The 2nd is a great way is still a great way to go. Collab and community bring major value

In fact, early GitHub is a prime example of this.

(more on GitHub another time, though product GitHub is interesting, platform GitHub is most important software company in the world)

Note: adding collaboration and leveraging community is different than monetizing the community. This is a huge difference in thinking.

Communities can create value, though the community itself is unlikely to monetize. Trying to make the community pay you is unlikely to work

The third is, imo, the perhaps the most intriguing at the moment.

Using open source/project to solve another problem that is real and needed.

A great example is Fastly. Fastly is *still* mostly a product offering though has platform tendencies. Compute@edge is full platform

Now, let's move those products to platforms.

What is a platform? There's no real one definition which will always infuriate people. In fact, if @chamath ever read this I'm sure he'd drop any number of f-bombs on me for this thread.

It's cool, I'm still a Warriors fan

I suggest people read @benthompson here:

https://stratechery.com/2018/the-bill-gates-line/

And dive into his concepts of Aggregators v Platforms here: https://stratechery.com/concept/aggregation-theory/platforms-vs-aggregators/

to start.

But I am also going to give you my suuuuuuper simple definition that I used to guide how I think about it.

1. Other people *need* to use you (aka you bring something to the party)
2. Other people *want* to use you (aka it's better for them to build on you instead of going alone)

3. People find value with you they otherwise couldn't
4. you cannot be replaced by single tool, competitor, ecosystem.
5. full programmatic interfaces

Most of these should be satisfied IMO, though not all need to be.

Think of this list as base requirements (IMV of course)

So let's talk about taking a product to platform.

First, strategy is important. I usually suggest people build out current product adjacent value *first* and then extend from there.

e.g. Datadog getting into project planning would not make sense to users.

Second, you have to find code level ways for folks to interact with you.

Think: API, functions etc.

Yes, products usually have this too though it's a requirement for platforms. Likely multiple ways for different purposes.

Third, look for ways to push your value *lower* into the stack, not higher initially.

If you are a platform, products can be built *on top* of you, and you want to become a more 'substrate' layer below them.

Think: cloud infra vs mobile app

Another way to think of this here is interfaces.

CLI, mobile, API, web, etc . If you have a platform, all these are essentially *interfaces* into or of the platform.

Fundamentally different from when you are a product and mobile app *is* the entire thing.

And interfaces can be entire companies. If companies are popping up built mostly on the value you are bringing you are on the right track.

Best in the biz at this has been Salesforce.

Interestingly Salesforce has some of the worst products but best platform for their ecosystem

Last, you have to be able to clearly identify the moat, and as a platform you have to have at least one and should have several.

I call these the unfair natural advantages.

If you can't identify these, work to discover and understand them.

And then build on them.

Twitter is the wrong medium for this topic though hope it helps someone as they think through their plans.

Also know, you don't have to have it all figured out right away though you *must* know how the project turns into a product w/ real user value. Why will some pay you $

tl;dr

Projects are the majority. Most likely don't make great products, particularly VC funded biz

Products find user value via the proj though not exclusively of the proj

Platforms move from initial proj to broader eco of value. Initial proj might even be least important bit