>

>

What Should Actually Be Paid For in Open Source?

>

>

What Should Actually Be Paid For in Open Source?

>

>

What Should Actually Be Paid For in Open Source?

Notes on open source & monetisation

Notes on
Growth, design for open source start-ups

What Should Actually Be Paid For in Open Source?

Designing the boundary between a commons and a business, and why I think good monetisation begins when the company takes on a new kind of responsibility.

by David Soh

14 Sep 2026

12 min read

Growth strategy

Business Models

Monetisation

Open Source

This is the third essay in a short series on growth and open source. The previous, What I Had to Unlearn About Growth in Open Source, was about why free usage, community participation and commercial demand cannot be treated as the same thing. This one asks the question that followed naturally: if that is true, what should an open-source company actually charge for?

This is the third essay in a short series on growth and open source. The previous, What I Had to Unlearn About Growth in Open Source, was about why free usage, community participation and commercial demand cannot be treated as the same thing.

This one asks the question that followed naturally: if that is true, what should an open-source company actually charge for?

When I was scouring the web on monetisation, I found a similar rhetoric across many content creators and it seemed like

the easiest way to think about monetising open source is to ask what can be put behind a paywall. I think that is often the wrong starting point.

the easiest way to think about monetising open source is to ask what can be put behind a paywall. I think that is often the wrong starting point.

By the time I had worked through the difference between adoption and commercial demand, I was left with a more uncomfortable question. If the open product is genuinely useful, if free users are creating value for the ecosystem, and if community participation is part of what makes the project stronger, then monetisation cannot simply mean finding more ways to restrict access.

The Premise

A move for monetisation should
always be able to answer this fundamental question:

A move for monetisation should always be able to answer this fundamental question:

"What new value should the commercial organisation be responsible for creating?"

In this essay, I show how I began thinking about monetisation less as a pricing exercise and more as a boundary-design problem. What belongs to the shared ecosystem? What becomes materially different when an individual tool becomes organisational infrastructure? And where can a paid relationship make both the customer and the wider project stronger?

I wanted to test that instinct against the research rather than simply turn my own experience into a universal rule.

01 – STARTING WITH THE VALUE

Open source can create enormous economic value without capturing it as revenue.

Open source can create enormous economic value without capturing it as revenue.

This sounds obvious, but it changes the framing of the business problem.

Harvard Business School research on the economic value of open-source software estimates that firms would have to spend roughly $8.8 trillion to recreate the software they currently use, and that building without open source would cost organisations about 3.5 times more. 1 The precise number matters less to me than the asymmetry it exposes: value creation in open source can be enormous while direct payment for the underlying software remains zero.

That means revenue is not a complete measure of whether value exists. It is a measure of how much of a particular kind of value the company has chosen, and managed to capture.

This is also why I am wary of treating a large free user base as unrealised revenue. Some of that value is supposed to remain distributed. Open access can create experimentation, familiarity, contribution, integrations, advocacy and trust. Harvard researchers Sonali Shah and Frank Nagle describe user communities as a distinct strategic structure for exchanging ideas and knowledge, rather than simply another marketing audience. 2


More recent Linux Foundation research makes the commercial relationship even more interesting: its 2025 analysis of 800 venture-backed startups found a close link between community health and company valuation, suggesting that community value and business value can reinforce one another rather than existing in opposition. 3

This sounds obvious, but it changes the framing of the business problem.

Harvard Business School research on the economic value of open-source software estimates that firms would have to spend roughly $8.8 trillion to recreate the software they currently use, and that building without open source would cost organisations about 3.5 times more. 1 The precise number matters less to me than the asymmetry it exposes: value creation in open source can be enormous while direct payment for the underlying software remains zero.

That means revenue is not a complete measure of whether value exists. It is a measure of how much of a particular kind of value the company has chosen, and managed to capture.

This is also why I am wary of treating a large free user base as unrealised revenue. Some of that value is supposed to remain distributed. Open access can create experimentation, familiarity, contribution, integrations, advocacy and trust. Harvard researchers Sonali Shah and Frank Nagle describe user communities as a distinct strategic structure for exchanging ideas and knowledge, rather than simply another marketing audience. 2


More recent Linux Foundation research makes the commercial relationship even more interesting: its 2025 analysis of 800 venture-backed startups found a close link between community health and company valuation, suggesting that community value and business value can reinforce one another rather than existing in opposition. 3

This sounds obvious, but it changes the framing of the business problem.

Harvard Business School research on the economic value of open-source software estimates that firms would have to spend roughly $8.8 trillion to recreate the software they currently use, and that building without open source would cost organisations about 3.5 times more. 1 The precise number matters less to me than the asymmetry it exposes: value creation in open source can be enormous while direct payment for the underlying software remains zero.

That means revenue is not a complete measure of whether value exists. It is a measure of how much of a particular kind of value the company has chosen, and managed to capture.

This is also why I am wary of treating a large free user base as unrealised revenue. Some of that value is supposed to remain distributed. Open access can create experimentation, familiarity, contribution, integrations, advocacy and trust. Harvard researchers Sonali Shah and Frank Nagle describe user communities as a distinct strategic structure for exchanging ideas and knowledge, rather than simply another marketing audience. 2


More recent Linux Foundation research makes the commercial relationship even more interesting: its 2025 analysis of 800 venture-backed startups found a close link between community health and company valuation, suggesting that community value and business value can reinforce one another rather than existing in opposition. 3

If the commons is part of the value engine, weakening it to create scarcity can destroy some of the thing you are trying to monetise.

If the commons is part of the value engine, weakening it to create scarcity can destroy some of the thing you are trying to monetise.

02– WHAT'S THE BOUNDARY?

I stopped asking “what can we charge for?” and started asking “where does the problem change?”

I stopped asking “what can we charge for?” and started asking “where does the problem change?”

There are many viable open-source business models. Harvard Business School's recent industry note describes open-source businesses as a group of commercial models, not a single formula. 4 Therefore, I do not believe there is a universal rule for which features should be free or paid.


That said, I have observed a pattern that consistently helps clarify these decisions.

The challenge often shifts when software transitions from being managed by an individual or small team to becoming critical infrastructure for an organisation.

There are many viable open-source business models. Harvard Business School's recent industry note describes open-source businesses as a group of commercial models, not a single formula. 4 Therefore, I do not believe there is a universal rule for which features should be free or paid.


That said, I have observed a pattern that consistently helps clarify these decisions.

The challenge often shifts when software transitions from being managed by an individual or small team to becoming critical infrastructure for an organisation.

Four value layers

Four value layers

*This is not a pricing ladder but instead a way to locate the problem

*This is not a pricing ladder but instead a way to locate the problem

01 Commons

Access & utility

The product solves a real problem and remains usable, inspectable and adaptable. This is where experimentation and initial trust can begin.

02 Ecosystem

Participation & knowledge

Users, contributors and partners add documentation, integrations, feedback, advocacy and know-how that make the system more valuable.

03 Organisation

Scale & coordination

Deployment creates new needs around governance, identity, security, lifecycle, administration, procurement and internal adoption.

04 Commercial

Assurance & accountability

The company takes responsibility for reducing risk, maintaining reliability, supporting operations and standing behind the product in ways the commons cannot be expected to.

In my opinion, the key transition occurs between the third and fourth layers.

An organisation does not always pay because it suddenly wants access to something technically unavailable. It may pay because the consequences of failure have changed. A hobby project can tolerate uncertainty. A production system serving thousands of employees may need someone accountable for security updates, compatibility, support, recovery and long-term operation.

This shift introduces a new set of challenges that require a different approach.

03 – RESPONSIBILITY

The most durable paid boundary may be where the company agrees to stand behind the system.

The most durable paid boundary may be where the company agrees to stand behind the system.

This is the idea I keep returning to: rather than seeing monetisation as just locking features, I find it more helpful to think of it as a responsibility boundary.

Increasing organisational dependence → increasing need for accountability

01
Use
I can run it and get value from it myself.
02
Adoption
My team begins coordinating around it.
03
Dependance
Work, data or operations now rely on it.
04
Accountability
Someone must be accountable when reliability, security or continuity matters.
01
Use
I can run it and get value from it myself.
02
Adoption
My team begins coordinating around it.
03
Dependance
Work, data or operations now rely on it.
04
Accountability
Someone must be accountable when reliability, security or continuity matters.
01
Use
I can run it and get value from it myself.
02
Adoption
My team begins coordinating around it.
03
Dependance
Work, data or operations now rely on it.
04
Accountability
Someone must be accountable when reliability, security or continuity matters.

Red Hat stands out as a strong example, as its subscription model recognises that the open-source code remains accessible. Instead, Red Hat’s own description of the subscription highlights the value in production-ready software, lifecycle management, security, interoperability, certifications, updates, and expert support. 5


The subscription's value goes beyond simply providing access to the software. Red Hat positions itself as a partner in making the software reliable for your organisation, and commits to supporting you throughout that relationship.


This distinction helped me put into words something I had felt before but wasn’t really able to name:

Red Hat stands out as a strong example, as its subscription model recognises that the open-source code remains accessible. Instead, Red Hat’s own description of the subscription highlights the value in production-ready software, lifecycle management, security, interoperability, certifications, updates, and expert support. 5


The subscription's value goes beyond simply providing access to the software. Red Hat positions itself as a partner in making the software reliable for your organisation, and commits to supporting you throughout that relationship.


This distinction helped me put into words something I had felt before but wasn’t really able to name:

Open access creates possibility.
Commercial responsibility creates confidence.

Confidence can be worth paying for when uncertainty becomes expensive.

04 – DIFFERENT MODELS

Responsibility does not have to be monetised in only one way.

Responsibility does not have to be monetised in only one way.

I am not suggesting that Red Hat's model should apply to every situation. Each product leads to its own costs, competition, and customer needs.


For example, GitLab has described itself as an open-core business. It offered an open-source product along with a paid Enterprise Edition sold by subscription. 6 This means the main difference between free and paid versions was in product features, not just in services or support.


These are very different approaches. Both show that asking “how does open source make money?” is too broad of a question.

I am not suggesting that Red Hat's model should apply to every situation. Each product leads to its own costs, competition, and customer needs.


For example, GitLab has described itself as an open-core business. It offered an open-source product along with a paid Enterprise Edition sold by subscription. 6 This means the main difference between free and paid versions was in product features, not just in services or support.


These are very different approaches. Both show that asking “how does open source make money?” is too broad of a question.

Two contrasting boundary choices

Two contrasting boundary choices

for example only

for example only

Subscription / assurance

Keep the software open.
Monetise confidence around operating it.

Customers pay for tested releases, long-term maintenance, support, security, certification, interoperability and a supplier relationship.

Illustrated by Red Hat's subscription model.
Their commercial value sits heavily in assurance, lifecycle and accountability.

Illustrated by Red Hat's subscription model.
Their commercial value sits heavily in assurance, lifecycle and accountability.

Question I learned to ask

What new problem appears when usage becomes organisational?

The boundary is partly product-based: the open project continues to provide meaningful utility while proprietary capabilities address additional organisational needs.

Historically used by GitLab, where the strategic challenge is deciding which capabilities strengthen the paid proposition without hollowing out the open core.

Historically used by GitLab, where the strategic challenge is deciding which capabilities strengthen the paid proposition without hollowing out the open core.

The issue is not about moral or commercial superiority. Much rather, the boundary itself represents a strategic design decision.

After I understood this, I stopped searching for a single open-source monetisation model. Instead, I began to look for the connections between four things: the commons, the user's problem, the organisation's new problem, and the unique responsibility the company can take on.

05 – MANUFACTURED SCARCITY – (A SUBTRACTIVE BUISNESS MODEL)

05 – MANUFACTURED SCARCITY
(A SUBTRACTIVE BUISNESS MODEL)

This is where my view goes beyond simply describing existing business models.

As monetisation that works primarily by making the free experience worse is quite suspicious.

As monetisation that works primarily by making the free experience worse is quite suspicious.

Every business needs economic sustainability. Open source is not exempt from salaries, infrastructure, security work, support costs or the simple fact that maintainers need time and resources. The Linux Foundation explicitly frames sustainability as a cycle in which commercial products generate profits that can be reinvested into open projects and communities. 7


But there is a difference between creating a paid layer and manufacturing pain to force conversion.

If a capability is central to why the community adopted, trusted or contributed to the project, deliberately degrading that value can create revenue in the short term while weakening distribution, goodwill or participation in the long term. However, that does not mean every feature must remain free forever, it instead means that the second-order effects matter more than we realise.

So the principle question that needs to be asked is:

Every business needs economic sustainability. Open source is not exempt from salaries, infrastructure, security work, support costs or the simple fact that maintainers need time and resources. The Linux Foundation explicitly frames sustainability as a cycle in which commercial products generate profits that can be reinvested into open projects and communities. 7


But there is a difference between creating a paid layer and manufacturing pain to force conversion.

If a capability is central to why the community adopted, trusted or contributed to the project, deliberately degrading that value can create revenue in the short term while weakening distribution, goodwill or participation in the long term. However, that does not mean every feature must remain free forever, it instead means that the second-order effects matter more than we realise.

So the principle question that needs to be asked is:

Does the paid layer create new value because the customer's responsibility has changed?

or

Does it only create scarcity because the company needs something to sell?

Does the paid layer create new value because the customer's responsibility has changed?

or

Does it only create scarcity because the company needs something to sell?

The first represents a value proposition, while the second may risk undermining existing trust.

06 – A WORKING FRAMEWORK

Five questions I would now use as guidelines to design the commercial boundary.

Five questions I would now use as guidelines to design the commercial boundary.

These questions are designed to surface the trade-offs involved, rather than to deliver a mechanically correct answer.

01

What value must remain open for the ecosystem to keep working?

Identify the utility, interoperability, experimentation or participation that drives adoption and community health.
Treat these as strategic assets, not automatically as unmonetised inventory.

02

What new problem appears when usage becomes organisational?

Look for needs created by scale, coordination, governance, risk, compliance, reliability, deployment and procurement rather than simply asking which existing feature could be gated.

03

What responsibility can the commercial organisation credibly take on?

The strongest paid value may come from commitments the community cannot reasonably be expected to guarantee: support, lifecycle, assurance, administration, accountability or organisation-specific capability

04

Is there an economic owner for that problem?

A monetisable problem needs someone with budget, authority and a consequence they are trying to avoid or an outcome they are trying to achieve.

05

Does the paid model reinforce the commons or quietly consume it?

Look beyond conversion. Find out if whether revenue creates resources, trust or capability that feed back into the wider ecosystem and what could happen to community behaviour if the boundary moves.

In my experience, conventional monetisation discussions tend to overlook the last question .


If community health and commercial value can reinforce each other, as the Linux Foundation's recent commercial open-source research suggests, then protecting the commons is not necessarily philanthropy sitting outside the business model 3 , it can be part of the commercial strategy itself.

07

The business does not need to monetise all the value it creates.

The business does not need to monetise all the value it creates.

This distinction has significantly influenced my current perspective.

In conventional businesses, uncaptured value is often seen as leakage. Within open source, however, the presence of uncaptured value can be fundamental to the system’s overall value and resilience.

The strategic focus shifts to understanding which value should be captured and which should remain open in order to sustain the ecosystem.

The aim would be to create a fair and healthy exchange among everyone involved.

The commons provides accessible utility, encourages experimentation, and invites participation, and the broader ecosystem extends the product’s reach, deepens knowledge, and builds legitimacy. As organisations adopt the product, new complexities and risks emerge. The commercial entity is then positioned to address these higher-order challenges.

Now, revenue enables ongoing investment in the product, strengthens customer relationships, and supports the ecosystem that underpins long-term success.

What I would takeaway

The question is not how much of the commons a company can successfully monetise. It is what new responsibility the commercial organisation can take on that makes both the customer and the commons stronger.

As it should already be apparent:
I included research in this piece because the first two essays in this series focused mostly on my own experiences. For this essay, I want to document how my ideas compare with those of researchers and companies that have spent more time studying open-source economics. The framework above is my synthesis; it is not presented as a framework from Harvard, the Linux Foundation, Red Hat, or GitLab.

I chose examples that show clear differences and not to cover every possible case. Eg., licensing, hosting, dual licensing, services, sponsorship, managed cloud, and other models all set different boundaries. My goal is not to rank these models, but to help readers think about the boundaries they create.

Research & references

Research & references

01

Harvard Business School — “Open Source Software: The $9 Trillion Resource Companies Take for Granted”

Summary of research by Manuel Hoffmann, Frank Nagle and Yanuo Zhou on the economic value of open-source software.

Summary of research by Manuel Hoffmann, Frank Nagle and Yanuo Zhou on the economic value of open-source software.

Summary of research by Manuel Hoffmann, Frank Nagle and Yanuo Zhou on the economic value of open-source software.

Read Source

Read Source

02

Harvard Business School — “Why Do User Communities Matter for Strategy?”

Sonali K. Shah and Frank Nagle on user communities as strategic structures for knowledge exchange, innovation and firm strategy.

Sonali K. Shah and Frank Nagle on user communities as strategic structures for knowledge exchange, innovation and firm strategy.

Sonali K. Shah and Frank Nagle on user communities as strategic structures for knowledge exchange, innovation and firm strategy.

Read Source

Read Source

03

Linux Foundation Research — The State of Commercial Open Source 2025

Analysis of 25 years of venture data across 800 VC-backed startups, including the relationship between community health and company value.

Analysis of 25 years of venture data across 800 VC-backed startups, including the relationship between community health and company value.

Analysis of 25 years of venture data across 800 VC-backed startups, including the relationship between community health and company value.

Read Source

Read Source

04

Harvard Business School — Open Source Software and Hardware Business Models

2025 industry/background note by Frank Nagle and Richie Zitomer describing the diversity of commercial models around open technologies.

2025 industry/background note by Frank Nagle and Richie Zitomer describing the diversity of commercial models around open technologies.

2025 industry/background note by Frank Nagle and Richie Zitomer describing the diversity of commercial models around open technologies.

Read Source

Read Source

05

Red Hat — Subscription Model FAQ

Primary-source description of how Red Hat creates paid value through production readiness, lifecycle management, security, certification, interoperability and support.

Primary-source description of how Red Hat creates paid value through production readiness, lifecycle management, security, certification, interoperability and support.

Primary-source description of how Red Hat creates paid value through production readiness, lifecycle management, security, certification, interoperability and support.

Read Source

Read Source

06

GitLab — Letter from Shareholders

A historical primary-source description of GitLab's open-core model and Enterprise Edition subscription strategy.

A historical primary-source description of GitLab's open-core model and Enterprise Edition subscription strategy.

A historical primary-source description of GitLab's open-core model and Enterprise Edition subscription strategy.

Read Source

Read Source

07

Linux Foundation — Setting an Open Source Strategy

Guidance on sustainability, commercial dependencies and the reinvestment cycle between products and open-source project communities.

Guidance on sustainability, commercial dependencies and the reinvestment cycle between products and open-source project communities.

Guidance on sustainability, commercial dependencies and the reinvestment cycle between products and open-source project communities.

Read Source

Read Source

These reflections come from my own experience working on growth and strategy around Open WebUI. They describe how my thinking changed during that work, rather than the company's current commercial strategy or an official view on how open-source companies should monetise.

Previously in this series

How I used service-design thinking to make an early commercial motion more deliberate, clarifying the problem, mapping the buying journey and building a growth system designed to learn.