• 

  Andrew ParkEditorial Lead
  Heavybit

Open Source vs. Proprietary: Development, Licensing, Business Models, Security, and More

Open-source software (OSS) is free to use, distribute, and inspect (depending on the licensing fine print), while proprietary software must be paid for up-front and is covered by much stricter licenses. Still, open-source software (OSS) is gaining incredible traction in the modern software marketplace, despite a massive field of proprietary software products continuing to see considerable success.

This article will briefly touch on the topline differences between the types of software, and dive deeply into what startups and technical teams need to know about the differences in how to develop, license, and go to market with OSS vs. closed-source software products.

In addition, we’ll also cover the impact of AI on open source and how new AI developments may affect the choices of founders to pursue open-source software vs. proprietary projects and tooling. AI is reportedly helping devs accelerate the coding, review, deployment, and maintenance of their software projects, but it also poses new challenges to open-source projects, particularly in terms of privacy and IP issues.

For more insights on the challenges and trends in modern OSS, join the DevGuild: Open Source event to learn from leading startup founders and investors in the space, or read this guide how to start an open-source project.

Key Differences Between Open-Source Software and Proprietary Software

Open-source software is distributed with its source code, and lets users freely inspect, modify, and enhance it per its licensing agreement, while proprietary software is the copyrighted property of an individual or company that limits its use, distribution, and modification by way of a proprietary license.

Prominent examples of open-source software include the web browser Firefox, the content management system WordPress (and its many, many plugins), and the open-source operating systems Android and Linux (including the popular Ubuntu distribution for Linux).

Most (though not all) closed-source projects tend to be commercial apps–some examples of proprietary software include the Microsoft Office productivity suite, Adobe’s creative cloud, and commercial operating systems such as MacOS and Windows.

Startups and technical teams considering open-source versus proprietary tooling should consider the following fundamental differences:

Open-Source Software Differences:

• Development: Community based
• Licensing: Copyleft or permissive, allowing some level of modification and distribution
• Security: Dependent on open-source community participation to manage vulnerabilities
• Flexibility: Significant flexibility with the freedom to inspect and modify OSS
• Maintenance: Community-driven across the software’s lifetime per the community’s priorities (with bug fixes potentially being provided by only a handful of dedicated community members)
• Support: Not provided as a service, and generally requires independently checking the documentation or crowdsourcing across the community
• Cost to Use: OSS generally has no initial fee, except in the case of companies using a SaaS model–see below. Over time, relying on OSS may incur ongoing costs in developer maintenance.
• Ease of Use: Getting started working with a new OSS project for the first time may depend entirely on how much clear community documentation is available for beginners

Proprietary Software Differences:

• Development: Closed-source, usually by a professional team
• Licensing: Restrictive, usually forbidding modification and distribution
• Security: Usually company-driven and requires professional development resources
• Flexibility: Restrictive licenses mean limited customization, preventing inspecting and modifying code
• Maintenance: Regular updates provided by the software’s publisher across its lifetime in response to customer needs and market conditions
• Support: Provided by the software’s publisher in direct response to customer needs
• Cost to Use: Typically requires purchase/subscription, particularly for commercial apps
• Ease of Use: Software vendors are incentivized to focus on user interface and user experience to encourage continued use

The Benefits of Open-Source Software

It’s impossible to ignore the increasing popularity of open-source software (OSS) despite the enormous market share of well-known proprietary software brands. As of 2015, 66% of companies reportedly created software for customers built on OSS, with 78% of companies running at least part of their infrastructure on OSS components. More-recent reports suggest that developers build up to 70%-90% of modern software on OSS.

So, why is open-source software ascendant? As we cover in our article on the benefits of open-source software, OSS offers many important advantages to developers and startup founders, including faster project start time, community-driven collaboration, no initial contract lock-in to a vendor, and lower initial costs including a typical lack of licensing fees–though over time, relying on open-source code rather than homegrown will tend to cost additional developer hours as teams invest extra cycles to improve compatibility by adapting OSS projects to their own specific use cases and infrastructure.

For all its benefits, OSS may not always inherently be a “better” way to deliver software for every possible use case. There is still a place for proprietary software in the modern business landscape–reportedly, about 45% of all software tools that companies use are proprietary. Next, we’ll cover some of the most significant differences between the two software models, and what they potentially mean for tech leaders and startup founders.

Differences: Developing Software Projects

As you might expect, software development for OSS projects differs drastically from development processes for proprietary software, particularly since the majority of closed-source products come from commercial, for-profit businesses.

Open-Source Software Development

OSS development tends to:

Offer a Faster, Lower-Friction Start

While established open-source projects have many important aspects (including project documentation and a highly-engaged community), to start a new project, you don’t need much more than:

• Your Project’s License - Which stipulates usage, distribution, and modification
• Your Project’s README File - Which acts as an initial set of documentation and guidelines for the project
• Guidelines for Contribution - To specify how people can contribute to your project
• Code of Conduct - To establish expectations for how participants should conduct themselves in what will hopefully become a vibrant community

Offer a Lower-Cost Entry

Over time, long-lived OSS projects that organizations pivot into commercial projects incur the kinds of business expenses you’d expect, including incorporation and hiring. And incorporating OSS projects into your existing startup’s operations also carries a lifetime cost of ownership as your team adapts and maintains those projects to play nicely with your infrastructure.

However, starting a new open-source project technically does not require any initial investment–no hired team of developers, no workstations, and likely no commercial software.

Offer Potentially Faster Development Cycles

Over time, development on OSS projects can frequently accelerate quickly as members of a growing community continuously make improvements and innovations independent of any formal product roadmap.

Depend on Building a Community

A foundational, non-negotiable step is building an open-source community for open collaboration. A strong, vibrant community serves many important purposes in an open-source project, acting not only as contributors, but as testers, QA experts, and documenters, and even product managers who may actively determine the development of new features, and even as evangelists who bring new contributors into the fold.

Proprietary Software Development

Proprietary software development tends to:

Begin With Potentially Fewer Resources:

Some of the biggest success stories in proprietary software famously began inside basements as the result of conversations between a few genius computer programmers. Proprietary software projects inherently begin their lives without a massive open-source community, or additional partners who are either directly compensated or on the hook for equity in an eventual startup built around the product.

Proprietary software projects tend to begin their lives as a minimum viable product (MVP) lovingly crafted by a small but dedicated group of developers (or a single, lone developer), creating everything–code, documentation, logs–by themselves. They can apply for venture capital funding, but in any case, their initial resources will be spread thinner across hardware, tooling, office space, and all the usual startup concerns.

Have a Relatively Small Team for Some Time

The most successful open-source communities grow to massive sizes over the years. GitHub reportedly has more than 100M users. A proprietary software project, particularly a commercial one, will have a much smaller initial team that will likely not grow much bigger for some time.

Having more developers gives teams the bandwidth to not only create more code faster but also the ability to improve their projects in multiple ways by assigning team specialists to important areas such as security, testing, and managing tools and infrastructure.

Startups working on proprietary software rely on bootstrapped budgets or VC fundraises–and in each case, they’ll have fewer overall developers on their project than a vibrant OSS community, having to manage hiring as yet another expense on their early startup budget. Fewer devs means less flexibility, more team members wearing multiple hats, and slower going.

Require an Investment up Front

Technical teams adding proprietary commercial tools to their stack can expect to pay upfront costs for contracts/seats. Founders looking to build a commercial proprietary software project may eventually need to build a company around said project, and will need to invest in traditional corporate overhead, such as incorporation and legal fees to protect their intellectual property. Founders that go the route of founding a startup will need to raise capital.

Manage Expectations Around Interoperability and Integrations

A proprietary software product that customers can use only after paying for it carries implicit expectations. For example, some reports suggest that more than 80% of companies consider product integrations a top priority.

Customers will expect to use your product in tandem with the rest of their stack, and they’ll increasingly expect it to play nicely with their other tooling, without the need for costly third-party solutions. In addition to the rest of the features on your emerging product roadmap, your startup may face increasing pressure to deliver integrations with your customers’ tools.

More Resources on Development

Learn more about open-source software development in the Heavybit Library:

• PODCAST: A Closer Look at Open Source Today with Paul Biggar and Edith Harbaugh
• ARTICLE: Project Management for Open-Source Projects by Ev Haus
• ARTICLE: How to Hire a Startup Lawyer by Rebecca Lee Whiting
• VIDEO: Security Process in Open Source Communities with Alex Gaynor

Differences: Licensing

Given the relatively restrictive distribution and modification requirements, particularly for commercial apps, licensing tends to be a bit more straightforward for proprietary software projects. In contrast, there’s a variety of popular open-source licenses already in use, and for developers who pivot an OSS project into a commercial product, licensing can become more complicated.

Open-Source Software Licensing

Open-source licenses are usually either copyleft–meaning they inherit the licensing terms of the previously-existing OSS project on which their project is based; or permissive–which provides more flexibility for distribution, modification, and reuse. (To briefly make the distinction, free and open-source software are not exactly identical, as free software is completely free to inspect, modify, and do with as users see fit, without any type of license restriction.)

Regardless of the specific licensing model its creators choose, open-source licenses generally tend to share a few common points:

• Storage in a publicly-shared repository - Generally speaking, open-source licenses require storing the source code in a publicly-accessible location.
• Document changes in source code - Open-source licenses also generally require documentation of any changes made to source code, and by what methods.

Examples of Copyleft Licenses:

• GNU General Public License (GPL): The original copyleft license crafted by open-source pioneer Richard Stallman, the GPL stipulates that when distributing any software built on GNU-licensed OSS, creators must also distribute the source code of their projects together.
• GPL variants, including Affero GPL (AGPL) and Lesser GPL (LGPL): AGPL closes a distribution loophole in the original GPL, making its restrictions apply to software made available via network as well as through general distribution. LGPL is largely similar to GPL and AGPL but relaxes distribution requirements for smaller projects accessed via licensed projects.
• Mozilla Public License (MPL): The MPL is perhaps the least restrictive copyleft license since it lets developers modify and use any code from MPL-licensed projects, so long as those developers keep the modified files separate and distribute them along with their software project.

Examples of Permissive Licenses:

• Apache License: The Apache License requires copyright and license notices, but otherwise allows for different licensing terms in derivative projects and doesn’t require creators to provide the source code of their work.
• MIT License: One of the most popular permissive OSS licenses, the MIT License simply requires that developers include the original license and copyright notice in any derivative products.Berkeley Source Distribution (BSD) License: BSD Licenses also require copyright and license notices but do not require source code distribution, with various additional restrictions in the 2-Clause, 3-Clause, and 4-Clause variants.

Proprietary Software Licenses

Proprietary software licenses usually restrict modification, inspection, and distribution. Creators of proprietary software projects usually don’t distribute the source code for their projects, and attempting to distribute commercial proprietary software without a reselling or distribution license is usually illegal. Proprietary software licenses are standard for commercial apps. Here are some of the more-common variants:

End-User License Agreements (EULA)

A software EULA tends to be a “clickwrap” agreement that users accept during the setup process that requires users to accept privacy and copyright policies before use. The primary purpose of a commercial app’s EULA is to set forth guidelines on what users can and can’t do with the app.

Specifically, to ensure they do not duplicate, reverse-engineer, or distribute copies of it. The EULA tends to be most relevant for single-user consumer applications, rather than for developer-first apps that provide ongoing service.

Terms of Service (ToS) / Terms of Use (ToU) / Terms and Conditions (T&C)

A ToS agreement is more common among business and developer-first software apps, and stipulates services a software publisher will provide to an end-user, provided users also comply. A ToS might include terms such as eligibility, content contained with the app and scope of services, billing and payment, liability, and dispute resolution.

Service-Level Agreements (SLA)

An SLA acts as a commitment from a software publisher to a customer that focuses on the service that the publisher provides, particularly in terms of quality and reliability. SLA agreements might include terms such as uptime, the length of time for an issue resolution window, and monitoring and reporting.

More Resources on Licensing

Learn more about software licensing in the Heavybit Library:

• PODCAST: Standing on the Shoulders of a License with Ilan Rabinovitch
• ARTICLE: What to Know About Open-Source Licenses

Differences: Business Models (and Additional Considerations)

The “software business” as we know it consists primarily of companies selling commercial versions of proprietary software apps. And while traditionally, open-source projects don’t begin their lives as commercial products, several have pivoted or spun off into commercial software offerings with their own ways to generate revenue. As you might expect, a great many of the thousands of net-new software products each year come from startup companies built around bringing them to market.

Open-Source Business Models

Since open-source software typically starts as a non-commercial, community-based project, OSS startups will frequently pivot from open source to a commercial business model.

These are some of the most common business models for OSS startups:

Professional services (ProServ)

Companies such as Red Hat, IBM, and Chef offer professional services, which include for-pay training, technical support, and consulting from solutions teams sharing their technical expertise to enhance a customer’s ability to use an open-source product.

Open core / “freemium”

Companies such as Docker, GitLab, and MongoDB use the open core model, which requires publishers to maintain multiple codebases, including a “core” codebase that remains open source, and a gated commercial codebase that offers additional premium functionality as “upgrades” for paying customers.

SaaS

Companies such as GitHub and Automattic (the vendor of WordPress) use the SaaS business model, which requires customers to accept a binding contract, but also tends to offer robust, dedicated support (training, troubleshooting, and one-on-one support hours) for their open-source solutions.

Donations/foundation

Perhaps the most well-known examples of companies that use the charitable foundation model are the Mozilla Foundation and Wikimedia (provider of Wikipedia). Rather than charge any fees for products or services, such organizations rely on charitable donations to continue their operations.

Additional Considerations for Teams Adopting Open-Source Tooling

If you’re leading a team that’s looking at adding OSS to its stack, here are some points to consider:

• Licensing Nuances: Pay careful attention to your OSS tools’ licensing–you may have restrictions on what you can and can’t build and sell.
• Licensing Changes: Prominent OSS orgs like HashiCorp, Elastic, and MongoDB have suddenly changed their license structure–your next OSS tools may as well.
• Following Community Updates: Team members will frequently need to invest extra cycles in following the latest changes in OSS projects.

Additional OSS Business Considerations for Startups:

• Managing multiple codebases with open core: Open-core projects eventually require their publishers to maintain separate core and commercial codebases, increasing the workload for development, testing, and deployment across two or more separate repositories while ensuring that gated features don’t accidentally end up in the core code.
• Maintaining robust core and commercial versions with open core: Organizations that do manage multiple codebases may find themselves making difficult decisions about which features to “gate” within commercial offerings and which to keep as part of the core codebase. “Giving away too much functionality” in the core offering arguably keeps the community vibrant but potentially devalues the commercial offering, while gating too many features may make the commercial offering seem more valuable, but potentially jeopardizes the health of a community that dislikes being denied important functionality behind a paywall.
• Potentially challenging pricing and packaging: Pricing for COSS products can be anything but straightforward, especially as a COSS startup scales and potentially scales and begins selling to larger, enterprise customers in deals that bring in more revenue, but also may need to invest significantly into staffing up with additional hires to provide pre-sales and post-sales support.
• Growing and maintaining a community: In all cases, OSS startups must keep the community alive to ensure the project remains relevant and that potential users can discover it both through traditional marketing and developer relations as well as through community word-of-mouth, forums, chat channels, or even live events. While some OSS communities are enormous, in many cases, much of the maintenance and updates come from a handful of engaged users who are the key to keeping the project alive.

Proprietary Software Business Models

While successfully selling proprietary software is by no means easy, it tends to have a more-straightforward path to market as it goes commercial. Even early-stage proprietary software companies will start with a basic sales function (even if the founders themselves have to perform double-duty as salespeople), and may also add early-stage marketing teams to attract more potential customers. You’ll generally find startups selling proprietary software products following one (or both) of these two paths for their go-to-market (GTM) motion:

• Inbound: The inbound model relies on bringing interest to your business. For early-stage startups, it’s common to use your website as a center of gravity for inbound and attempt to attract interested leads via search engine optimization (SEO) of content on your website and promotion through other channels, including social media, video, and podcasts.
• Outbound: As the name suggests, this model relies on having sales professionals perform outreach to pitch their products to prospects. In the earliest stages, it’s common for software companies to focus on closing early sales deals via networks and references, and eventually scale up to having coverage across all stages of their sales funnel, including marketing supporting the top-of-funnel (TOFU); “inside sales,” also known as business development representatives (BDR) or sales development representatives (SDR) to qualify leads around the middle-of-funnel (MOFU); and “outside sales,” also known as account executives (AE), who manage active discovery, demo, and closing deals (hopefully as “closed won deals”) at the bottom-of-funnel (BOFU).

Additional Considerations for Teams Considering Adopting Proprietary Tooling

If your team is looking at purchasing proprietary software tooling, here are some important considerations:

• Integrations: Managing data across apps is crucial. New apps that lack integrations to your tools may require hours of tedious copy-pasting.
• Negotiating Terms: Most software is sold with flexible terms. It can be worthwhile to negotiate items like price, contract length, and support options.
• Vendor Lock-In: Proprietary software products “lock you in” for months/years after you sign a contract, limiting your ability to switch.

Additional Proprietary Software Business Considerations for Startups

• Scaling can depend on your ability to hire: Those looking to create new commercial closed-source products not only need to budget for the salaries and benefits of new developers to employ on their projects, but also for the sourcing, interviewing, and hiring of said developers.
• Finding “product-market fit” may depend on design partners: The most successful commercial open-source projects tend to begin as clear, distinctive, and valuable solutions to a significant problem many people experience. In a best-case scenario, an OSS community forges a robust project over time with continuous iteration to create something that real human beings genuinely need.

Startups launching proprietary software apps don’t benefit from the communal testing and feedback OSS projects typically provide, so they may benefit from working with design partners–trusted early adopters that can provide clear feedback on how to create a better, more-usable product while honing in on the business problem your software should solve.

Learn more about business models and other software startup challenges in the Heavybit library:

• VIDEO: Commercial Open-Source Business Strategies with Sid Sijbrandij
• VIDEO: Open-Source Business Strategy with Peter Reinhardt
• VIDEO: Lessons on Pricing and Selling Open Core with Aghi Marietti
• VIDEO: Value-Based Pricing with Christina Noren
• VIDEO: Pricing a New Product: Adventures in Research & Rollouts
• ARTICLE: Early-Stage Sales by Nick Beecroft

Open-Source Versus Proprietary: AI Considerations

Modern-day AI (specifically, generative AI–which spins up massive amounts of content in response to prompts fed to a large-language model) is a popular topic that has made its way into conversations about open-source projects.

Below, we’ll cover what teams utilizing open-source AI products vs. proprietary AI products need to know.

What to Consider About Adopting Open-Source AI Products for Teams

Technical teams thinking about adding open-source AI tools, such as open language models, open machine learning platforms, and other open frameworks should account for these potential issues:

Lack of Support

Open-source AI projects are like other open-source projects–they belong to the communities that built and participate in them. Since open-source support is community support, f you have questions, be prepared to wade through documentation or (hopefully) ask in an active community space. The answers you get may come from enthusiasts rather than subject matter experts who are paid to solve your problems.

Lack of a Machine Learning PhD

AI/ML tooling for developers is advancing rapidly, but there are still many fundamental concepts–and many more advanced concepts–which means solving for the specific needs of your use case could be the domain of those with academic training. If you never got that PhD in machine learning, your team’s learning journey may be that much longer.

A La Carte: Many Tools In Various States

Unlike a finely-crafted premium product, open-source AI projects are provided as-is. Some may be supported by a vibrant community of developers that dutifully document everything and generously share updates. Some may have been built by a PhD candidate years ago and never updated to be production-ready.

Evaluation

There are many open models out there waiting to be tried. Which actually makes sense for your use case? You’ll need to evaluate the models you’re looking at to ensure it’s a good fit, using metrics like relevancy, coherence, accuracy, and hallucinations.

Privacy Issues

Generative AI is hungry for data to train on, and feeding data into an AI model could cause it to incorporate your private data into its training corpus. As of this writing, there’s no universal privacy standard for AI projects, so you’ll have to read the fine print.

Hidden Costs

Open-source AI products, like LLMs, don’t have an initial sticker price. At a glance, they’re cost-effective. However, the cost of running inference jobs on datasets, particularly at a large scale, isn’t free. The bigger the job, the bigger the cost.

What to Consider About Adopting Proprietary AI Products for Teams

Teams looking to add proprietary AI tools to their toolchain may want to consider these issues:

Evaluation, Again

Not all proprietary AI tools are equal–you’ll want to do your due diligence before purchasing one, just like with any proprietary software product. Particularly since you’ll be locked in with your vendor after you sign the contract.

Privacy, Again

Most proprietary AI providers at least have some kind of contractual fine print around data privacy–in many cases, for your protection as a customer, but also for their own protection should their products unwittingly get hacked or leak your data due to unforeseen circumstances.

Hidden Costs, Again

After signing up with the vendor of a proprietary AI product, particularly one that can run high-volume jobs (such as a LLM), if you’re not careful, you may find yourself looking at an eye-watering bill. Third-party API vendors of LLMs in particular are aware of price sensitivities and frequently offer new improvements to cut down on token costs, but be aware that untethered data jobs may lead to big bills.

What’s Happening With the Vendor Itself

Without naming names, we’ll politely mention again that the AI space is dynamic and exciting. Which you would expect from the billions of dollars in investment monies sloshing around. As a result, major shakeups can and have happened among AI vendors. For example, executives and key team members have departed companies they founded, sometimes coming back.

Conclusion and Takeaways

If you’re a startup founder, or considering starting a new project, you probably already have a fairly clear idea about whether going open source or proprietary is “right” for you. With any luck, this article should hopefully provide more clarity into how each path differs in terms of development, licensing, and GTM strategy.

And as exciting new technologies like AI emerge, it’s also important to have a firm grasp of the potential risks and hidden costs associated with them–whether for open-source or for proprietary products.

Learn more about the state of open source and future opportunities and challenges at the DevGuild: Open Source event.

Additional Resources

• Opensource.Guide - Starting an Open Source Project
• The Register - Enterprise Open-Source vs. Proprietary Software
• Snyk’s Guide to Open-Source Licenses
• Thomson Reuters - Software License Agreements