What is the difference between SaaS and Open Core?

My previous post in this series was a brief introduction of SaaS and Open Core business models in the B2B space. To summarize: I concluded that SaaS is the dominant and default model for selling software but there are signs that a new model is getting traction, Commercial Open Source Software (“COSS”), of which a common monetization strategy is Open Core (proprietary features on the crust of an open core).

I also received some feedback on that post that the SaaS and COSS models are orthogonal and you could do both. This is quite true! However, when starting a company or even an idea, it is usually wise to be laser focused on what you are building.

As an example, we can use Teleport (our modern replacement for OpenSSH). It is open source software that is monetized through an Open Core model. One of its most desirable features is the unified audit log with session replay. The storage of this audit log is configurable and this configurability comes with increased development costs and design trade-offs.

If Teleport was a SaaS product, we could have picked a suitable storage back-end from the vast menu of options available on AWS. On the other hand, we would have had to develop a secure, multi-tenant control plane to allow thousands of customers to manage SSH and Kubernetes access to their environments.

Doing all of these things initially would be expensive. So most startups or developers make the correct decision of starting with either a proprietary SaaS or an open source, installable / on-prem offering. Proprietary, installable software may also be an option but I don’t see that as a popular choice these days.

Generally speaking, in the B2B market, the advantage of the Open Source model is that it is easier to build awareness; the advantage of the proprietary, SaaS model is that it is easier to monetize.

Once an initial offering has some traction, it may be wise to attempt to take advantage of the other business model’s strengths:

• Open Source starts to offers additional proprietary features on the crust of the open source core as a monetization strategy (eg. MapR Enterprise, InfluxDB Enterprise).
• Perhaps to a lesser extent, but some B2B SaaS companies open source ancillary tools built to deliver the solution in order to build awareness for product marketing or recruiting (eg. Twilio Open Source Projects, Segment Open Source). This practice is much more common with large B2C companies and may even become the foundation of future Open Core companies (eg. Uber Open Source, Facebook Open Source).

Likewise, a different delivery method may be used to unlock a new market:

• In the case of Open Core, hosting the offering may improve ease of use and make it easier to offer usage based billing and sell to smaller buyers (eg. Elastic Cloud, Gitlab SaaS).
• In the case of SaaS, selling an on-premises version (which may mean in the customer’s cloud account) may make it easier to sell to the enterprise, public sector or other security minded customers (eg. Github Enterprise, Mulesoft Anypoint Private Cloud).

So we can see a simple distinction between Open Core and proprietary SaaS is difficult to make.

Joseph Jacks offers a straightforward way of defining commercial open source companies as those that would not exist if their underlying open source core didn’t exist. (By the way, I recommend reading the OSS Capital blog and especially this post if you are looking for a broader (and more nuanced) market perspective on these topics.)

Since I am focused on the operator’s perspective, I’d like to keep this comparison limited to the most common initial offerings - proprietary SaaS and Open Core that runs on-premises (not through SaaS delivery). An additional point of clarification: running on-premises may just mean the customer’s cloud account and not actually on their own servers.

#saas #open core