Three options for the future of Open-Core licensing

Although I hesitate to write another post about business models, Open-Core and open source, it was clear from the recent Open World Forum and Open Source Think Tank in Paris that the issue of how to make money from open source software is still of considerable concern to vendors.

Open Source Think Tank attendees were split into teams and asked to come up with a list of the three biggest issues facing open source; pick the biggest priority; and suggest a potential solution. Looking around the room, several groups had business models somewhere in their top three*.

It was also clear during the session that there was considerable concern amongst those involved about the dilution of the open source brand, with a couple of groups naming it as the biggest issue (their solutions: a coordinated campaign of education; and an OSI-led effort to define approved development and business strategies – which is precisely what Simon Phipps is advocating for).

Concern over the dilution of the open source brand and confusion over the best way to make money from open source are two issues with a common thread: the Open-Core licensing model (an open source project combined with proprietary extensions for paying customers).

Those that wish to protect the open source brand see Open-Core as a dilution of the meaning of open source that is spreading confusion; while those that wish to build sustainable businesses around open source are aware that many vendors and VCs see Open-Core as the solution to their problem. Can the two be reconciled?

During his speech to the Open World Forum, Michael Tiemann commented that the move to free eradicates the traditional concept of product, and shifts the value that a company can provide to customers to services.

In this context it is possible to see that pure open source software approaches to making money from open source are service-led, while the Open-Core approach remains product-led.

Or, as Mark Shuttleworth put it at today’s BT Accelerating Enterprise adoption of Open Source Software event in London, “the move to services is not being led by open source companies themselves – they have modeled themselves on the companies they are trying to dislodge.”

This is perhaps understandable since, as Mark also explained, it takes time for user companies to create strategies around consuming technology as a service, rather than as a product.

Either way, if Michael Tiemann is correct, the Open-Core model is likely to be short-lived, since it mistakenly continues to focus on product, rather than service as the main way to deliver value.

I would agree with my CAOS colleague Jay Lyman’s assessment that we may be starting to see a return to support and other services, rather than commercial code and licensing, as the preferred mode to directly monetize open source.

However, we are also seeing the increased use of open source within the development of proprietary products by companies like IBM, Day Software and Oracle. The theory that free necessitates a services approach assumes that the vendor is trying to directly monetize open source software. Clearly this is not always the case, and there is still a market for products – proprietary products – that benefit from components developed by collaborative open source communities.

I have also previously questioned whether Open-Core is sustainable in the long-term, although for me the issue is that if it is used with GPL software and vendor-dominated development models (which it typically is, under the vendor-controlled Open-Core model as defined by Andrew Lampitt) then the vendor does not enjoy the benefits of a true collaborative open source development community.

Or, as John Mark Walker recently put it, by failing to enable the community to flourish, vendors are “putting a cap on the possible success and losing the innovation opportunity that comes with a wildly successful community.”

If that is the case, then what is the future of the vendor-controlled Open-Core model? Before considering the options, I want to make four quick statements:

a) Regardless of whether you consider it to be true to “open source” or not, the Open-Core model clearly does work for some companies that are succeeding in generating revenue and fostering community. Balancing the two is hard, but it can be done, and there are examples of companies that are making it work.

b) By “services” we are not simply referring to support services, training, implementation services and consulting. The term also includes software delivered as a service, and features delivered as a managed service – such as via Red Hat Network. It includes anything you can think of, in fact, that can be considered a service, rather than a product.

c) Vendors using the GPL to prevent potential rivals building proprietary products from their code are also often (though not always) locking out potential contributors and missing out on the benefits of a collaborative development community.

d) I’m assuming that any company following the Open-Core model has already worked out which features are appropriate to community users and which are appropriate to paying customers, and has a clear idea of the difference between the two. If not, the community will likely build open source projects that rival the proprietary features. Sorting that out is the priority.

That said, what are the options for vendor-controlled Open-Core vendors faced with trying to build a sustainable business while growing and serving a community, and (potentially) staying true to the definitions of the Open Source Initiative?

To my mind there are three options:

Option 1: Open-Core but not necessarily OSI™-approved
Companies that are successfully executing an Open-Core model by growing revenue and building community have no need to change what they are doing. However, they should be prepared that any attempt to refer to themselves as an “open source vendor” is going to elicit disapproval, or be prepared to fight the plans to use the OSI to define open source development models and business strategies.

Option 2: The open source core should be released under a more permissive license, or better still via an existing community/foundation in order to benefit from and encourage a collaborative development community. This requires a shift to the community-controlled Open-Core model, as defined by Andrew Lampitt. There are two potential options for monetizing this approach:

    Option 2a: Open-Core, paid services
    Features currently delivered via proprietary licensing are delivered as a managed service to paying customers. This might sound like a giant step, and it does require something of a leap of faith from both the executives and investors, but I spoke to the CEO of one Open-Core company during OWF which is actively considering shifting its strategy in this way. Remember, you’re not competing on product anymore, you’re competing on service. If you’re good enough you ought to be able to compete with proprietary rivals even if they are using what was your code to build rival products.
    Option 2b: Open code, proprietary products
    Stop kidding yourself that you are an open source vendor and focus your attention on developing the proprietary extensions as a complete product that complements and builds on the open source code. If the proprietary extensions are valuable enough you ought to be able to compete with proprietary rivals even if they are using what was your code to build rival products.

For more discussion on the benefits and challenges of Vendor-controlled and Community-controlled Open-Core, see Andew Lampitt’s post. For more discussion on Option 2b, see this post on the patron model of open source commericalisation

*The group I was in certainly rated business models as a key issue as the vendors present wondered how it was possible to build a sustainable business around open source that could grow at a rate that could support a public offering without also involving VCs that would likely sell the company to a proprietary rival if the offer was right. (An impossible dream? We certainly couldn’t think of a solution, although we did manage to “solve” the issues of license confusion and the lack of a single lobbying effort to represent all open source constituents, so the time wasn’t completely wasted.)

Tags: , , , , , , , , , , , , , , , , , ,


#1 Twitter Trackbacks for 451 CAOS Theory » Three options for the future of Open-Core licensing [] on on 10.05.09 at 6:07 am

[…] 451 CAOS Theory » Three options for the future of Open-Core licensing – view page – cached An open source blog by The 451 Group. — From the page […]

#2 Three options for the future of Open-Core licensing IM Consultant on 10.05.09 at 6:49 am

[…] Three options for the future of Open-Core licensing By admin | category: consulting companies | tags: Action, biggest, biggest-issues, […]

#3 451 CAOS Theory » Three options for the future of Open-Core licensing | Open Hacking on 10.05.09 at 6:52 am

[…] original here:  451 CAOS Theory » Three options for the future of Open-Core licensing This entry was posted on Monday, October 5th, 2009 at 5:56 am and is filed under Linux, News, […]

#4 Michael Tiemann on 10.05.09 at 4:43 pm


I think you’ve done a great job of delineating the issues, and of course only time will tell which position will be best for which companies. I’d just like to add one thing which became clear to me as I discussed this topic with other delegates to the Open World Forum: it may be that a successful service-based open source company depends on an aspect of scale, diversity, and complexity. That is to say it may be quite difficult for a thousand such companies to ever make any money each supporting one single package, but it is possible for a single company to make quite a bit of money supporting a thousand packages.

So many entrepreneurs have it drilled into their heads that they need to focus on some narrow slice of the market, demonstrate dominance there, and only then expand. These entrepreneurs who have staked out narrow territories, be it content management, CRM, database, etc., are finding that in the real world such focus is contraindicated.

I think that the radical success of open source companies can better found by understanding how to deliver the appearance of simplicity to complex problems via services rather than trying to build and sell piece parts that may become more or less valuable with proprietary widgets around them.

#5 Matthew Aslett on 10.06.09 at 3:40 am

Thanks Michael,

One of the topics that cropped up at the Open Source Think Tank and also at an event in London yesterday was the idea that the concept of services needs to evolve to maximize the potential returns for open source service providers (both supporting single and multiple packages). Perhaps this focus on the product has distracted attention from the creation of more innovative approaches to services.

#6 Lessons Learned: A Services-Based Open Source Business Model | OpenLogic Blogs on 10.06.09 at 11:50 am

[…] source, showed that they are not big fans of the open-core business model.  Recent blogs by Matthew Aslett, Jay Lyman, John Mark Walker center around this issue of the future of the open core business […]

#7 451 CAOS Theory » The rejuvenation of community-controlled open source on 10.15.09 at 5:33 am

[…] open source Matthew Aslett, October 15, 2009 @ 5:26 am ET Jay recently speculated, and I agreed, that we may be starting to see a return to support and other services, rather than commercial code […]

#8 451 CAOS Theory » Out of control on 10.16.09 at 8:47 am

[…] of control Matthew Aslett, October 16, 2009 @ 8:47 am ET I recently suggested that open core vendors should consider releasing the source code for their core open source project […]

#9 Only Compare rival white mountain ice cream maker Products and Resources | Ice Cream Maker Deals and Discounts on 11.09.09 at 5:46 am

[…] Three options for the future of Open-Core licensing […]

#10 451 CAOS Theory » How Day Software stumbled upon an open source business strategy on 11.09.09 at 10:17 am

[…] source business strategy Matthew Aslett, November 9, 2009 @ 10:15 am ET I’ve written a few times recently about the fact that I think the open source engagement model practiced by companies such […]

#11 451 CAOS Theory » What SpringSource dm Server licensing tells us about control versus community on 01.20.10 at 7:21 pm

[…] also: Three options for the future of Open-Core licensing The rejuvenation of community-controlled open source Out of control Losing control Permalink | […]

#12 451 CAOS Theory » Alfresco’s new Activiti en route to Apache on 05.17.10 at 6:58 am

[…] is something that I previously argued open core vendors should be considering last year. Of course it is not a simple matter. […]

#13 451 CAOS Theory » Winning and losing with open core on 01.04.11 at 7:37 am

[…] vendors thanks to their installed base, larger resources, and product maturity. I have previously argued that open core vendors actually need to become more open to encourage wider adoption and […]