Paper:FOSS Business Models

From Wiki

The topic of business models and sustainability was an area of keen interest for many of the participants at Good to Great FOSS. Projects already distributing code under FOSS licenses were looking for insight on ways in which to make their projects more self-sufficient. Projects that had not yet committed to FOSS licensing were looking for justification and motivation to consider FOSS distribution models.

A range of business models were considered in the discussion:

• Support-based revenue model: Firms like Red Hat distribute their code as FOSS, but then enter into support contracts with customers. In this model, the market differentiator is quality and ubiquity of the support. Customizations and integrations with other systems can also generate additional revenue.

• Consulting models: In this model, business practices are built around FOSS applications which require customization and substantial experience in order to deploy. Consultants generate revenue based on the FOSS code by selling installation and development services to clients, as well as support and maintenance. Perhaps the most pervasive example of this approach are the FOSS content management systems (CMS) used in web site publishing. Applications like Drupal, Joomla! and Plone have rich ecologies of consultants who install, customize and support individual instances of the software. A key difference in this type of model is that the consultants aren't necessarily involved in developing and maintaining the core application code.

• Data-centric model: In this model the code is FOSS, but the application depends on non-open or difficult-to-acquire data sets. An example of this from the US software market is the DemocracyInAction platform, which is licensed under the Affero GPL. The platform is used for online advocacy targeted at media and congressional representatives, and the databases of legislative and media contacts are not distributed with the platform.

• Hosting services: Some businesses are built around value-added hosting using FOSS platforms: the value proposition is in the quality of hosting, excellent up-time, support, maintenance. In these scenarios, customers know they can migrate at any time, but in the meantime can delegate cumbersome responsibilities that require full-time attention (such as machine maintenance and security patching) to vendors with dedicated facilities. An example of this business model is www.opensourcehost.com, which provides Linux-based hosting of web applications for content management, blogging, site management and other services.

• Hardware-based business models: While cast in a different context than the projects at Good to Great FOSS, business models based on selling hardware running FOSS software are becoming more common. Perhaps the best known examples in this regard are the wireless network routers based on FOSS. These routers run stripped-down versions of the Linux operating system customized for networking. Linux-based cellphones are also making inroads in markets such as Japan. In these models, vendors make money on hardware sales and/or subscriptions where open source is just an ingredient in the product offering.

• Dual licensing models: As described in the FOSS Licensing section of this document, projects such as MySQL distribute their code under multiple licenses. The product is distributed as free GPL software to other GPL projects, but licensed with associated revenue to vendors who want to distribute the code in proprietary products.

A general theme in the business models dialog centered on "generic" versus unique software functionality: FOSS business models may tend to work better with applications where unique intellectual property and/or unique business strategy and logic are not embedded in the application. This is because it is clearly difficult to maintain a market advantage when the differentiating product attributes are available for any competing party to study and emulate. This is borne out in product sectors such as web content mangement systems; platforms such as Drupal, Joomla! and Plone offer very similar features, and compete on core architectural traits including module capabilities, underlying technologies, and usability. On the other hand, a project like the Asterisk VOIP (Voice over Internet Protocols) platform serves as a counter-example; Asterisk possesses a number of unique features and capabilities embedded in the code, but maintains its market edge by consistently out-innovating the competition.

The question was raised as to how well each of the discussed business models might work in the African context. The consensus was that making money in Africa is hard in general. Since the majority of IT expenditures in developing countries are made by government and large telecoms, the implication was that each model should be vetted for viability in those markets. But it was also pointed out that trying to have government understand the difference between FOSS and proprietary offerings is quite difficult.

Several challenges in successful FOSS business models for Africa were enumerated:

• Support for FOSS offerings is still costly in Africa; the tipping point where demand drives down prices has not yet come. Microsoft, for example, is lower quality, but it's more cost effective.

• There is a pervasive issue of “certific-itis”; the Africa markets tend to prioritize vendor certification over unique and strategic technical knowledge and tools. Microsoft has a large practice in certifying Microsoft expertise, and FOSS products like Linux lag in this regard on the continent.

• Available cash to bootstrap FOSS ventures is also less available to private enterprise; NGO's can use grants and other sources in ways that companies might not.

As exercise, participants in the discussion decided to build out a scenario using the DrumNet project and talk through business model and sustainability. While session time constraints did not allow the scenario discussion to fully play out, many interesting ideas were exchanged in the dialog.

The scenario was laid out as follows:

• DrumNet wants to connect 4 actors: buyer, farmer/producer, banks, and input suppliers (fertilizer, etc). They want to have a seamless interface between these suppliers in a Market / Finance / Information (MFI) model.

• They are developing a platform to use phones and the internet to connect these actors. In theory this is a win-win: banks loan and make money, farmers get support and sales, buyers get better organized local markets, suppliers sell product.

• The particular use case for the discussion was as follows: Farmers join the network, indicating the number of sunflower acres they intend to plant. SMS and email are sent to suppliers, telling them what quantity of supplies to ship to each farmer Suppliers send goods to a stockist, and farmers pick up goods. Farmers then report on crop progress and yields, which are then sold through the network at harvest.

The overall opportunity is that agricultural markets are poorly served, and system like above can introduce efficiencies and transparency into the market place. The struggle is in defining the business model: is it transactional, is it ASP (Application Service Provider), or is it training and support?

A matrix was then developed to consider how to apply various business models for the scenario based on a FOSS platform:

Scenario	Drumnet Value Add	Revenue Sources	Risks	Advantages
Single Instance	Market and Actor knowledge and relationships, Trust Factor, System knowledge	Transaction & interest fees, Certifications, Training & doc, Access fees/sponsorships	Cloned by competition	Goodwill, Trust, leadership by example, 3rd party contributions
Franchise (contractual)	Markets knowledge, Strategic consulting, Organization development	Consulting, % of volume, Support, Customization	Contract non-compliance via hacked code or fake data
ASP model	content/data integration, scale in markets (a/k/a critical mass)	Hosting fees, meta-consulting, partnering, advertising
Data-centric	Unique data, “Secret sauce”	Selling data, consulting on data	Reverse engineering

All participants agreed it would be interesting to see where DrumNet actually takes their business model, and what licensing they decide upon to support their approach.