Free. A word that grabs attention. In an age of austerity, free software has to be especially attractive, doesn’t it?
The attraction is easy to see. Licence fees are a highly visible cost for many software projects. Ongoing support costs, based on the initial licence fee, add to the pain. So as Free and Open Source software (FOSS) gains credibility in areas such as content management and customer relationship management, it’s not surprising that people are looking to eliminate those licence fees in order to reduce costs.
Sadly, licence fees are only the tip of the iceberg. Projects also incur costs to select the right product, install and configure it, train people to use it, manage the organisational change necessary to deliver benefits with it. These can outweigh licence costs by a factor of five or more. If FOSS inflates these costs, it may not be so cheap after all.
So where do the costs of open source lie? Consider these factors:
Product Selection. As more open source options gain credibility, the cost to choose the right product grows. For example, you might want to consider half a dozen open source products for a web content management project. And you’ll need to do the analysis yourself – open source systems rarely have pre-sales engineers to provide support. (Licence fees fund those pre-sales engineers.)
Proofs of Concept. To understand an open source product, you tend to rely on pilots and proofs of concept rather than documentation. This is probably good practice for any product, proprietary or open source, but open source requires a greater emphasis on hands-on analysis, with its associated costs.
Configuration and integration. People who build open source software have to pay their mortgages. If they can’t make money from licence fees, then they probably make it from configuration and integration services. And this is often the cost that’s hardest to pin down: licence fees are fixed, but integration has a habit of expanding as the project proceeds.
Customisation. Access to source code creates a hidden risk: developers are tempted to spend time tweaking it. This increases costs for the initial project. More importantly, it makes it more difficult to apply upgrades in future.
Support. Some open source products have excellent community support: responsive, knowledgeable and prompt. But many organisations want the certainty that comes with paid-for service level agreements. This is another place where open source vendors can make their money.
Examine these factors and you may find that open source has some substantial costs attached. It may still be cheaper than proprietary software. Or it may not. You need to do the analysis for your context.
This analysis may reveal something else. Open source implementations tend to spread their expenditure over time in a very different way to proprietary ones.
Proprietary software carries that upfront licence fee. Open source involves a more gradual spend. You download the software and run a pilot. If that succeeds, you ramp up. If it doesn’t, you back out. Commitment comes in small chunks. How many organisations have paid large licence fees for proprietary software they no longer use?
And initial commitment creates a psychological risk. You’re making a big upfront investment, so you search for more benefits. Each of these requires small extensions to the system. Scope creeps. You end up with a much larger project than you started out with. Yet there’s one thing we know for certain about software projects: big projects are far more likely to fail. If it helps you avoid this trap, open source will save you a lot of money.
Open source can provide other benefits too. For example:
It sits naturally with open standards. Vendors can’t lock you in with proprietary data formats and APIs. This reduces integration costs. It also reduces end-of-lifecycle costs to move from one system to another.
Access to source code makes it easier to configure the software. Sometimes you need to see just how something works in order to work with it.
There are fewer licensing problems when deploying to the “cloud”. Deploying multiple virtual servers is rarely a problem with open source. It can be a headache for proprietary software that is licensed on a per-server basis. This is a growing issue as cloud infrastructure goes mainstream.
But for my money, open source is a risk avoidance strategy. Choose it to save licence costs and you may be making a false economy. Choose it to reduce the risk of project bloat, and there’s every chance you’ll achieve significant cost savings.