Most self-funded shrinkwrap software companies I know begin their lives making development tools.  This is not surprising: they are following that old saw that says write what you know.  And, of course, developers know about development tools.  If it were up to software developers, there wouldn't even be a class of people called dumb users because the only users would be software developers using one development tool to create more development tools.

And, if you recursively reduce this economic equation to its simplest form¹, you prove that 1 equals 0 and the universe explodes.  Boom.

Still, for many self-funded software companies, building software development tools makes some amount of sense.

By definition, the development tool market is very familiar to software engineers.  While all the uses of the software cannot be predicted, companies started by software engineers have an intuitive understanding of this market by virtue of being a part of the same industry.  Sure, a software engineer might be able to learn enough about, say, baby baboons to write decent software for zookeepers but he'll probably never be interested enough in zookeeping to be as familiar with it as software development tools.  Write what you know -- software development is what software engineers know about.

Furthermore, software engineers as customers know the value of software and are willing to pay for software.  Others might not.  For example, I knew a guy who wrote a sophisticated suite of software for managing lifeguard stations along the California coast.  But he had a terrible time convincing lifeguards of its value, even though it was obvious to him that computer software was much better than pencil and paper.  At least with software developers, you can assume that they understand the value of software as a concept.

On the minus side, software development tools are a crowded market and not a particular large market, at that.  But, considering how many self-funded companies make at least a semi-successful go at it, it is a reasonably good starting place for a new company.

If you are interested in starting a new company yourself but are still casting around for product ideas, an analysis of which other companies are self-funded, what they produce and, in some cases, what happened to them might help.

A quick analysis shows that it is easy to break down the development tools market into five groups.  They are:

1. Component companies
2. Common development tools companies
3. Database manipulation tools companies
4. Uncommon development tools companies
5. Copy protection companies

Component companies create components for sale to application developers.  Examples of self-funded component companies include Dundas Software, BCGSoft and Stingray Software (before being bought for $21M by Rogue Wave in 1998).  Major sub-categories are: grid components, chart components, fancy GUI components, image file manipulation components and complex mathematics components.

Common development tools companies create tools that are less popular versions of popular mainstream development tools.  Examples of self-funded common development tools companies include Perforce, SourceGear and Fog Creek Software.  Major sub-categories are: code editors, source code control systems and bug tracking systems.   Professional shops always use these products and small common development tools companies cater to developers who prefer (for whatever reason) to buy them from a small company than to buy from a big company or use open source.

Database manipulation tools companies create tools to query and organize databases.  Examples of self-funded database manipulation tools companies include Mentat Technologies.

Uncommon development tools companies create exotic development tools that are either not popularized yet or are niche products.  The two major sub-categories are: automating programs and exotic niche programs.  Examples of self-funded uncommon development tools companies include SandStone Technology and MicroGOLD. Sophisticated build management applications, UML drawing applications, testing tools and documentation generators are automating programs; they automate what is done by hand or with simple command line tools on most teams.  Exotic niche programs include weird real-time operating systems and databases built around exotic themes (such as small footprint databases, XML databases and object-oriented databases).

Copy protection companies create software and/or hardware schemes for sale to other software developers to prevent software piracy (theft).

Another important company defies categorization although many use its product for software development.  VMWare (before being bought for $625M by EMC in 2003) is an example of a self-funded software company.

Please e-mail me any self-funded companies that I have failed to mention here so I can update this article.

¹ The existence of a single development tool company pre-supposes the existence of an infinite number of development tool companies by inevitably taking the divergent customer series out to infinite generations. By prima facie reasoning, however, infinite companies cannot exist in a finite universe, proving that development tool companies do not exist. This contradicts our a priori assumption, that is, that development tool companies exist, leading us to conclude that one development tool company equals zero development tool companies, or 1 equals 0.