"Man is a microcosm, or a little world, because he is an extract from all the stars and planets of the whole firmament, from the earth and the elements; and so he is their quintessence."
Philipus Aureolus Paracelsus
Open source has seen great success in general information processing, but does it have a future in vertical markets? In this article, we examine how geospatial open source provides an example of the market challenges of a mid-sized vertical market.
Open Source and Decision Makers
General purpose open source software (OSS) has been on the radar of decision makers for almost a decade. Big projects like Linux, Apache, Firefox and Open Office are supported by Fortune 500 companies like IBM and Sun. Everyone knows about open source, it is in their plans, books have been written. In general information technology (IT), there is little more to say about open source.
However, the IT world does not end at databases, operating systems, and office automation. IT is like a fractal form, each major facet can be subdivided and re-subdivided down into particular shapes that fit the needs of unique markets.
The economy is full of niche markets with very particular information processing requirements. Examples of niche markets can be found in health care, education, natural resources, manufacturing, and telephony. Each of these fields makes use of the generic open source building blocks that have already swept through IT. They also have their own distinct ecosystems of dominant proprietary vendors and de facto standards that shape decision making and software acquisition. Open source is growing in these niche markets, but much more slowly than in general purpose IT. The reasons are pretty straightforward: smaller markets mean fewer users, fewer developers, and fewer resources for open source.
History of Geospatial Markets
Geospatial is one of the niche fields that is being slowly colonized by open source. Geospatial software is used by natural resource managers, cartographers, fleet managers, and anyone with a location component in their data.
Geospatial is a recent term for what has traditionally been known as "geographic information systems", or GIS. The original users of GIS software were government environmental scientists and land managers. As early as the mid-1960s, governments were building their own GIS systems, writing the code in-house.
In the 1980s, as computing hardware became cheaper and more interoperable, the economics of GIS shifted. Rather than buying computers and writing GIS systems from scratch, governments bought computers and then bought GIS software. An ecosystem of proprietary GIS vendors emerged quickly. Some vendors were regionally based, others were specialized in particular fields like forestry or oil and gas.
Some of the last in-house systems were written by the government. The Geographic Resources Analysis Support System (GRASS) was written by the US Army Corps of Engineers after an evaluation process that determined none of the current proprietary systems met their needs. The Map Overlay and Statistical System (MOSS) was written by US Fish and Wildlife Service, after a similar market evaluation.
Both GRASS and MOSS were released as public domain works, effectively becoming the earliest examples of open source geospatial software. GRASS was re-licensed as a GPL project in the mid-1990s. However, throughout the 1990s, the OSS largely languished in academia while proprietary software filled the entire government and corporate ecosystem for workstation-based GIS.
Just as the computer and operating system market consolidated to Intel and Microsoft through the 1980s, the GIS market consolidated through the 1990s. The GIS market consolidation battle in North America and, increasingly, the rest of the world, was won by a proprietary vendor, ESRI, which started out from a market base in the US federal government and gradually displaced other competitors in the North American market. By 2000, ESRI had achieved a geospatial market position similar to that of Oracle or Microsoft in the general IT marketplace. They were the default vendor and the only safe choice for decision makers.
Like Microsoft and Oracle, ESRI's market dominance was built on a narrow but important product category. Microsoft dominated operating systems and office automation. Oracle dominated relational databases. ESRI's dominant category was, and still is, desktop GIS software. Desktop GIS software provides users with the capability to create, edit, analyze, and cartographically print geospatial data.
The rise of the Internet was famously disruptive to the Microsoft business model. In 1995, Bill Gates radically revised the company's software strategy to focus on networked communication. The new strategy was intended to pre-empt new competitors arising who could take advantage of the Internet software arena. On the desktop side, Microsoft was successful, Netscape was crushed, and RealNetworks is mostly gone.
In geospatial, ESRI suffered a similar disruption from the Internet. As their users got used to accessing non-spatial data over the Internet, they began to ask an obvious question: "how can we provide access to our geospatial data over the Internet?".
The immediate question didn't involve analysis, cartographic printing, or data capture, all of which are ESRI's core desktop strengths. The immediate request was for simple data access. Like Oracle, ESRI made some early strategic errors in providing Internet-enabled versions of their software. Their pricing model locked smaller organizations out as the cost of entry was too high. However, rather than throwing up their hands and not providing Internet services, smaller organizations and individuals simply looked for alternatives. In the case of Oracle, they found MySQL and PostgreSQL. In the case of ESRI, they found MapServer.
Arguably, the price points that individuals and small organizations wanted couldn't be rationally provided by ESRI or Oracle. But, by driving individuals and smaller organizations away, the dominant vendors seeded a new marketplace that quickly developed alternative open source product suites.
Dynamics of Open Source Geospatial
In 2000, the market for "web map servers" was brand new. The high cost and poor performance of the offerings from ESRI slowed adoption. The low technical bar to entry needed to "display maps on the Internet" allowed a new entrant to gain traction. MapServer was originally developed in an academic setting, the University of Minnesota, using grant money. Released as open source in 2000, it had just enough existing functionality to begin attracting new developers.
Once MapServer began providing useful services to users, it started to attract more open source development. Unlike the "pay to play" model of proprietary products, the open source MapServer allowed organizations to get their feet wet with existing capabilities for free, then pay to add new capabilities if and when they needed them. Organizations that adopted MapServer for free began using funds to add improvements around the margins.
Among the features that were added by the MapServer user community were the ability to read data from additional GIS file formats and from satellite imagery, and support for international interoperability standards. The improvements, and particularly the standards support, served to make MapServer useful to a larger audience, which drove market growth even more.
The characteristics of the adopters of open source geospatial are familiar to any student of open source market dynamics. Open source tools are generally evangelized by technical staff, who have the ability to acquire and test the tools themselves. This effectively limits early adoption to startups and larger organizations with pockets of progressive technical expertise. Conservative organizations tend to gather intelligence through vendors and trade magazines, which serve mainly to reinforce existing purchasing patterns. Organizations that have already chosen a proprietary product for a functional category will rarely switch to an open source equivalent until the proprietary product hits end-of-life.
The exceptions to this rule have been open source phenomenons, like Linux or Firefox, which have received publicity outside the technical trade press. The publicity around Linux has forced even conservative organizations to make some formal consideration of using general open source.
Open source geospatial is a small niche and its software will never receive the popular press coverage of the sort that has made Linux and Firefox well-known. Most existing geospatial software customers, such as governments and resource industries, remain comfortably within the arms of the dominant proprietary vendor, ESRI.
The market area where open source geospatial has been most quickly taken up is among organizations in which there is no existing bias towards the dominant proprietary vendor. These are usually companies or agencies making their first foray into Internet mapping, motivated by the geospatial renaissance triggered by the introduction of Google Earth and Google Maps. Their technical staff research the options and find that they can get what they want from open source. The available open source includes:
- the PostGIS geospatial database
- GeoServer to provide web map services and web feature services
- the GDAL programming library
- the Geotools Java GIS toolkit
- the SharpMap geospatial application framework
- the web interfaces provided by OpenLayers
Providers of web-based services, which have to scale their infrastructure based on potentially exponential changes in users, are particularly enthusiastic about open source. The capital cost of scaling an infrastructure that uses proprietary software is $X dollars times the number of nodes. For open source, the capital cost of scaling is zero. The annual infrastructure support costs for proprietary software are again locked linearly to the number of nodes. For open source, the support costs in consulting fees to open source programmers or companies generally increase with the number of nodes, because the odds of needing new features or exposing bugs increase as use increases, but the increase is generally less than linearly.
Examples of new Internet companies using open source geospatial include:
- Redfin, a real-estate information company
- GeoCommons, a data sharing community
- Zonar, a fleet management and vehicle tracking company
- GlobeXplorer, a satellite imagery re-seller
- Urban Spoon, a restaurant review site
- Whereyougonnabe, a spatial add-on for Facebook
The other area where open source geospatial has been taken up is in technically savvy pockets of large organizations. Like the Internet companies, these organizations have enough of a user-side demand that deploying proprietary Internet service software creates a large financial burden. Unlike startup companies, large organizations can potentially afford to pay for the proprietary software.
In some large, conservative organizations, visionary managers adopt and deploy open source. However, there is nothing systematic about the adoption, it is a consequence of the personality and personal interests of the manager. If he or she leaves the company, the open source pocket may disappear. Because the progress of open source geospatial in these organizations is so personality-based, it tends to be rare and to run in fits and starts.
Open Source Geospatial Challenges
The progress of open source geospatial on the desktop has been very slow. Desktop geospatial software already has an entrenched proprietary incumbent, ESRI's ArcGIS, with a long history. The amount of quality code required to reach feature-parity with the incumbent is very high as ESRI has been working on their desktop software for decades. Simple desktop implementations are available with QGIS, uDig, and gvSIG, but are mostly consigned to the niche of extremely low budget organizations. As a result, financial resources are not available to speed up development, and the pace of progress remains slow. The exception has been gvSIG, which is heavily funded by Spanish government organizations, but it is still mostly a niche development used in Spain.
In all cases, the growth of open source geospatial has been slowed by matters of scale. Open source products generate much smaller revenue streams from user populations than proprietary products. In large markets with well-capitalized customers, companies can make good money even on the smaller revenue flow of open source.
However, in a small vertical market, it is difficult for companies to get a foothold. A customer will usually deploy several open source geospatial products to create a solution, so a support provider has to have extensive in-house experience to support the whole solution. In a traditional model, start-up costs would be capitalized by a venture funder, but the size of the geospatial market-place makes the 10:1 returns required by venture capital unlikely.
OpenGeo is breaking the geospatial support log-jam by building a social enterprise using philanthropic funding to bootstrap an organization that contains the breadth of expertise necessary to support a variety of open source geospatial applications. OpenGeo's motivation is not to maximize profit, but to maximize social good, while covering costs. This allows the organization to build a sustainable market while surviving on the smaller revenue streams available in the open source geospatial arena. The products OpenGeo supports such as Geoserver, OpenLayers, PostGIS, and GeoExt, make a top-to-bottom deployment stack for geospatial applications.
Lessons for Other Markets
Open source geospatial holds a number of lessons for other vertical markets. First, frontal assaults on the leading proprietary vendor are unlikely to succeed. In their core areas, the leading vendor has an advantage in technology development and existing mind-share. Usually, building enough technology to compete with a leading vendor head-to-head takes years of development, and a partially functional product will be ignored.
Second, disruptive changes in technology provide opportunities for open source. Most leading vendors carved out their advantage on the desktop during the 1980s and 1990s. The transition to web-based services has opened a temporary gap in the marketplace where existing vendors have a smaller technology advantage, and their marketing advantage is limited to their existing universe of customers. Open source can become the core for new service-based companies competing with proprietary software vendors.
Finally, new markets for capabilities are the most fertile opportunity of all. In geospatial, its expansion into daily life, through vehicle and device tracking, low cost aerial imaging, and handheld mapping, is growing the market exponentially. New developers and managers, without long-held preconceptions, are making technology choices. On a level playing field, open source Internet technologies are regularly winning.