The Navy PWC Yokosuka began its GIS implementation journey in 1995. Since that time, millions of dollars have been invested by PWC and its customers in equipment, training, data acquisition, and application development. Yet, demand for the technology continues to increase and has spread throughout the U.S. Navy in Japan. This presentation will cover measurable and intangible benefits of GIS technology, cost factors and pitfalls, organizational issues, customers, products and long-term sustenance considerations; all based on the real world experiences of PWC Yokosuka in a facilities management environment.
Implementing GIS technology is not a “one size fits all” proposition. Therefore, the information in this paper dealing with potential costs, pitfalls and benefits must be understood in the context of our particular operating environment.
The mission of PWC Yokosuka (hereafter referred to as “PWC”) is to provide many types of services including maintenance & repair, utilities, transportation, engineering and planning to the military operating forces (primarily Navy), shore establishment and other Federal agencies in our area of responsibility (AOR). Our customers expect us to provide these services in a responsive, high quality and cost-effective manner. As a part of our continuous effort to improve, PWC initiated a program in 1995 to begin implementing GIS technology throughout the Command.
The Japanese islands of Honshu and Kyushu are home to numerous United States military installations belonging to all the DOD services. The various Naval installations are clustered into Naval Complexes (NCs). As a result of our GIS successes during the first couple of years, PWC Yokosuka was asked by other NCs to implement the technology in like manner. Thus, in 1998, PWC began its current role as a GIS service provider to the Japan Navy Region. The region’s GIS implementation now consists of 21 sites organized into four NCs as shown below:
The objective behind the GIS technology is integrated installation management. This involves enabling many diverse functions to work together, sharing data and maps.
The full-time regional GIS support staff resides in Yokosuka and consists of an American Civil Service (USCS) manager and two (soon to be four) Japanese national employees. The remote NCs designate a primary GIS point of contact (POC) with whom the regional staff can coordinate their implementation. All NCs also designate departmental POCs throughout their respective organizations. GIS and departmental POCs perform their functions as collateral duties. The regional support staff coordinates all data development, training, etc. for the NCs.
In terms of equipment, the Yokosuka NC hosts a master GIS data server, an Oracle/ArcSDE server and an ArcIMS server. The remote NCs each manage a local GIS data server for direct query and editing of their data. This remote data is periodically reconciled back to the master data server in Yokosuka. The ArcIMS server serves as the regional map server for all NCs and provides basic functionality to all users in the region.
Before making a commitment to implement GIS technology, you must answer the following questions to get to “Level 1”:
a) Do we need GIS technology? This is the most basic question. How will the system be used? What problems will it help solve? Are these practical uses? Have you seen others using the technology in this manner? Who are the potential users & viewers in your organization? Who needs it most?
b) Is management committed to this? Without solid upper management commitment for at least two to three years, you risk having the plug pulled on the implementation before it can get going. Management needs to understand that results will take some time.
c) Is adequate funding available? Later in this paper, I will lay out some guidelines for estimating your startup costs. You must have money available to invest.
d) What is the staffing plan? Will GIS support be a full-time or collateral duty? Will your organization re-allocate positions or hire on-site contract support? You should be aware that some sort of full-time commitment is essential. GIS implementations with only collateral staffing move very slowly and risk being crowded out altogether by the normal urgencies that fill people’s workday.
If you are able to answer these questions and the picture looks encouraging, then consider the following questions to get to “Level 2”:
a) Do we have a suitable map? If one is not available, it will take longer to get started (and cost more money). However, almost every site is lacking in some respect. There are six criteria I use in answering this question:
- Up-to-date. The map should be reasonably accurate or else a lot of manual effort will be required up front to make the updates. This could take just as long as creating a new map.
- Spatially accurate. Maps developed from scanned drawings can exhibit spatial distortion due to the warping introduced during the scanning process or poor digitizing practice (re-positioning the source document while digitizing).
- Geo-referenced. It must be tied to the appropriate coordinate system and projected accordingly.
- Topologically correct. This is usually a problem with CAD maps. Building features should be drawn as closed polygons; utility features should be digitized in the direction of flow and should have precise coincidence between related features.
- Spatially contiguous. The map should be seamless, not broken out into tiles.
- Spatial Data Standards (SDS) compliant. This is particularly important for Department of Defense users. These standards are maintained and published out by the CADD/GIS Center for Facilities, Infrastructure, and the Environment (http://tsc.wes.army.mil).
b) What data formats will be used? This may depend on your current digital editing environment, if any. If your organization has strong CAD expertise, you may wish to keep most of your data in CAD format and convert only certain features to GIS formats (shape, coverage, etc.) as necessary. Otherwise, you may wish to begin right away with ArcInfo and the geodatabase format.
c) What software will be used? There are many choices and they depend mostly on the capabilities that users require and how you will maintain your mapping layers.
d) How will viewers (“low-intensity” users) get to the information? Will you set up a web service to make it widely available? Or will you create a smaller portable application using MapObjects? Or you could deploy the ArcExplorer or ArcReader tools, which are free of charge. It is best to simplify the strategy in the beginning; try not to roll out too may different tools at once.
e) Where should we start (“quick kills”)? Based on the GIS capabilities you have seen, where can your organization benefit the most right away? In our case, we started with an adequate map, linked property record information, floor plans and photographs. This is our most basic level of functionality and it brings tremendous benefit immediately.
f) How will we keep the data current? Normally, the best approach is to push data maintenance out to the end users that are the natural data owners. This way, your GIS staff can focus more on the technology, new functionality and data improvements rather than being consumed with trying to keep data up-to-date. However, keep in mind that end users will require training or they will likely become frustrated.
g) How much emphasis should be placed on custom coding? There is a delicate balance between using straight commercial software (out-of-the-box) capabilities and writing your own code. Unless a function is immediately required, it may be best to defer it and see if the software manufacturer is willing to add the functionality to the commercial product. Also, any coding should conform to sensible development practices. This will be discussed more under “pitfalls”.
h) Can we partner with other sites? If you know of other sites with similar requirements, you can learn from them and save yourself a lot of time and trouble during what I call the “figuring out” process. Also, it is often possible to jointly fund initiatives that will benefit two or more sites at once.
If after considering these issues you are ready to seriously consider an implementation, that’s great! Now it is time to start considering the actual costs that will be incurred. The following discussion focuses on the most significant cost factors that will apply in most cases. However, remember that these numbers may vary greatly depending on your particular situation.
a) Staffing. Does it make sense to devote in-house personnel to the implementation? Can someone be transferred to this function or should we hire someone new? I will discuss staffing options in further detail below.
b) Equipment. Is my existing network capable of supporting GIS data traffic? Are our workstations capable of running the software? How should I license my software?
c) Training. Who needs training? Should we bring an instructor on-site or send people to a central training facility? Can we partner with other sites in providing training?
d) Consulting & support services. This will almost always be necessary to some degree, especially if you do not maintain contractor personnel on-site at all times.
e) Data development and cleanup. This is a guaranteed requirement. Most GIS implementations do not consider the poor state of their data in the beginning and end up being surprised at what it takes to organize and clean it. Without good data, a GIS can be rendered practically useless no matter how much money you spend elsewhere.
f) Application development. Like I mentioned before, this involves balancing commercial capabilities and additional functions you may need. Also remember that anytime an application is developed, you are assuming an associated maintenance cost – you will likely be required to use additional funds to maintain and enhance the capabilities over time; or even port it to a new platform.
g) Survey monuments. This is often overlooked, but it is best to consider it up front. In order to maintain your data over time, you will need a sufficient number of monuments for using field surveys to update and add new layers to your GIS database.
There are three basic approaches to staffing. You can use in-house resources for 100% of the requirements, hire contractors for 100%, or settle on some combination both. 100% contract staffing normally implies that one or more contractors will remain on-site at all times. A combination effort normally implies that a contractor will be brought on-site a few times a year for a couple weeks at a time and will be “retained” for various tasks in between the site visits. The relative advantages and disadvantages of each staffing option are listed below:
Less expensive and more flexible.
Familiar with the organization; can push for progress.
Can continue to operate on a limited basis in periods of “lean” funding.
Must establish positions and find people.
Difficult to terminate if it doesn’t work out.
Will likely have limited access to broader resources and technical expertise.
No need to establish permanent positions.
The burden of finding capable people is placed with the contract firm.
Usually has access to broader resources and technical expertise.
More expensive, especially for simple tasks.
Unfamiliar with the organization; limited authority to drive the implementation.
Risk of discontinuity if funding becomes a problem or the assigned person leaves the company.
The in-house staff can drive the implementation and stay actively engaged with the contractor’s efforts.
Smaller tasks can be handled in-house; larger or more complicated ones can be passed to the contractor as necessary.
Lower cost for broad implementations.
Same issues with in-house staffing.
Can be more expensive for smaller implementations.
Must be more proactive in tracking and managing the contractor’s tasks remotely.
Here are some very rough guidelines that you can apply to determine what it might take to implement a long-term GIS solution. Again, please keep in mind that these costs can vary widely depending on the approach taken and the particular operating environment.
a) Staffing (annual costs): estimate $100k/year per position for in-house support. About twice as much for full-time contractor support.
b) Equipment startup (assuming a suitable network is already in place): estimate $10k/person for advanced users, $2k/person for viewers.
c) Equipment maintenance & upgrades (annual costs): estimate $1k/person per year.
d) Training (annual costs): estimate $5k/person for advanced users, $1k/person for viewers.
e) Consultation & support (annual costs): estimate about $200k/year (about ˝ for contract staff full-time on-site).
f) Data development and/or cleanup: To create a new base map from aerial photography, assume $400/acre for installations at least 200 acres in size. For adding legacy layers from existing CAD or hardcopy documentation, estimate about $150/acre.
g) Application development (annual costs): estimate about $100k/year (about ˝ for contract staff full-time on-site).
h) Survey monuments: estimate one monument required for every 10 acres of developed area @ $6k each.
Unfortunately, there are many ways to waste money during the course of a GIS implementation. Here are some common pitfalls to watch out for:
a) Poor planning; unclear objectives. Consider all Level 1 and Level 2 questions before starting. Be sure management understands the plan, the time it will take and what results to expect. Lay out a plan of action and milestones.
b) Lack of organizational commitment. This applies to end users as well as management. The benefits of GIS will not be realized if users are not willing to use the system. Spend time demonstrating the technology and explaining the benefits to them well in advance of installing the software on their computer.
c) Failure to guide contractor efforts. Don’t let the contractor work in a vacuum – work together in setting goals and objectives. Communicate regularly to ensure that assigned tasks are on-track and meet your requirements.
d) Shooting for flash (short term) versus substance (long term). Although demonstrations of advanced GIS capabilities are often helpful in gaining upper management approvals, it may lead to unrealistic short-term expectations. Be sure to focus on the foundational things in the beginning – good data, data maintenance, basic functionality.
e) Failing to regularly evaluate progress and adjust course. It is almost certain that priorities or objectives will change as the implementation progresses. Technical decisions made in the beginning may have proven too expensive or unrealistic. Don’t be afraid to adjust course as long as the destination is still in front of you.
a) Failing to start with a suitable land base. It is rare that an organization has a suitable map in the very beginning. It is usually necessary to develop a new one that meets the six criteria outlined earlier. Because this requirement could delay the implementation, it is tempting to skip this step. Sometimes it may be possible to obtain a land base through means other than aerial photography. Just realize that as time passes it may become apparent that a new map is necessary.
b) Failure to consider the big picture (all potential functions using the system). This comes into play when you are concerned about satisfying many functional requirements with your implementation. A big picture mentality is required when formatting data and writing applications that will potentially be shared by different departments.
c) Spending time & money to add or clean data that is not useful. During a GIS analysis, data is found tucked away all over the place. It is tempting to try to add everything to the system. However, the expense in re-formatting data or adjusting users’ work processes may not be justified by the potential benefit to be gained. Don’t add data simply because it is there.
d) Failure to consider existing work processes and data maintenance considerations. Another important consideration when adding data is to think about it long-term. Will users naturally keep it up to date? For this to happen, you must tie the maintenance to their existing process somehow. Double-entry data maintenance will not help your users – try to avoid it.
a) Duplicating commercial features. The advantage of purchasing commercial software is that the vendor has taken the time to build the functionality. You are paying for this! Most times it is better to leverage what is already there even if it is not the most convenient approach for your users. Provide feedback to the vendors if usability is an issue and urge them to make improvements to their product.
b) Creating applications that look cool but are not of practical use. Again, sometimes it is necessary to do this on a limited scale in order to help upper management see the potential benefits. But these types of applications often end up as “throw-aways”.
c) “Stovepipe” coding (platform or function-specific). This relates to not keeping the “big picture” in mind. For example, why develop a different document retrieval (hotlink) function for different departments? If this is necessary, code it in such a way that it works for any potential user for any feature on any layer.
d) Failure to consider future software compatibility. The ArcView Avenue language has become a classic example here. It does not make a lot of sense to invest much more in these applications unless you plan to be committed to that platform for a long time.
With all this said, what can you realistically expect to gain? I’ve compiled a short list of gains that can be expected in the near-term. While I stopped short of applying dollar values to these, you can think of them in terms of your organization to determine potential savings over time.
a) GIS forces good data management practices, thus instituting a framework for continuous data improvement over time. All organizations have information scattered all over the place in diverse formats. Because GIS can relate this information together and make it more usable by others, it naturally forces a manageable schema and rules for organizing data.
b) Ongoing data collection efforts can be captured for future use. Many one-time data collection efforts are conducted in response to data calls. This is often thrown away when the need has passed. Then a year or so later, it is done again from scratch. If the users know how to capture and store the data in the new GIS schema, it can be easily retained and retrieved by others as necessary.
c) GIS facilitates data upkeep by making it easier, more systematic and simplifying the validation process. Better than most any tool, GIS technology helps point out data deficiencies. There is thus an incentive to correct errors or complete missing information because more people now notice these problems.
a) Data is more accessible, thus it is leveraged among a broader user base.
b) Data repetition and data redundancy will decrease. A good GIS schema partitions the data and places its maintenance with the true owners (those that know it best and have a vested interest in its accuracy).
c) Users can focus on their own data. Users often spend time keeping data that they need but aren’t actually responsible for. Once the owner’s version of the data is accessible, those users can focus on keeping their primary information current.
d) Decisions can be made quicker and more accurately. The natural outcome of this progression is that information is more accurate and immediately accessible to those who require it.
a) The quality of your products and services will greatly improve. What report or diagram is not enhanced by the inclusion of a color map? These types of enhancements all reflect positively on your organization.
b) Users' morale will improve as certain aspects of their jobs are simplified and they work with state-of-the-art tools. Users opposed to these types of tools in their work are diminishing. Most users now welcome them and enjoy the new capabilities. They will find themselves more productive.
c) New types of analysis are available that were previously not possible. Being able to relate information across functional areas, overlaying layers that were previously on separate drawings and performing buffer analysis are a few of many examples.
GIS is an effective technology, but it is definitely expensive. Before starting on the “road to GIS”, it is essential to realize and accept the commitment that is required. In many cases, about two to three years is required to start seeing significant results.
Management often wants to know the return on investment (ROI) in real dollars. However, most benefits are intangible and hard to quantify. What I have found is that “seeing is believing” and typically speaks louder than ROI numbers. This is the dilemma faced by most. If you can get far enough to show basic functionality with your organization’s data, you will be almost over the hump.
You should also realize that this is a great time for implementing GIS technology. Costs are lower than ever, the technology is much more mature and intuitive, standards are better defined and it is highly likely that you can find other organizations that are already doing what you want to do. You can learn from them without having to reinvent the wheel. Good luck!