This article reports on the development of a Strategic Technology Plan for the Santa Barbara Museum of Art (SBMA). "Strategic Technology Plan" (STPTM) is the author's name for the plan that guides a museum's implementation of information systems. These plans are also known as Systems Architectures. The purpose of this article is to show how long-range information planning can be done in a short time with minimal cost.
An organization's STP outlines the overall computing models to be implemented, such where minicomputers should be used, and where micros; what databases are required and where they should reside; what computers should be networked; software and hardware standards; implementation priorities; and similar guidelines.
Several important museums in the past few years have developed Information Architectures (or Strategic Data Plans). In general these serve the same purpose, but they focus on the details of data and of information flows in the museum. The result is a comprehensive information-management review of the museum, and a detailed picture of the museum's information and data, and descriptions of all functions and data.
While Information Architectures are of value, we believe that most of the value can be gained for a fraction of the cost by developing an STP instead. This is because an Information Architecture provides not only the overview of data needed for planning, but a great deal of detail about data that is not needed until systems are implemented. Developing an Information Architecture means that costs for the analysis of specific systems are shifted to the planning phase. The drawback of this is that since details about the data can change before implementation, the details must be reviewed later.
Thus we have found the STP to provide a comprehensive blueprint at a minimum cost.
This study was prompted by the goal of the Museum director, Paul N. Perrot, to have a long-range architecture to assist in future decision making. It provided a framework for the Museum to build toward compatibility, connectivity, and efficiency in order to make most effective use of staff, systems, and information. (In this article, only the methodology is discussed in detail, since the recommendations are the property of the Museum. The author gratefully acknowledges Mr Perrot's permission to publish this article, and his review of a draft of it.)
We have characterized an STP as a planning tool. Planning should be the first step in any systems improvement program, which might consist of these four steps:
In other words, a technology investment -- whether a purchase of technology or the development of a new system -- may be cost-effective in relation to its original justification, but be wasteful when seen in the context of all the other programs of the museum. Thus managers' responsibility to make sound decisions about individual systems has been complicated by the need to ensure long-range compatibility and connectivity. Making sound decisions now requires looking at the technology of the museum as a whole, rather than program-by-program or task-by-task. An STP provides conceptual models for making decisions where wide and long-range compatibility is required.
The purpose of using structured analysis techniques is to ensure that all important aspects of the Museum's information needs have been considered. However, techniques alone do not solve problems. All they do is provide a means for attention to be focused on the problems, so that the intelligence and experience of the staff and consultants can be brought to bear.
In addition to structured analysis, which attempts to be purely objective, Systems Planning employed its Critical Information Needs technique to incorporate the subjective views of staff on their information needs.
While some of the techniques of a Strategic Technology Planning process can be substituted for by similar techniques, the ones used for Santa Barbara were chosen for these reasons:
It was not feasible for the consultants to work with the entire Museum staff on all tasks, or even with all staff involved with computers. Instead, various groups of Museum staff were brought into the process as necessary. The composition of the team was determined jointly by the Museum director and consultants before each project task.
In addition to the work done during consultants' site visits, Museum staff performed additional work between visits. This gave the staff additional experience in thinking about their information needs. Furthermore, when staff are involved in a study of this kind, they become more informed about the nature of information management, which is to the long-term benefit of the organization.
Systems obviously exist to further the goals of the museum, so the first step is to determine what these goals are. If the museum already has a strategic plan, this step can be simplified to a review of the existing plan. Otherwise, one begins by developing a mission statement, goals, project plans, and priorities for the museum.
In order to understand the Museum's functions, a one-day session with the director and nine other managers was held to prepare a Functions Hierarchy.
The original list of functions came directly from the Goals developed in the Strategic Planning step. It was then organized into a hierarchy of functions and subfunctions and refined by the project team. The hierarchy was amplified by adding codes to indicate which functions involve existing computer systems.
Dataflow Diagrams show how information relates functions together, and where information is created, used, and stored. The diagrams are critical, since they are used as the conceptual framework for the reminder of the study.
It should be noted that the term "data" is used in a very broad sense in Dataflow Diagrams, to include ideas, plans, intentions, and even art objects, since all of these communicate information, or can be contained in a form, a database record, a document, etc.
Dataflows are a powerful way of looking at functional relationships in the Museum. In fact, it is by developing dataflows that refinement of the Functions Hierarchy occurs. This is because functions in many organizations are defined by the information they create or use.
To initiate the discussion, initial Dataflow Diagrams are generated from the Functions Hierarchy.
During a series of six sessions held over three days, the consultants worked with eighteen Museum staff to develop Dataflow Diagrams. (Each session focused on a different functional area of the Museum, so that no one had to attend the full three days.) The final session on the third day included all staff; the entire set of diagrams was reviewed and some interface questions resolved.
In order to determine the museum's priorities for information, a Critical Information Needs analysis was performed. Critical Information Needs (CINs) are the key data that museum staff require in order to do their jobs.
CINs provide a subjective view of information, and thus act as a check on the formalistic approach of the Dataflow Diagrams.
In order to determine the Museum's priorities for information, a Critical Information Needs analysis was performed.
The Critical Information Needs (CIN) technique is a Systems Planning extension of the Critical Success Factors technique developed by John Rockart at MIT's Sloan School of Management. Critical Success Factors are the activities in which favorable results are critical to the organization.
Critical Information Needs are discovered by first finding what the Critical Success Factors of each job are, and then determining the information needs behind the key activities.
Worksheets were completed by all Museum staff to record their Critical Information Needs, present and future. Staff coded each CIN to indicate the desired format for the CIN (online, printed, etc.), the currency required (how up-to-date), how many people needed this information, and similar characteristics.
CINs indicate the priorities for systems development by calculating a total score based on the importance, the urgency, and the number of persons who need data. This total is in theory the overall benefit to the Museum. Of course, this kind of mechanical process must always be interpreted through management perspective before decisions are made.
A side-benefit of CINs is that the worksheets filled out by staff can form the basis for a requirements statement for new systems, since they describe the major information needs and characteristics.
The Data Matrices are a set of tables showing relationships between people and data, people and systems, and systems and data. This step shows which systems each person needs access to, and thus where additional PCs, software, or terminals are needed.
In the future analysis and design stages for new systems, the tables provide a wealth of information about what is really going on in the museum.
Study of the dataflows revealed that there were nine entities important to the SBMA:
For each of these nine, two spreadsheets were developed. For the first spreadsheet, every dataflow for the entity became a column heading, and the SBMA staff became row headings. This matrix was used to show the data that each person Creates, Updates, Reads (uses), or Disposes of (manages), by entering the codes C, U, R, and D.
A second spreadsheet for each entity had the same dataflows along the top, but down the left side had names of software systems and of physical document storage (such as photo archives and filing cabinets). This matrix was used to show how data related to software and to storage by entering an "X" in the appropriate cells.
Every staff member was given the matrices for entities important to his/her job, and asked to correct and add to the codes.
The matrices are used to study information flow. For example, every data element should have at least one C, R, and D, since otherwise it means that no one is creating it, using it, or has responsibility for its disposition. Most data elements should have at least one X, to indicate what software processes them, or where they are filed. Data that moves from one department to another will generally have CU for the originating department, and RD for the receiving department. Data that has C, U, or D in more than one department is a warning that there may be duplication of data or duplication of effort. Systems or files that have no Xs are obviously either incorrectly coded or no longer in use.
One of the shortcuts that saves time and money in an STP is eliminating the step of building a data dictionary defining every dataflow. A data dictionary would improve the accuracy of these Data Matrices, but for understanding the big picture, they are sufficiently accurate without a one. During the detailed analysis and design for specific systems, these matrices should be refined and a data dictionary built, since in those phases, accuracy of details is critical.
This step is purely administrative. First is an inventory of the museum's hardware and software. This shows the current state of the museum's technology, and what is available for use in new systems. Second is a summary table that show who needs access to shared systems.
The hardware/software inventory is straightforward; data is entered into a spreadsheet or database.
Access needs were tabulated from the Data Matrices for each of the major future systems required by SBMA. This table contributes to determining the cost of each of these system.
One of the most important aspects of a Systems Architecture is the recommendation of appropriate projects to pursue. Over the past several years, Systems Planning has been developing a model for the evaluation of technology-based projects. The model consists of four factors that a manager must consider when determining the risk of failure of a specific project:
Criticality refers to whether the project is Critical, Important, or Noncritical to the Museum's mission. For example, a collections information system is a Critical project; an interactive visitor information system is typically Noncritical (or Important); an accounting system is Critical; an employee cafeteria is Noncritical.
Critical systems are higher risk than Noncritical, because if they fail, the mission of the museum is affected.
Technology Type describes to what extent the technology proposed requires cooperation and interrelationship between the organizational units of the Museum. "Long-linked" technologies imply close coordination among Museum departments. "Intensive" technologies are used by individuals. "Mediating" technologies are used by individuals but require common standards. (These terms are from Organizational Theory.)
Technology Age describes whether the technology proposed is New, Old, or Current. New technologies have the highest risk of failure, Old the lowest. (However, old technologies are usually less cost-effective.)
Experience Level describes the experience of the Museum in the use of the proposed technology (High, Medium, or Low). Low experience means high risk.
This step looks at these risk factors, as well as at costs and benefits associated with developing or acquiring new systems.
Each potential new SBMA system was first evaluated in terms of the four risk factors.
Each of these factors was given a score (1 to 4), and the scores were multiplied to give a total "Risk Score" (1 to 256) for the proposed project. High risk scores mean failure is likely; low scores indicate that projects may be trivial or be using obsolete technology.
However, risk scores alone do not suggest what projects to pursue. In addition, there are the cost aspects, and most important, the value to the Museum. We scored these using numbers ranging from 0 to 250, and calculated a "Total Score" that integrated risk, cost, and benefits.
We estimated benefit on the basis of the number of staff who would use the proposed system, how often they would use it, and whether there were feasible alternatives that would make the system less critical.
The calculation of "Total Score" was (100 * Benefit) / (Risk * Cost). Thus higher costs worsen the Total Score, higher benefits better it. (The 100 factor is just to get numbers that are not fractional.)
It must be repeated here that these kinds of numerical scores must always be interpreted through management perspective before decisions are made.
There is no methodology for this step. However, it is of critical importance. Since policies are the broadest statement of operational decisions, policies are truly the core of the STP. Typical policies describe the museum's standard platform (such as processor type, software, etc.), its network configuration, its position in regard to very new or old technology, and similar statements of broad guidance to future development work.
A critical exercise for a complex organization is to determine where the boundaries of its databases should be. For example, this step would determine whether to combine accounting and membership databases, or whether they should be separate.
For SBMA, as for nearly every museum of its size, the value of using off-the-shelf software supercedes any possible gains from combining databases. (This may not be true for large, complex museums.) Had this not been the case, databases would have been partitioned based on a series of technical procedures which are beyond the scope of this article.
This step consists of scheduling systems development depending on priorities and budgets.
We began by outlining the major activities needed to implement each system, and estimated the time they would take and the resources (money and people) they would need.
We then looked at all the systems and activities for dependencies among them. For example, a LAN-based system obviously depends on implementation of a LAN. We then put each activity into a spreadsheet, with one row for each major system (plus rows for "Information Management" and "Miscellaneous") and one column for each year (except that the first two years had columns for "First half" and "Second half"). Into the resulting cells we put the activities, with liberal comments. The cells also included dollar costs for each activity, which the spreadsheet totalled in both directions.
Project-management software can help here, to manage the dependencies and resource allocations. It is essential for the detailed planning of each project.
The methodology described here is appropriate for small to moderately large museums, and where most systems will be purchased, not developed. For very large or complex museums, the techniques are equally valuable, but supplementary techniques may be necessary to provide higher levels of summary, or for specifying custom software design.
The value of these techniques is not that they can tell you what to do, or automatically develop a plan. Instead, their value is this:
It is difficult to determine precise costs for a study of this nature. However, some of the specific costs for the SBMA were these.
Staff Museum staff spent an estimated 81.5 person-days on the study, including all meetings, information gathering, and document review. Of this, about 3.5 days were the time of the Director. This does not include time for planning the study or choosing a consultant.
Consultants Consultants spent an estimated 32 person-days on the study. This figure does not include proposal writing, project planning, methodology development, or travel time.
Other costs There were no significant other costs besides travel, copying, courier, and similar costs (required because the consultants were not in the same city as the Museum). No special hardware or software was needed -- a desktop computer with modern word-processor and spreadsheet were used for all work. A CASE tool would have been useful to produce prettier diagrams than hand-drawing (Computer Aided Software Engineering tools assist in managing the structured planning, such as balancing dataflows, managing the data dictionary, etc.). Project-management software could have automated the final step.