Category: R2O

Listen to the community and refine R2O

Research to operations requires an astute and routine identification and evaluation of the stakeholders. The most efficient way to do this is holding conferences, in person, that pull unique participants from the breadth of researchers, practitioners, and leaders within, and beyond, the community. This means engaging colleagues beyond the geographic and disciplinary reaches of current R2O activities. There are usually many international partners and adjacent sectors that can provide a fresh perspective on how to optimize R2O activities based on their own trials. Taken to heart, these perspectives can challenge the status quo through reframing R2O discussions. After all, R2O is about collaboration. It is about the people involved more than the process that institutes and prescribes it.

It is not as easy as calling a meeting and buying some coffee and muffins though. Planning a conference and establishing an agenda that focuses on R2O necessitates placing a notable emphasis on who is in the audience, not who should be presenting. Conferences that seek to serve as a springboard for new R2O ideas or concepts for improving R2O itself must hear from those “in the trenches” instead of “on the hill”. Successful R2O is in its details.

There is a clear way to identify whether an agenda has strayed too far from the ideals of connecting people and exploiting opportunities. If the program on paper appears like an itemized organizational chart, with the senior leadership presenting first, then the mid-level management, and finally project managers, then the conference will be more about affirming the present than changing for the future. The audience is likely to consist of a mixture of researchers and practitioners that do not necessary identify with the process-centric perspective that managers usually proffer anyway. Hearing the internal community perspectives as well as those from aboard and beyond can stimulate a thorough discussion about future directions and energize everyone involved.

Conferences that connect both players in the R2O cycle must translate the discussion to a consensus so that the promising ideas for future R2O transition or process improvement are identified. Once they are identified, further conference interactions should aim to produce actionable partnerships and proposals. This requires less talking to the audience and more hearing from the audience, or giving the members of the audience the opportunity to talk amongst each other. In other words, the program should consist of fewer presentations and more town hall meetings, panel discussions, and breakout groups.

Beyond that, inviting international collaborators to partake in events at R2O conferences can assure a different perspective is shared. This could be particularly advantageous for R2O projects with substantial government involvement. After all, government processes can be resistant to change because there are no external market forces to demand it, even if budgets are tight. Yet, international government partners likely have similar challenges and may resolve them differently.

But it has to be easy for all of the R2O players to air their opinions and provide potential solutions. This is difficult to force, but encouraging the willing conference attendees to address how they see the present climate and future evolution of the R2O environment in the perspective of their own R2O projects instead of what they are doing is a major step. The “how” provides the details. There are endless examples of R2O but examples of R2O, even successful ones, are not really telling of best practices on their face.

Furthermore, from a science perspective, conferences, including those with international collaborations, can be beneficial to the entire community in producing the best research approaches that can then translate to improved research byproducts. Parallel algorithm development teams may be in competition, whether formally or informally, but there is joint interest in determining the merits of each solution, scientific or otherwise. The best criticism and ideas can sometimes come from the outside.

Bringing people together to build collaborations and expand the community requires use of that capacity in demonstrating its benefit. To that end, the largest community conference can still be unsuccessful if it is nothing more than presentations directed to an audience. Encouraging attendees to contribute to the forum and build their personal networks beyond their discipline, organization, or nation is essential for the growth and refinement of R2O. Listen to them.


Beginning, evolving, and concluding R2O projects

An enterprise or organization could become throttled by its own success when research is routinely transitioning to operations, new ideas are arising from both research and operational workforce segments, and improvements within existing projects are continuing. Leaders with fixed budgets and workforces with little unallocated time will complain that they cannot do everything or they cannot continue everything. This is a particular challenge for enterprises with substantial government participation because market conditions do not necessary guide the path ahead. Should we put time, money, and effort into this, or that? Where is the greatest benefit? What will improve our services the most? When should we wind down a project?

Beginning R2O projects

For new R2O projects, it is important to decide whether an idea or concept has the potential to evolve an enterprise or organization strategically or resolve an existing challenge tactically. R2O projects that are strategic in nature will require a substantial degree of involvement and support from senior leadership, whereas tactical R2O projects can succeed with relatively minor management support. The major reasons for this are scope and impact.

Strategic R2O projects keep an organization on the front edge of an enterprise in terms of ideation and innovation, and seek to remedy prospective shifts in community or market conditions before an organization falls behind the rest of the enterprise. As such, some strategic R2O projects could, in the end, impact many aspects of an organization. These types of projects are those that rethink paradigms for delivering information, interpreting data, and making decisions. They have a substantial cost and time commitment and the end state will have a bearing on the types of products or services that customers will receive.

Though they may take several years, strategic R2O projects are not just “corporate” change initiatives. Projects of this nature will require broad participation from the research and operational sectors of an organization or enterprise. In addition, strategic R2O projects do not presume or initially prescribe an end state; that evolves over time. Early in the project, there will be numerous ideas, and in a “fail fast” model, eventually the tractable ones will continue to be refined and improved as the R2O cycle continues. In other words, strategic R2O projects start with the “what” in order to define the “how” as part of the process.

In contrast, tactical R2O projects solve existing problems that people recognize as hindering an organization or limiting the potential of its services. Tactical R2O projects start with the “why” and work toward the “how”. Problems vary in size and type, and, as such, the amount of time and effort required to produce a solution will depend on the situation. Generally, tactical R2O projects should not require a substantial degree of funding beyond the personnel time to tackle the issue though. Projects of this kind should produce tangible results in about one year. A tactical R2O project team may be sufficient with a single research representative, a technical assistant, someone from operations, and possibly a manager familiar with the challenge at hand to sponsor any modest resource requests from the rest of the team, as well as keep senior leadership aware of the project.

Assembling teams of three or four ensures that enough people, and potentially their managers, feel it is worth their time, and that the work will be efficient (i.e., all of the components for productive R2O are in place). An organizational requirement of a minimum of three people on a R2O team requires that someone who identifies a problem is able to recruit at least two other people and convince them that it deserves their attention. It is better yet if the team members are geographically disperse, especially for organizations of regional, national, or global scale. Three or four is not an upper limit. Some tactical R2O projects can grow upscale with time.

Evolving and concluding R2O projects

As R2O projects become mature enough, leaders should adjust the requirements behind the mission of an organization or enterprise based on the successes of R2O. This helps keep the scope of R2O projects in check over time, and identifies future directions. This is particularly important because people may like or opt to continue to work on a successful project until they see it as complete, which may differ from when additional cycles are no longer valuable to meeting the requirement. A requirement also ensures that there is funding to support the R2O project deliverable in subsequent years. That is, R2O project completion, or conversion to maintenance, should be defined as a metric from the requirement, not from the members of the team or the initial charter. This is tricky because neither strategic nor tactical R2O projects need to, or should, begin with a specific requirement. The cyclic nature of R2O ensures that it remains exploratory, particularly near a project’s inception. But eventually the pursuit of perfection may create too much “tunnel vision” toward the existing project and obscure new R2O opportunities.


Operations-driven R2O with science-integrated training

The ideas that initiate a research to operations transition cycle need not always originate as part of the research process. Some suitable concepts for research development can and, with the right governance structure in place, do come from operations. This constitutes operations-driven R2O. Operations can serve as a strong driver for R2O in cases where practitioners proffer an idea because the potential application and need is understood. Therefore, early adoption of byproducts or techniques resulting from operations-inspired research is easier.

Arguably, well-founded ideas—those formulated with a preliminary “how” to accompany the “why”—for R2O projects that originate in operations have the potential to be more successful and more impactful than those that come from the normal progression of research projects because they escape an improvement sequence. However, the challenge is that operations may not have a sufficient knowledgebase to fully appreciate how certain observations and derived research byproducts could be combined, improved, incorporated, or refined, even if the approximate nexus between research and an operational application is evident. Well-intentioned practitioners desire to assist R2O efforts at the “ground floor” to shape the nature of the eventual application early in the process. Ensuring the proper scientific background of partnering practitioners in such efforts is important to forming a mutual working relationship between research and operations.

The best way to accomplish this is through training. Though training approaches are diverse in their approach and reach, routine training on new observations and derived research products is important in establishing operational functions that are performed with the greatest possible capability. Training enables an evolution of the workforce and an enterprise because the end result is personal and organizational potential for higher performance, through individual abilities and collaboration.

With proper training methods, the amount of learning is directly proportional to the amount of training. When there is more training, there is more generally more learning, as long as the training approach is designed appropriately for its purpose. There are two general purposes for training: (1) developing and/or assuring familiarity with set applications and (2) building expertise to envision potential additional applications. The latter case for training, which is generally more comprehensive, involves a greater inclusion of scientific principles and studies in those fields with an academic component. Training of this nature can support operations-driven R2O. It can also ensure that a workforce is ready to fulfill an organizational mission with high-quality services in complex scenarios.

Learning growth with science-integrated training. Figure developed by Jordan Gerth.
Figure: A greater fraction of training activities for building expertise must incorporate scientific principles. In comparison, smaller applications-based training efforts are designed to meet current operational needs but do not ready a workforce for an expansion or improvement of services.

This concept is illustrated with the learning growth scale. The learning growth scale depicts a growing fraction of science-integrated training relative to a training battery. In order to develop expertise, practitioners must not only engage in a high degree of training, but also a large amount of science-based training. This contrasts to training toward an application, where a training program can minimize scientific background and advanced theories.

Training to applications focuses on the “what”, “when”, and “where”, while the science-based training focuses on the “how” and the “why”. All are important, but the “how” and “why” position the practitioner beyond a common set of scenarios. Consider an example from meteorology: a tropical cyclone, or hurricane. An applications-based training approach would provide the practitioner with several examples of tropical cyclones and explain their time and space characteristics. A science-based approach would explain why tropical cyclones form and how they are sustained, with a greater amount of explanation of the water temperature and wind shear contributions to a tropical cyclone lifecycle.

Some practitioners may scoff that they “don’t need a Ph.D.” to succeed at operational tasks. But comments like this are misplaced, unless the training program truly does have practitioners completing entire graduate-level courses, which is probably ill advised. Comments like this most likely result from limited or no training content to connect the science back to applications. Science-based training should not lack applications. An advanced and thorough training program should be comprised of scientific lessons that are related the operational setting. Operational interests should still dominate the training, though a large amount of introductory material may be rooted in science. The training should be presented in a way that creates an opportunity for discussion, which can only occur with a suitable time allotment for training.

Successful R2O requires a time investment, especially for operations to drive the transition. Science-integrated training not only benefits R2O process, but also increases the capacity of a workforce. This, in turn, improves and evolves services.


Science first, service always

In the weather enterprise, there are numerous service organizations. The services that these organizations provide vary, but generally involve analyzing, interpreting, and, in many cases, forecasting meteorological and environmental parameters. Service organizations not only exist in the private sector, but in the public sector as well. The National Weather Service and National Environmental Satellite, Data, and Information Service are large service organizations in the federal government. Service is in their name. Service organizations in scientific fields, like meteorology, can struggle to find an internal identity because, when it comes to core service improvement, there is a tension between focusing on development to better services (e.g., physical science research-based enhancements) and improvements to the delivery of those services (i.e., the nature of the communication).

While rarely is a service organization in a science-related enterprise completely devoid of science, budgetary and management priorities toward improving service delivery can set a tone that discourages substantive research progress. Research is the best way to produce incremental change in the accuracy, and arguably quality, of scientific services. The end user experience with a prospective service improvement should not be overshadowed though.

“Research to operations” could be restated as “science to service” in order to emphasize the importance of science in the service organization. In other words, the R2O cycle is not exclusive of the services that an organization provides. How a research byproduct caters to a service deliverable must be considered. After all, the purpose of R2O is for research to meet a tailored need in operations. Operational entities cater to users or customers. In the weather enterprise, the end users are often the public, businesses, and other government agencies that consume information. And these consumers should acutely guide operational activities with their needs.

After many meteorological and environmental disasters, we often hear how it “struck without warning” as if no one knew it was coming. Other than earthquakes, there is almost always some early notification of impending hazards provided via traditional means of communication. And, as most people receive it and take heed, they likely pass that information to people who missed it. Some threats are easy to convey, such as a tornado or volcano. People understand that these are always life threatening. But how about floods and heat waves? Do these types of disasters—that have varying magnitudes, from barely an inconvenience to deadly—result in the public widely taking preventative action?

Unfortunately, recent history tells us that they generally do not. But is the prevailing problem that the means of communication were insufficient, or that the message was not conveyed properly? The breakdown probably occurred because the messaging lacked some combination of pragmatism, credibility, and certainty. These are three items that science can improve substantially and messaging can only improve marginally.

For example, a weather forecast could call for ten inches of rain, with at least six inches highly likely. This may be a credible forecast but very difficult to convert to certain impacts because the impacts are reliant on the scenario. In other words, it might be obvious to anticipate ponding of water in low spots, but what roads will become impassable with that amount of rain? How many homes will be underwater? With the highest geographic precision, the forecast must be certain and the impacts must be certain.

Improving the meteorological prediction and developing specific environmental models to convert observations to impacts is inherently scientific. It is impossible to communicate or base a service on information you do not have, or cannot be stated easily, confidently, and specifically. In this sense, in order to become a strong service organization, an advanced scientific research program is a necessary component. A connection must exist between this program and the rest of the organization. Let an organizationally integrated approach to R2O promote science first in support of services always.


Data sparse or information poor?

DRIP (data rich, information poor) has been used for decades to describe the ineffective or absent use of data that an organization or enterprise collects daily to inform important decisions involving certain conditions that could impact the mission of that entity. Breaking the curse of DRIP involves collecting, storing, organizing, and accessing data so that it can be converted into information. That information must subsequently be archived, analyzed, and ultimately used to create a DAIR (data and information rich) environment. Gleaning information from data is challenging, especially if the data is collected for a purpose other than converting it to information, or the information sought requires combining multiple data sources.

Correcting for DRIP postures requires both technical and scientific acumen. In science, data is collected in search of information. In many cases, research is conducted to provide information from data. The R2O cycle seeks to transition the new research information that is apposite for the practitioner. The transition that the R2O process facilitates serves to refine the information, and consolidates or reformats data into parameters or quantities that the practitioner understands. A practitioner may also convert data to information through their mental model (e.g., a day with a temperature of 95 °F and a dew point of 75 °F is hot and muggy).

In meteorology, many consider the oceans of the world to be “data sparse”, meaning there is not enough data collected from the high seas to fully characterize the atmosphere and its underlying dynamical and physical processes that are at play there. This is in spite of the numerous observations from weather satellites, as well as the occasional reporting buoy or overhead aircraft with the appropriate sensor package. To that end, the atmosphere over the oceans is really “information poor”; the current observing systems, despite the data they provide, are not producing, or contributing to, actionable information in the eyes of the practitioner. However, there is quite a substantial amount of data.

To be fair, there are certainly portions of the earth where the observations are so lacking that more data is necessary. But more data is not always a potent medicine for limited information. Nor is it always easy to determine whether there is a fundamental limitation on the data that already exists. In the context of R2O, it is necessary to examine how researchers are interrogating the data that is available and how practitioners are applying the information to serve the consumers. New analytical research methods could provide clues that develop additional information from existing data. Similarly, altering the nature and/or amount of the information (or data) that practitioners review could change its interpretation.

It is not obvious whether animating a time sequence of spatial data/information at different speeds may somehow alter how a practitioner perceives the evolving scenario. For example, does a growing thunderstorm appear more severe if it is viewed in a fast animation instead of a slow animation? Does a different palette used to colorize the images in the animation alter that perception? Does the temporal frequency at which the images were collected matter? An interesting recent piece of research investigated how the playback speed of videos depicting violent conduct impacts viewers’ judgment of the intent of the actor. How do different frame rates influence scientific practitioners attempting to understand a complex atmosphere?

For successful R2O that respects the value of observations, we must learn more about how people process information. And information must be confined in the context in decisions. To reach that point, there must be adequate research with ample data. But challenges will continually compound as a desire for greater storage capacity, faster computers, and innovative methods accompany the “big data” future. Even still, the continuum between the most data and the best decisions (via information) is not a linear function. Instead, information must be tailored and actionable to be rich.