Editor’s Note: The best model for any designer or developer is the one that works well for a particular organization—a model that consistently produces effective learning outcomes on time and on budget. Here,Michael Allen outlines four necessary criteria for the ideal process model, each of which are met byhis Successive Approximation Model (SAM) as an alternative to the ADDIE instructional design model.
By Michael W. Allen
Although for the inventive project manager there’s a temptation to start from scratch, it’s not necessary to bypass the considerable experience and wisdom gained from the millions of projects that have come before. History warns that frustration, late completion, and budget overruns are likely outcomes, even with an experienced team, when using many of today’s traditional processes. Starting from scratch is a great idea when exploring processes, but not when trying to complete projects and honor commitments.
Indeed, while it seems that intuition should lead toward an effective process, one wouldn’t consider such an approach for engineering a highway bridge. And while it may be a case of my problems are tougher than yours, many learning projects, especially those of significant size and those involving technology, are frightfully complex. When the team grows, with concomitant communication and coordination needs, problems snowball. Issues that may have been quickly and easily addressed by a couple of people become problems requiring meetings and much discussion.
We can learn by making mistakes. Giving learners an opportunity to make mistakes and see the consequences is an excellent instructional design strategy when there is little physical danger and when learners have the time necessary to make the mistakes and witness consequences. Typically, there isn’t a lot of extra time to makes mistakes just to learn from consequences when tasked with putting out an instructional product. Although some mistakes will generate immediate feedback, some knowledge of results wouldn’t come until learners actually were using the products. Then the cost of mistakes could be devastating.
Thankfully, we also can learn from the mistakes made and the successes achieved by others. From such mistakes—so many we’ve made personally and also mistakes we’ve recognized as we observed teams work—we recommend the following criteria for model selection.
Criterion #1—The Process Must Be Iterative
A preferred process reveals as much about the developing product as early and continuously as possible. It allows frequent evaluation and course correction at times when corrections cost the least. It prevents investment of the majority of a project’s time and budget before assessments can be made—before learners and stakeholders can get a clear picture of the product (download Figure 4-1: Basic Iterative Process below).
As opposed to a linear or waterfall process in which each step—develop, evaluate, design—is done once, fully, and to the highest level of perfection possible (because subsequent steps build on previous steps), the iterative process takes small, somewhat experimental steps that can easily be reversed or modified several times.
Criterion #2—The Process Must Support Collaboration
Although one-person design and development “teams” are common, others usually need to be involved. Learners, for example, while too frequently are not involved until a product has been nearly or fully finished, have much to offer and should be involved early and continually throughout the process.
Of course, there’s no guarantee that a team of professionals, including sponsors, stakeholders, managers, and learners, will do better than a devoted and passionate individual. In fact, there’s some evidence that individuals eventually produce superior products when given enough time. But it does take a single individual longer than a team, unless the team is poorly coordinated and managed. There are teams that seem almost totally unable to work effectively, whether through lack of leadership, clear process, deep differences of opinion, or other paralyzing factors. Coordinating expectations and roles can make the difference between success and failure when there are multiple team members and stakeholders.
An effective collaboration model is one that makes the most of the ideas, opinions, experiences, and knowledge of people with value to offer. It avoids bureaucracy and indecision that can cause projects to lose their way and waste precious time. It does this by being clear about who is doing what, when decisions are made and by whom, how the work is documented and communicated, and how the process flows.
Criterion #3—The Process Must Be Efficient and Effective
Constraints always exist within which instructional products must be designed, constructed, and delivered. An effective process must work within constraints and also provide breathing room for addressing the unexpected issues that should always be expected.
No project is perfect or will ever become perfect. While the pursuit of perfection is a noble thought, it’s a dangerous goal. The goal cannot be reached, but infinite time and money can be invested only to prove once again that perfection is like infinity itself; one can continually get closer but never arrive. With so many components in a course of instruction, there are many items to perfect. As one component changes, it may well require changes in other components. Attention skitters from one improvement to another while time passes too quickly, and yet perfection fails to arrive. The selected model, therefore, should clarify where and when to focus energy and resources for maximum benefit. It should produce usable products as quickly as possible and allow time for improvements, even those identified late, that will significantly improve impact.
Perhaps most important is that the model produce the best learning experiences possible while staying within given constraints. We should reject a process that creates products with which no one wants to be associated.
Criterion #4—The Process Must Be Manageable
One of the attractive attributes, perhaps even the most attractive attribute, of ADDIE is that it appears manageable. While evidence is lacking that ADDIE actually reduces cost and schedule overruns better than other processes, its defined phases and tasks make PERT chart production possible (download Figure 4-2 below). ADDIE PERT charts spanning 30 pages of tasks have been seen and offer impressive detail. Such charts are helpful to management (and everyone) because of the task requirements they reveal. And manageability is important.
But while clarity is important to manageability, a process doesn’t work well just because we can manage it. We must have both—a process that works and is manageable. And by manageable, we mean it’s possible to complete projects within time and budget expectations, predict the impact of in-process changes, and produce a product that meets established criteria for quality.
So, there we have it—four primary criteria of an ideal process model:
Let’s see how successive approximation measures up.
Excerpt from “Leaving ADDIE for SAM” by Michael W. Allen (ASTD, September 2012). For more information, visit http://www.astd.org/Publications/Books/Leaving-ADDIE-for-SAM.
Michael W. Allen is chairman and CEO of Allen Interactions, whose studios build custom e-learning, provide strategic learning consulting, and train e-learning professionals in collaboration with ASTD. With studios in Minneapolis, Tampa, and San Francisco, Allen Interactions delivers interactive solutions that enhance knowledge, skills, and performance. With a Ph.D. in educational psychology from The Ohio State University, Dr. Allen has pioneered multimedia learning technologies, interactive instructional paradigms, and rapid-prototyping processes. He is an adjunct associate professor at the University of Minnesota Medical School. In May 2011, he received ASTD’s Distinguished Contribution to Workplace Learning and Performance Award. In May 2012, The National Ethnic Coalition of Organizations (NECO) Advisory Committee selected Allen as a recipient of the 2012 Ellis Island Medal of Honor.