Michael Allen's Guide to e-Learning: CCAF and SAM Explained – A Summary

Michael Allen spent decades at WICAT and then Macromedia working on the technology and instructional design of computer-based learning before founding Allen Interactions. His book is neither a beginner's primer nor an academic overview — it's an argument, built from experience, about what makes e-learning work and why most of it doesn't.

The central claim is that e-learning has been systematically built wrong. Not because designers lack skill, but because the prevailing frameworks — and the organizational pressures behind them — push toward information delivery rather than skill development. Allen's book is an extended case for doing it differently.

#What's wrong with most e-learning

Allen's diagnosis is direct: most e-learning fails because it's designed to present information, not to develop capability. The typical pattern is a content screen followed by a multiple-choice question that tests whether the learner was paying attention. This is not learning — it's coverage.

The problems compound. Click-through navigation means learners can advance without engaging. Feedback limited to "Correct" or "Incorrect" tells learners nothing about why they were wrong or what they should do differently. And the entire experience takes place outside any meaningful context — abstract principles disconnected from the situations where they'll need to be applied.

Allen argues that this pattern survives not because it works, but because it's cheap to produce, easy to measure (completion rates), and hard to challenge organizationally. The alternative requires more upfront thought about what learners actually need to be able to do.

#The CCAF model

Allen's framework for designing effective e-learning has four components: Context, Challenge, Activity, and Feedback.

Context is the situation in which the learning takes place. It's not background information — it's the circumstances that make the task meaningful and that allow the learner to bring relevant knowledge and judgment to bear. Good context creates the emotional and cognitive conditions for learning. Poor context presents information in a vacuum and hopes for the best.

Challenge is the problem the learner must solve. It's distinct from an objective and distinct from a question. A challenge creates genuine uncertainty: the learner has to think, draw on what they know, and make a decision. A typical multiple-choice question after a content screen isn't a challenge in this sense — it's a comprehension check.

Activity is what the learner actually does. Allen consistently pushes for activities that mirror real-world tasks rather than knowledge recall. This is where instructional design most often falls short: the activity is designed around what's easy to build (text inputs, multiple choice) rather than what develops the target skill.

Feedback is the most underappreciated component. Allen devotes substantial attention to it because effective feedback does more than confirm correctness — it explains consequences, makes the learning visible, and connects the activity back to real-world performance. "That's correct. Good job." is not feedback in any meaningful instructional sense.

#SAM: the Successive Approximation Model

The book's second major contribution is the Successive Approximation Model, a development process Allen positions explicitly as an alternative to ADDIE. The critique of ADDIE is not that analysis, design, and development are bad activities, but that organizing them as sequential phases creates serious practical problems.

In a linear ADDIE process, significant design decisions are made early — often before anyone has seen working content. Stakeholder review happens late. By the time problems are identified, the cost of fixing them is high and the organizational pressure to ship is intense.

SAM replaces sequential phases with iterative cycles. The process starts with a Savvy Start — a structured collaborative session where stakeholders, subject matter experts, and designers work together rapidly, identifying examples, generating design ideas, and producing rough prototypes. The goal is to get something tangible into review as quickly as possible, so that feedback is responding to actual content rather than abstract specifications.

Design and development then proceed through short iterative cycles, with each cycle producing progressively more refined content that stakeholders can react to. Problems surface early, when they're cheap to fix. The final product has typically gone through multiple rounds of real review, not a single sign-off at the end.

#Who benefits most from SAM

SAM is most valuable in projects where requirements are uncertain, stakeholders are hard to align in advance, or the risk of late-stage rework is high. It requires a different relationship with clients and stakeholders — one comfortable with showing rough work early — but organizations that adopt it typically report faster delivery times and less costly revision cycles.

#The deeper argument

What holds the book together is a consistent insistence that e-learning should be held to the same standard as any other performance improvement intervention. The question isn't "did learners complete it?" but "can they do the thing the training was designed to develop?"

Allen is particularly sharp on the gap between what organizations measure and what matters. Completion rates, satisfaction scores, and time-on-task are easy to capture and easy to report. Whether employees can apply the skill six weeks after training is harder to measure and rarely attempted. The book argues that this measurement gap shapes what gets built — and not for the better.

Applying the CCAF model and SAM to your organization's learning program requires both design skill and a supportive infrastructure. Scibly is built to support structured, goal-oriented learning — making it easier to deploy the kinds of courses Allen describes without fighting against the platform.