Agile, Waterfall, or Hybrid: An IF4IT Framework for Choosing Delivery Methodology

Several of the framework’s indicators are expressed using relative terms — work is partitioned into small units, increments are delivered in consistent cycles, a failure causes severe harm. The words small, consistent, and severe have no universal, absolute value. They must be calibrated. The framework does not supply universal thresholds for them; instead, it assigns the responsibility for calibration explicitly to the enterprise, and it identifies who within the enterprise is responsible for each.

What counts as small is determined by the enterprise. A unit of work that is small for one enterprise may be very large for another, because enterprises differ in scale, in the size of their typical delivery efforts, and in the granularity at which they manage work. The framework does not declare a universal definition of small. It instructs each enterprise to calibrate the term against its own typical unit of delivery, and to apply that calibration consistently.

What counts as a consistent delivery cycle is determined by the Product or Service Owner. The framework observes that delivery cycles are most commonly one to four weeks in length, but it stresses that the length is not the point. The defining property of a delivery cycle is the ability to commit to it consistently. A team at a major pharmaceutical enterprise, for example, may begin with a four-week delivery cycle and, as it becomes more capable and more familiar with iterative delivery, shorten that cycle to three weeks. The change in length does not change the team’s answer to the Time of Delivery indicator; the team could commit to a consistent cycle at four weeks and can still commit to one at three. The one-to-four-week range is a useful reference point, not a rule. The genuine test is the ability to commit to a consistent, repeatable cycle.

What counts as severe is also determined by the enterprise, but with an important structural nuance. Some criteria for severe harm are self-evident and require no enterprise calibration. The question of whether a failure will injure or kill a human has an obvious answer, and a harm of that kind is always severe, in every enterprise and every industry. This is the self-evident floor of the Consequence of Failure indicator. Other criteria are genuinely enterprise-relative. Whether a failure will significantly damage the enterprise’s brand depends on the industry the enterprise serves, its public exposure, and its regulatory environment, and a harm that is severe for one enterprise may be minor for another. The enterprise must calibrate this enterprise-relative remainder for itself.

Because Consequence of Failure is the framework’s gating indicator — the one indicator that can determine the outcome on its own — the enterprise-relative portion of severe must be calibrated deliberately and in advance, as a matter of policy. An enterprise should settle what severe brand, financial, regulatory, and reputational harm means for it before it assesses any specific Product or Service, not in the moment of assessing a particular Product or Service it is eager to deliver quickly. Calibrating severe in advance, as policy, protects the framework from the pressure to rationalize a high-consequence Product or Service into a lower-consequence classification simply because the enterprise would prefer the faster methodology.

The burden of calibration is lighter than it may appear. Because the framework is applied chiefly by recognizing the type of work, precise calibration is rarely needed at the extremes — no enterprise needs a precise definition of small to recognize that an aircraft is not small. Calibration matters mainly in genuinely ambiguous cases, and even there the four indicators reinforce one another, so that a borderline judgment on one indicator is usually settled by the others.