A standing intelligence requirement has returned nothing in two years - how do you decide whether to retire it?

## Why silence is the hardest input An intelligence requirement that produced nothing looks like a saving waiting to be made. It is also the shape of a requirement whose collection was never actually performed. The three explanations are indistinguishable from the output alone: 1. **Nothing happened.** The activity the requirement asks about genuinely did not occur in your sector or against you. 2. **Nobody looked.** No source was ever tasked, or the task lapsed - a contract renewal dropped the feed, the analyst who owned it moved, the provider deprioritised it. 3. **Somebody looked in the wrong place.** Collection existed but against a source that could never have seen the activity. The first is a finding. The second and third are collection failures wearing a finding's clothes. Retiring on silence alone means you cannot tell which one you are acting on. ## The three tests ### Was collection real? Go to the plan, not to the reports. Was a source named, was it tasked, is that task still in the provider's scope after the last contract change, did the feed keep delivering. A useful cross-check: did that source publish on this activity *for anyone* in the period? A vendor that reported this behaviour against peer organisations and never against you is evidence of genuine absence; a vendor that has published nothing on it at all tells you only about the vendor. ### Does the decision still exist? Requirements die with their decisions. If the requirement was written to inform a choice that has since been made, deferred indefinitely, or absorbed into a programme nobody runs any more, it is dead however lively the reporting is. Equally, if the consumer left and no successor claims it, there is no one to act on an answer - retire it and say so, rather than continuing to collect for an empty chair. ### Did the targeting rationale hold? The most common dead requirement is one that entered the list because a vendor tracked the group, not because anything you operate was ever plausibly in its target set. That is worth checking honestly: what did you conclude this adversary wanted, which of your systems or data matched, and does that match still exist after two years of estate change. If the rationale was never written down, that is itself the answer. ## Retiring properly Retirement is a record, not a deletion: - **Date and rationale.** Which of the three tests failed, and on what evidence. This is what stops the same requirement reappearing next quarter because somebody read a new write-up. - **A reopening trigger.** Name the condition that would put it back: sector reporting placing this activity against manufacturers of your size, an acquisition that changes your exposure, an incident with matching characteristics. A requirement retired with a trigger is a decision; one retired silently is an omission. - **Where the budget went.** Collection has a price - broker subscriptions, contracted provider hours, analyst attention. Retirement is only worth doing if the freed capacity moves to a requirement with a live consumer and a real decision. - **What you keep.** Retiring the question does not mean discarding any cheap standing monitoring behind it, and it does not mean deleting the reporting you already collected. Keep the artefacts and the historical answers; they are the evidence for the retirement itself. ## The judgement to demonstrate The strong answer separates *no evidence of activity* from *no evidence, because we did not look*, refuses to treat a quiet requirement as validation of anything, and then retires on the decision or the rationale rather than on the silence. The weak answer either keeps every requirement forever because 'you never know', which is how a list reaches thirty unranked items, or cancels the quiet ones because they look cheap to cancel - which reliably removes the requirements that were never collected against in the first place, leaving exactly the blind spot untouched. ## Cadence Do this on a schedule with the consumers, not opportunistically. A quarterly review that walks each requirement, asks the owner what they decided with it, and forces an explicit keep-or-retire on each one keeps the list short enough to rank. Requirements that survive three reviews with no answer and no decision are the ones to interrogate first.