Build from the output back to the sources
The usual mistake is to start with “which tools let us capture more things”. We reverse the order: first who must decide, then which question needs an answer, which product they expect to receive, which events could change their judgement and, only then, which sources and procedures we need.
A minimal operational chain can be expressed as:
Decision → question → indicators/signals → sources and coverage → capture → filtering → validation → analysis → product → dissemination → recipient response → adjustment.
It is not a line that is run only once. New findings can change the question, reveal missing sources or show that the chosen frequency was insufficient.
Interactive · Put in order
Put the operating chain in order
Start with whoever decides, not with the tools.
Activity: put the operating chain in order, from decision to adjustment.
- Decision
- Question
- Indicators and signals
- Sources and coverage
- Capture and filtering
- Validation and analysis
- Product and dissemination
- Recipient response and adjustment
Design the service, not just the content
For each intelligence product, describe its service agreement: recipient; decision supported; geographic/temporal scope; inclusion criteria; minimum reliability; level of interpretation; format; delivery date; channel; handling of urgent cases; owner; review. Differentiate products: a critical alert requires speed and essential evidence; a monthly briefing demands comparison, evolution and uncertainty; a quarterly radar needs trends and scenario changes.
Coverage becomes an operational log: active, failed, inaccessible and to-be-reviewed sources. “No news” only means something if we know which sources were checked successfully. The absence of a signal never proves, on its own, the absence of change.
Company A case · simulation
Fictional caseCompany A case: three outputs, three rhythms
Company A decides to produce: (a) a one-off alert if a verified industrial investment announcement exists; (b) a monthly briefing for innovation on players, pilots and patents, distinguishing evidence from inference; (c) a quarterly review of strategic hypotheses with management. Each output has a different audience and a different time requirement. Sending the same newsletter to everyone would be confusing dissemination with usefulness.
The team adds a critical control: if a priority source becomes inaccessible, the incident is logged and the coverage limitation is reported in the next product that depends on it.
Educational simulation: fictional information and organisations.
Do it yourself
Do it yourself · Design a four-week cycle
Choose a decision. Define two key questions, five observable events, between three and six specific sources you would consult, a criterion for validating evidence, a weekly output and a monthly review. Also build an exception procedure: what happens if a critical signal appears or an essential source goes down?
Deliverable: a one-page operational monitoring plan, with quality criteria and an escalation rule.
Common mistakes: assigning arbitrary frequencies; using only general-interest media to monitor official obligations; confusing a source that has been “added” with a source that is “monitored correctly”; not designing how the decision-maker receives and responds.
Key takeaway: the quality of the process is measured as much by what it observes as by what it can recognise it has not observed.
Operational monitoring plan
Turn the design into a sustainable routine.
See the template →Log of signals and evidence
Distinguish observation, source and interpretation.
See the template →Check it · self-assessment
Check your four-week cycle
This part needs professional judgement: there is no automatic marking. Complete your work in the practice, rate it with the lesson's rubric and compare it with the indicative solution.
What your answer should include
- Two key questions, five observable events and between three and six sources.
- Procedure for a failed source: "coverage temporarily incomplete".
- When there is a contradiction, you keep both observations.
- If a reviewer is absent, you assign a substitute or hold back high-risk claims.
Indicative solution
Model solution: sources in R05, a log of checks, an escalation rule proportional to risk, one monthly product and an extraordinary alert.
Mistake to avoid: Interpreting the silence of a failed source as an absence of changes.
Illustrative solution for a fictional case: reasoned alternatives are valid.
Optional extensionExtension · Controlled operation with exceptions
The normal design is not enough. A plan that only describes the week when everything works is not operational. Besides sources and products, three exception states must be described: a critical source does not respond; two pieces of evidence contradict each other; a responsible person is absent. Each exception needs detection, a contingency decision and a reassessment date.
Company A table · Fictional four-week pilot
| Week | Observable work | Evidence that must remain | Control decision |
|---|---|---|---|
| 1 | Agree on the question and a map of 4 families of sources | R05 v1, acknowledged gaps, roles | Approve pilot scope |
| 2 | Run queries and verify critical sources | Execution log, failures and duplicates | Adjust one filter using a sample |
| 3 | Prepare an alert and a briefing with explicit doubts | R06, version and reviews | Decide whether to escalate a signal |
| 4 | Ask for feedback and test the substitute arrangement | Change log and operational test | Continue, correct or stop |
Scenario 1 — regulatory silence: an official source records no news. We cannot infer “no changes” before checking that access, the query and the time period work. Record no news after checking, not no demonstrated activity.
Scenario 2 — contradiction: a headline reports a sale; the original document talks about a pilot. Keeping both records, prioritising the nature of each source and making the uncertainty explicit produces more value than imposing a single label without justification.
Scenario 3 — absence: the person in charge of validating news is absent. For high-impact information, define a substitute or postpone a conclusion. Do not treat a report as approved because the responsible people stay silent.
Optional extension · Process acceptance test
Design a realistic week of operation with two questions and five possible events. Add a control sheet with the columns planned date, source, result, health check, incident, owner and action. Introduce two failures artificially and explain which alerts the team could or could not issue. You pass the test if another person can run the plan and the capture failure is not turned into a false absence of signals. This exercise refers to R05 and L02: it does not replace the detailed design of filters in learning path 02.
Interactive · Decide
Three exceptions, three responses
A plan that only describes the week when everything works is not operational.
Activity: decide what to do about three operating exceptions.
Regulatory silence: the official source records no news.
Contradiction: a headline mentions a sale and the original document, a pilot.
Absence: the person who validates high-impact news is away.
References and limits
- UNE · Asociación Española de Normalización. UNE 166006:2018. Gestión de la I+D+i: Sistema de vigilancia e inteligencia (2018). DocumentReference framework for watch and intelligence systems. It is cited only as context, never as a source of requirements.Reference consulted on 2026-10-18.
- Innguma · internal training material. Construyendo un sistema de vigilancia (presentación, 2022-04-28) (2022).Sources, filtering, management and dissemination.Innguma training material, not public.
- Company A is a fictional case: its data, companies and figures do not describe real markets or organisations.
- The five dimensions and the 0–3 scale are a pedagogical rubric of the Academy: they do not reproduce clauses of UNE 166006, do not replace an audit and do not allow you to declare conformity.
Intelligence Lab · L02
Monitoring plan builder
Ready to apply it? Open the Monitoring plan builder in the Lab and build an editable plan you can download. Output: editable monitoring plan. Works in the browser, no server, no account and no connected AI. The result may need reviewing and adapting to your case.