[Skip to content](https://academy.innguma.com/en/analysis-foresight-and-anticipation/weak-signals-and-trends-avoiding-both-noise-and-false-certainty/#contenido) Lesson 5.3 · Learning path 05 # Weak signals and trends: avoiding both noise and false certainty How do you treat an early indication that could be relevant but is still ambiguous? **You take away** a signal log with mechanisms for cross-checking and follow-up. ≈ 30 min · Advanced [Do the practice](https://academy.innguma.com/en/analysis-foresight-and-anticipation/weak-signals-and-trends-avoiding-both-noise-and-false-certainty/#hazlo-tu-registro-de-senales-y-falsadores-20-min)[Exercise](https://academy.innguma.com/en/analysis-foresight-and-anticipation/weak-signals-and-trends-avoiding-both-noise-and-false-certainty/#compruebalo)[Read the lesson](https://academy.innguma.com/en/analysis-foresight-and-anticipation/weak-signals-and-trends-avoiding-both-noise-and-false-certainty/#senal-patron-y-tendencia-no-son-sinonimos) In 1 minute Signal, pattern and trend are not synonyms. ## Signal, pattern and trend are not synonyms A **signal** is a potentially relevant observation; a **weak signal** is an indication whose meaning, scope or trajectory is still uncertain and may go unnoticed; a **pattern** is a set of related observations; a **trend** is a direction of change that requires sufficient evidence over time and explicit criteria. The word «weak» does not necessarily mean low impact: it means that the interpretation or its support is still limited. Weakness has different components: **low visibility, little confirmation, novelty, ambiguity, temporal distance and scarcity of data**. Separating them helps decide what to investigate. A highly visible phenomenon can still be ambiguous; a little-known fact can be perfectly well documented. Interactive · Cards ### Signal, weak signal, pattern and trend Flip each card to see the definition and a Company A example. Signal, weak signal, pattern and trend: definitions. ## Five questions for prioritising an indication 1. **What exactly do we observe?** Reduce the description to a verifiable fact. 2. **What is its origin?** One primary source and ten republications are not eleven corroborations. 3. **Why might it matter?** Link it to a KIQ, a decision or a defined risk. 4. **What alternative explanation is possible?** Tell a structural change apart from an isolated anomaly. 5. **What confirmation or refutation do we expect?** Design indicators and a review date. The purpose of logging signals is not to get the future right, but to **detect what deserves targeted follow-up before an explanation becomes entrenched**. ## A signal log that learns Each entry must contain: identifier, observation date, source and URL, neutral description, topic and actor, possible mechanism of change, hypothesis, alternative explanation, relevance for a decision, uncertainty, follow-up indicator, owner, next review and status. Useful statuses: observed, under cross-checking, reinforced, weakened, closed. Closing a hypothesis does not mean erasing what was learned. It is also worth logging **observation biases**: if we capture only Spanish-language press, we will see some events more likely than others; if we consult only news, we may miss regulatory or technical documentation. Alerts on social networks or private platforms do not prove full coverage either. Interactive · Simulation ### The life of a signal in the log Simulation: how a signal in the log evolves over three months. Company A case · simulation Fictional case ## Company A case: three indications and two narratives In the fictional case, Company A observes a growing number of mentions of thermal storage at events, a supplier note about a pilot and an indirect regulatory reference. Narrative A holds that an opportunity for industrial adoption is emerging. Narrative B interprets a temporal concentration of communication with no substantial change in real operation. Indicators are agreed: verified contracts, installations with measured performance and regulatory changes actually approved. At the monthly review, if there is no confirmation, the team **does not automatically turn the absence of news into falsification**: it first checks the health of the sources and then reviews qualitative probabilities or the status of the hypotheses. Educational simulation: fictional information and organisations. Do it yourself ## Do it yourself · Signal log and falsifiers (20 minutes) Start from six publications from the fictional Company A corpus or from a well-documented set of your own. Identify which are original and which repeat other facts. Formulate two rival hypotheses about a signal and define **one finding that would strengthen each hypothesis and another that would weaken it**. Assign a weekly, monthly or quarterly review, justifying the cadence. **Deliverable:** a signal log with mechanisms for cross-checking and follow-up. **Common mistakes:** prioritising by virality; labelling everything as a «trend»; deleting discarded signals; ignoring source coverage; asking a model to invent signals or probability percentages. R06 · Evidence ### Log of signals and evidence Distinguish observation, source and interpretation. [See the template →](https://academy.innguma.com/en/library/log-of-signals-and-evidence/) R10 · Analysis ### Matrix of hypotheses and alternatives Assess rival explanations of a signal. [See the template →](https://academy.innguma.com/en/library/matrix-of-hypotheses-and-alternatives/) Check it · closed answer ## Signal triage Before escalating, identify the origin and the status of each signal. 1. A supplier announcement and three echoes in the media. **See the answer** **Answer: A single origin.** One announcement and three echoes make up one origin; calling republications a trend is the critical error. 2. A regulatory proposal under consultation. **See the answer** **Answer: Provisional status.** It keeps its provisional status: it can justify an urgent cross-check, but not a claim that it is already in force. 3. A commercial presentation at a trade fair. **See the answer** **Answer: Does not prove contracts.** A commercial presentation does not prove contracts. A confirmed change is communicated to management only if it affects a decision and has sufficient support. ### And now, your own answer 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 rival hypotheses about a signal. - One finding that would strengthen and another that would weaken each hypothesis. - A justified review cadence. Self-assessment rubric (0–2 per criterion) An answer that asserts unsupported facts cannot reach the maximum score. It is not a professional certification. Traceable evidence Alternative interpretations Proportionate decision Review and limits **Indicative solution** At the monthly review, if there is no confirmation, the team does not automatically turn the absence of news into falsification: it first checks the health of the sources and then reviews the status of the hypotheses. Illustrative solution for a fictional case: reasoned alternatives are valid. **Optional extension Going deeper · A weak signal needs a mechanism, not an adjective** A weekly review turns up seven entries on thermal storage. Five come from the same press release, one is a technical standard under consultation and another reports a pilot installation. If the analyst concludes «seven signals reinforce a trend», they have confused **the volume of publications** with **the number of independent facts**. The first step is to deduplicate; the second, to formulate the mechanism that would connect the facts to a business decision. ### Two-step triage protocol **Step A: quality of the fact.** Document the primary source, date, jurisdiction, context, possibility of replication and limitations. If a piece is missing, it is not compensated for by arbitrarily raising relevance. **Step B: possible consequence.** State which condition of the environment could change, how it would affect Company A and what other explanation would be consistent with the same indication. | Simulated entry | Independence | Possible meaning | Future test | | --- | --- | --- | --- | | S-A: original pilot announcement | Original | Adoption trial | Evidence of continuous operation | | S-B/S-C/S-D: republications of S-A | Dependent | Spread of the announcement | They do not count as three corroborations | | S-E: draft regulation under consultation | Original, provisional | Possible regulatory change | Final text and entry into force | | S-F: commercial presentation | Interested origin | Declared intent or positioning | Contract, installation or external evidence | A **follow-up indicator** is not a prediction. «Check every month whether identifiable operating installations appear with a primary source» is observable; «monitor whether the market takes off» does not define an operational criterion. ### Deciding when to escalate There is a difference between *logging*, *opening targeted follow-up* and *escalating to management*. Escalating requires a link to a decision, sufficient credibility and a relevant time window. On high-impact topics, an insufficiently proven fact can justify **verifying urgently** without justifying acting as if it were true. Conversely, a signal that is already confirmed but irrelevant to current decisions can stay under ordinary follow-up. **Optional extension (30 min):** from six entries (one original, three echoes, one preliminary document, one interested source) create at most four useful records in R06, with rival hypotheses, the condition that refutes them, an owner and a review date. Justify whether you send an alert to management or keep the cross-check internal. **Acceptance:** the echoes are grouped and each step has a criterion and a date. ## References and limits - Richards J. Heuer, Jr. · Center for the Study of Intelligence, CIA. *Psychology of Intelligence Analysis (cap. 8, Analysis of Competing Hypotheses)* (1999). [Document](https://www.cia.gov/resources/csi/static/Pyschology-of-Intelligence-Analysis.pdf) Basis of the rival-hypotheses method: the aim is to refute hypotheses with inconsistent evidence, not to confirm one. Reference verified on 2026-10-18. - Company A is a fictional case: its data, companies and figures do not describe real markets or organisations. - The analyses do not assign probabilities or scores without an explicit rule. Scenarios are plausible worlds, not forecasts. Intelligence Lab · L04 ## Signal log Ready to apply it? Open the Signal log in the Lab and keep observation, source and interpretation separate in your log. Output: signal log. Works in the browser, no server, no account and no connected AI. The result may need reviewing and adapting to your case. - [Open the tool](https://academy.innguma.com/en/lab/signal-log/) - [Back to the learning path](https://academy.innguma.com/en/analysis-foresight-and-anticipation/) **Download** - [R10 · Matrix of hypotheses and alternatives](https://academy.innguma.com/en/library/matrix-of-hypotheses-and-alternatives/) - [R06 · Log of signals and evidence](https://academy.innguma.com/en/library/log-of-signals-and-evidence/) Open reading, no registration.