Horizon Scanning

Description

Systematic identification of weak signals, emerging trends, wild cards, and discontinuities across a broad environmental landscape. Originated in military intelligence and defense planning, later adopted by foresight institutions (UK Government Office for Science, Finnish Parliament Committee for the Future). The method scans widely rather than deeply, prioritizing breadth and early detection over precision. It serves as the radar system for foresight work — detecting what is coming before it arrives.

When to Use

At the beginning of a foresight exercise, before scenario construction, to generate raw material.
When the topic is broad and the relevant signals are dispersed across multiple domains (e.g., space sustainability involves technology, law, economics, ecology, geopolitics).
When there is a need to challenge assumptions by surfacing overlooked or counterintuitive developments.
When the analyst suspects that the most important factors may not be the most obvious ones.
As a periodic monitoring activity to update an ongoing strategic assessment.

How to Apply

Define the scanning scope. Specify the topic domain, geographic breadth, and time horizon. Decide whether the scan is 360-degree (all domains) or targeted (specific sectors).
Establish scanning categories. Use a structured framework (e.g., STEEP: Social, Technological, Economic, Environmental, Political) to ensure no domain is missed. Add domain-specific categories as needed (e.g., Legal/Regulatory for space topics).
Collect signals. Gather data points from diverse sources: academic papers, patents, policy documents, industry reports, news, social media, expert interviews. Prioritize heterogeneous sources to avoid echo chambers.
Classify each signal. For each item, determine: Is it a weak signal (isolated, ambiguous, possibly important), an emerging trend (pattern forming, direction visible), a wild card (low probability, high impact), or a known trend (established, widely recognized)?
Assess signal strength and relevance. Rate each signal on: novelty, potential impact on the topic, speed of development, degree of uncertainty. Flag signals that appear in multiple independent sources.
Cluster and map signals. Group related signals into thematic clusters. Map them on an impact-vs-uncertainty matrix or a timeline to visualize the landscape.
Identify key implications. For the highest-impact clusters, write brief implication statements: “If this signal strengthens, it could mean X for Y.” These become inputs for scenario planning or trend analysis.
Document and update. Create a living signal database that can be revisited and updated as new information emerges.

Key Dimensions

Signal type — weak signal, emerging trend, wild card, established trend
STEEP+ domains — Social, Technological, Economic, Environmental, Political, Legal
Signal strength — number of independent sources, consistency, corroboration
Novelty — how new and unexpected the signal is
Potential impact — magnitude of consequences if the signal materializes
Speed of development — how quickly the signal could become a dominant force
Cross-domain interactions — signals from one domain that amplify or dampen signals in another

Expected Output

A structured signal inventory (table or database) with classification, source, and assessment for each signal.
A signal map (impact-vs-uncertainty matrix or cluster visualization).
A shortlist of highest-priority signals and wild cards with brief implication statements.
Recommended signals to monitor over time (watchlist).

Limitations

Breadth comes at the cost of depth: horizon scanning identifies signals but does not deeply analyze them.
High noise-to-signal ratio: most collected items will turn out to be irrelevant.
Vulnerable to source bias: if scanning sources are homogeneous, blind spots persist.
Wild cards are by definition hard to detect — the method can systematize the search but cannot guarantee discovery.
Requires regular updating; a one-time scan decays rapidly in fast-moving domains like space technology.
Weak signals are inherently ambiguous — analysts may over-interpret or under-interpret them depending on prior beliefs.