Attribution, in the competitive-intelligence sense, is the work of connecting an observed competitor action back to the strategic intent or root cause behind it. A single hire, a pricing-page edit, or a new landing page is evidence; attribution is the inference that this cluster of moves adds up to an upmarket push, or that this lone signal is noise. The output is a stated cause plus a confidence level, not a fact.
The concept inherits from two older fields. Attribution theory in psychology, developed by Fritz Heider and extended by Harold Kelley and Bernard Weiner, studies how people infer the causes of behavior, distinguishing internal from external factors and stable from unstable ones. National-security and cybersecurity intelligence apply the same causal move to adversaries, linking observed tactics, techniques, and infrastructure to a named threat actor and motive, the practice encoded in public frameworks such as MITRE ATT&CK. Competitive intelligence borrows from both. The object differs, commercial rivals rather than threat actors, but the discipline is the same: assign the most defensible cause to the evidence at hand.
Without attribution, competitive monitoring collapses into a feed of disconnected events. A pricing change alone might signal cost pressure, a segment test, or a deliberate strategic shift, and the appropriate response differs in each case. CI teams do attribution to convert raw signals into the small set of strategic narratives that leadership can act on, and to make the reasoning behind those narratives auditable when they turn out to be wrong.
How attribution works in CI practice
Attribution is a triangulation exercise. A CI analyst clusters signals across three dimensions: the surface they appeared on (jobs, pricing page, website copy, blog, news, filings), the time window in which they clustered, and the customer segment the moves appear to serve. The analyst then lists candidate explanations and weighs them by how much of the evidence each one accounts for and how much it leaves unexplained. The winner is the explanation that covers the most evidence and contradicts the least.
The result is recorded with an explicit confidence level, low, medium, or high, and the evidence that would change the call. Continuous monitoring of competitor websites, pricing pages, and job postings, the kind of feed meertrack maintains, supplies the raw material, but the attribution judgment itself is a human inference, not a metric the tool emits.
Attribution vs. revenue attribution vs. data lineage
The glossary treats three attribution-like concepts that answer different questions. Revenue attribution assigns closed-won financial credit across marketing and sales touchpoints, a backwards-looking accounting of which interactions earned pipeline credit. Data lineage and source provenance track where a data value came from, an origin question about a record rather than an actor. CI attribution, the subject of this page, asks why a competitor did what they did, a forward-looking inference about strategic intent.
Confusing the three wastes effort. Treating an attribution call as financial credit leads to false precision; treating it as lineage leads to a paper trail with no strategic punchline. Each belongs to its own workflow and answers its own question, cross-linked where they genuinely meet.
Cluster attribution: a worked example
Consider a SaaS competitor that, within one quarter, quietly hired five enterprise sales reps with named-account backgrounds, rewrote its pricing page around usage-based tiers, and published a new for-enterprise landing page. Read individually, each signal has a trivial explanation: recruiting, a pricing experiment, a content refresh. Read as a cluster, the three moves cohere around a single intent: an upmarket push into the enterprise segment.
The attribution call cites the cluster, names the intent, and states what would falsify it. If the pricing rewrite is reverted within a month or the enterprise reps are quietly reassigned, the upmarket narrative loses support and the attribution is downgraded. Holding the falsifying evidence in view is what separates attribution from storytelling.
Common mistakes and limitations
The most common failure is single-signal attribution, overclaiming intent from one data point. A single senior hire does not prove a strategic shift; a single pricing change does not prove a new segment. Pair every attribution with at least one corroborating signal on a different surface before committing.
Two biases compound the problem. Confirmation bias selects the evidence that fits a preferred narrative and ignores the rest. Motivated attribution, common when a sales team wants a threat narrative to support a deal, cherry-picks the supporting signals and escalates confidence past what the evidence carries. Discipline helps but does not eliminate the risk; recording the alternative explanations that were weighed and rejected is the cheapest safeguard, and the only one that survives when the original analyst leaves the team.