The principle

Software handles scale. Humans handle context. Every clip you receive has been reviewed, tagged, and contextualised by an analyst who understands your industry.

The unit of analysis

The clip — a single piece of coverage — scored across three dimensions and aggregated into a 1–5 impact score for strategic comparison.

The output

Reports that lead with implications, not data. Every chart answers a question. Every paragraph carries a recommendation.

The five-phase pipeline

From signal capture to executive briefing — what happens between the broadcast tower and the boardroom.

One pipeline, four media formats, twenty years of iteration.

The same five-phase architecture runs across online, print, broadcast, and social. The format-specific layers (OCR for print, voice-to-text for broadcast, public-page scraping for social) feed a unified classification and analysis pass.

01

Capture

Continuous collection from licensed publisher feeds, OCR scanning of print, voice-to-text on broadcast streams, and public-page coverage on social.

02

Classify

Software-driven tagging — sentiment, topics, entities, prominence, language. Confidence thresholds determine what flows directly to delivery and what requires human review.

03

Validate

Trained analysts review classifications below confidence threshold and audit a sample of automated decisions. The validation rate is calibrated per client and per content type.

04

Score

The impact score is calculated per clip — prominence × sentiment × reach — yielding a 1.0–5.0 composite for cross-clip comparison.

05

Synthesise

Quarterly and ad-hoc analysis pulls patterns across the period. Executive briefings translate findings into recommendations.

The impact score, in detail

A composite metric — not a black box.

01

Prominence

Where in the publication does the brand appear? Three tiers: Dedicated (the story is about the entity), Mentioned (referenced in context), Shared (one entity among several). Tier weight: 1.0 / 0.6 / 0.3.

02

Sentiment

Tone classification — positive, neutral, negative, mixed. Software classifies; analysts validate every clip below the confidence threshold and a 10% sample above it. Tier weight: +1.0 / 0 / −1.0 / contextual.

03

Reach

Audience exposure — circulation for print, listenership / viewership for broadcast, traffic for online, follower-weighted exposure for social. Logarithmic scaling so dominant outlets do not swamp the score.

The composite

Impact score = Prominence × (Sentiment + 1) × log(Reach), normalised to a 1.0–5.0 scale per client cohort. The formula is calibrated against a quarterly benchmark dataset; the calibration parameters are reviewed every six months by a working group that includes our senior analysts and at least one external academic reviewer.

What we measure beyond the score

The standard metric set every client receives, alongside the composite.

Volume

Total mentions across all monitored channels — the scale of media interest.

  • Daily / weekly / monthly cadence
  • Per-channel breakdown
  • Topic-tag rollups
  • Year-over-year deltas

Share of Voice

Your percentage of total coverage versus a defined competitor set.

  • Volume-based SOV
  • Reach-weighted SOV
  • Sentiment-weighted SOV
  • Topic-restricted SOV

AVE

Advertising Value Equivalent — used judiciously, with the methodology documented per report.

  • Per-clip AVE
  • Multipliers per channel
  • Prominence-adjusted AVE
  • Confidence intervals

Reach

Cumulative potential audience exposure — unique viewers who could have seen mentions.

  • De-duplicated unique reach
  • Channel-specific reach
  • Geographic distribution
  • Demographic estimates

Tone & Sentiment

Software-classified, human-validated. Drift over time is visible in every dashboard.

  • Per-clip tone
  • Aggregate sentiment trend
  • Sentiment by topic
  • Spokesperson sentiment

Topic and Theme

Thematic taxonomies tailored per client — yours, your competitors, your industry.

  • Custom taxonomies
  • Emerging-theme detection
  • Cross-topic correlation
  • Topic / sentiment heatmaps

What our methodology is not

Honest constraints. Things we do not pretend to do.

Not a real-time crisis tool.

For sub-five-minute alerting on social fire-storms, dedicated social-listening platforms are better suited. Our strength is depth and methodology rigour, not raw latency.

Not a substitute for primary research.

Media analysis tells you what is being said. It does not tell you what your stakeholders believe. Pair our outputs with survey work for the full picture.

Not "AI-only".

We use software extensively, but every client deliverable carries an analyst's name on it. If the methodology were fully automatable, we would have automated it.

Not industry-agnostic without onboarding.

The taxonomy, the spokesperson list, the competitor set — these require client onboarding to deliver value. The first 30 days are configuration, not coverage.

See the methodology applied to your industry.

Request a sample report and we will run the impact score against a representative dataset in your sector — methodology notes included.