Media Analysis

Media Intelligence Platform

Signicat Media Intelligence

Mission Control

Build quarterly media-analysis jobs with fewer repeated steps: choose scope, drop the source files, review readiness, and run with confidence.

Next best action

Select regions and topics to begin.

Quarter Not set

Regions 0 selected

Sources 0 / 0 ready

Topics 0 selected

Queue 0 jobs

Ready when everything is green

The page remembers setup metadata in this browser. Research queue, progress, and results are not saved or restored.

📅 Quarter & Settings

Year

Quarter

Autosave active Setup choices, source-file names and URLs, and context are remembered in this browser. Research jobs and results are not saved.

1 Scope & sources

Select the market scope once, then upload matching CSV source files. Bulk upload will map files by region code or country name.

Regions 0

URL Source Files (CSV) 0/0

Upload a CSV file with URLs for each selected region.

Drop all regional CSVs here

Files named with a region code or country name are matched automatically. Individual uploads remain available below.

Select regions first, then upload one CSV per selected region.

Topics 0

Select one or more topics for analysis.

2 Competitor watchlist (optional)

Edit the watchlist below to guide the competitive lens. Leave empty to skip competitor analysis. Deep Research uses this as context, not as a hard filter.

Edit freely. This text is injected into the prompt as-is.

Traditional online media analysis (STRICT CORPUS) → narrative & content strategy for Signicat

PARAMETERS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
• TOPIC: {TOPIC}
• REGION_LABEL: {REGION}
• COUNTRIES_IN_SCOPE: {COUNTRIES_IN_SCOPE}
• START_DATE: {START_DATE}
• END_DATE: {END_DATE}
• OUTPUT_LANGUAGE: {OUTPUT_LANGUAGE}
• COMPANY: Signicat
• PRIORITY_SECTORS: {PRIORITY_SECTORS}
• STRICT_CORPUS_MODE: {STRICT_CORPUS_MODE}
• MAX_PAGES: {MAX_PAGES}

{COMPETITOR_SECTION}

{PILLARS_SECTION}

ROLE
Act as a senior content, brand, and communications specialist for regulated industries (enterprise + public sector), experienced in brand positioning, demand creation, and demand capture.

OBJECTIVE
Analyse traditional online media coverage of {TOPIC} in {COUNTRIES_IN_SCOPE} ({REGION}) between {START_DATE} and {END_DATE} using ONLY the URLs in the provided Excel file.
Exclude social media, forums, and UGC platforms. Allow trade press and government press releases.

STYLE / TONE (non-negotiable)
Write in the style of a McKinsey or Forrester consulting report. Use crisp, declarative sentences. Avoid adjectives like "game-changing", "revolutionary", or "critical". Focus on verbs and nouns.

NON-NEGOTIABLE CORPUS RULES (auditability)
1) Use ONLY the URLs in the Excel file. No other sources, even for paywalls.
2) Aggregators: resolve (e.g., Google News, moreover, etc.) to the final publisher URL; strip tracking (UTM etc.).
3) Canonical date: publisher publication date (not crawl/aggregator date).
4) Deduplication: cluster near-duplicates (wire reprints/syndication/identical bodies). Count ONE canonical article per cluster.
   - Credit outlet/journalist to the canonical publisher version.
   - Exclude pure aggregator pages from outlet/journalist rankings.
5) Paywalls: use visible preview only. If insufficient to code sentiment/framing, mark "Unknown" and set confidence = Low.
6) Language: translate relevant non-English excerpts into {OUTPUT_LANGUAGE} for reporting; keep original in the Evidence Pack.

NEGATIVE CONSTRAINTS (hard limit; prevent gap-filling)
If a theme, claim, event, competitor activity, or Signicat capability is NOT evidenced in the provided corpus text, you must explicitly state: "Insufficient data in corpus." Do not fill gaps with general knowledge, prior training data, or "reasonable assumptions." Prefer "No data found" over generic or speculative advice.

DELIVERABLES
Produce TWO deliverables:

DELIVERABLE 1 — PR/Marketing-ready report (NO links, NO citations, NO bracket references)
- Write in {OUTPUT_LANGUAGE}.
- If {OUTPUT_LANGUAGE} is English: use UK English spelling and tone.
- Professional and accessible; avoid technical ML jargon in the main report.
- Max {MAX_PAGES} pages when pasted into Word; aim ~5–6 pages core narrative.
- Clear headings + tables; no word clouds.

DELIVERABLE 2 — Internal Evidence Pack + Dataset Export (links allowed)
- This is for auditability and internal use.
- Include: quote log with URLs; methodology; QA checks; CSV export of coded dataset.

WORKFLOW (do in order)

IMPORTANT: You have access to Python code execution. Use Python for:
- CSV parsing and data manipulation
- URL processing and deduplication
- Statistical calculations and aggregations
- Date parsing and timeline analysis
- Sentiment scoring and data coding
- Any computational or data processing tasks
This will ensure accuracy and reproducibility.

PHASE 1 — Data handling, normalisation, and coding
A) Ingest all Excel URLs. For each URL, attempt access; if dead/inaccessible, record status and exclude from analysis counts (report how many).
B) Resolve canonical URLs + strip tracking. Store both original URL and canonical URL.
C) Deduplicate into clusters. Choose a canonical article per cluster and justify rule (e.g., earliest publisher version; most complete text).
D) Use your internal Python coding capabilities to ingest this URL list, deduplicate it, and calculate the frequency tables. Do not rely on mental counting. For each CANONICAL article, code the dataset with:
   - canonical_url
   - original_url
   - publisher
   - publication_date (dd/mm/yyyy)
   - country (from {COUNTRIES_IN_SCOPE})
   - outlet_type (national; regional; trade; government; business press; other)
   - byline (exact as shown; if absent: "N/A")
   - stakeholder_voice (who is speaking most: regulator; vendor; bank; insurer; consumer advocate; journalist analysis; etc.)
   - sector_frame (banking; insurance; public sector; vendor/tech; consumer/trade press; fintech/payments; other)
   - themes (multi-label; max 5; use a stable taxonomy you define)
   - entities (orgs/regulators/products; keep concise)
   - competitor_mentions (from watchlist + any other vendors mentioned in-corpus)
   - signicat_presence: none / mention / quoted spokesperson / authored / featured case study
   - sentiment (towards {TOPIC}): Positive / Neutral / Negative / Mixed / Unknown
   - framing_tag: Risk/Threat; Regulatory/Policy; Innovation/Opportunity; Consumer Impact; Market/Competition; Operational/Implementation; Security/Fraud
   - tone_confidence: High / Medium / Low
   - explicit_tag_flags (true/false + short context note): regulation; AI; identity; trust; fraud; authentication; eIDAS/national eIDs
   - duplicate_cluster_id

E) Quality controls (appendix + summary)
- Define sentiment scale and framing taxonomy (with short definitions + examples).
- Define theme taxonomy (10–15 themes) and sector mapping rules.
- Document dedup logic (what counts as "near duplicate"; how canonical chosen).
- "Second-pass consistency check": re-code 30 items and report agreement % for sentiment + framing + primary theme.
- List assumptions/limitations: paywalls, missing bylines, inaccessible URLs, translation limits.

PHASE 2 — Analysis components (Deliverable 1 must include EXACTLY these headings)

A) Executive summary (≤ 1 page)
- Total deduplicated coverage (and how many URLs were excluded/inaccessible)
- Week-by-week trend table (weeks within {START_DATE}–{END_DATE})
- Top 3–6 storylines (written as narratives, not keywords)
- Overall sentiment balance (simple distribution)
- Top 5 outlets by volume
- One-line implications for Signicat's marketing/PR

B) What is being said
- Themes & narratives (explain the "story" and the implied beliefs/assumptions)
- Concerns by stakeholder type (companies; public institutions; consumer advocates)
- Quote 5–10 emblematic lines (translated into {OUTPUT_LANGUAGE}) WITHOUT URLs in this deliverable
- Competitor/other vendor concerns and positioning ONLY where evidenced in corpus (do not invent; otherwise say "Insufficient data in corpus.")

C) Trends over time
- Weekly timeline with annotated peaks (what triggered attention, based only on corpus)
- Emerging angles: first appearance date + subsequent lift (what changed and why it matters)

D) Topics driving the discourse
- Top issues/entities/bigrams grouped into narratives
- Distinguish: (1) prevalence (how often it appears) (2) intensity (how central it is when present)
- Summarise explicit tag occurrences: regulation; AI; identity; trust; fraud; authentication; eIDAS/national eIDs

E) Media & tone
- Top 15 outlets: deduped article count + share of voice (estimated reach ONLY if it exists in the Excel or article itself; otherwise "N/A")
- Top journalists: for each top outlet, 2–3 most active bylines + article counts (or N/A)
- Sentiment + framing distribution by outlet (table)

F) Sector perspectives
- Compare how banking, insurance, public sector, tech vendors, and consumer/trade press frame {TOPIC}
- Highlight consensus vs contention (where sectors disagree, and implications for enterprise/public sector buyers)

G) Conclusions & recommendations (Signicat-oriented, next 12 months)
- Comms gaps ("whitespace"): high external salience + low Signicat presence/quoting in this corpus
- Competitive narrative mapping:
  - What narratives are "owned" in this media window (ONLY if evidenced in corpus; otherwise say "Insufficient data in corpus.")
  - Where ownership is weak/fragmented (opportunity)
- Narrative opportunity scoring (simple, explainable rubric):
  For each opportunity theme, score 0–5 on:
  (1) External salience (prevalence + intensity)
  (2) Buyer relevance (priority sectors)
  (3) Signicat credibility fit (product pillars/certifications)
  (4) Competitor ownership (in-corpus evidence of competitor visibility)
  (5) PR hookability (timely triggers, regulatory moments, incident cycles)
  Then recommend the top opportunities and explain trade-offs.
- Three narrative territories Signicat can credibly lead (12 months):
  For each territory include:
  - positioning statement (1–2 lines)
  - why it matters now (based on corpus trend; otherwise "Insufficient data in corpus.")
  - "red lines" (what NOT to claim / where scrutiny is likely)
  - example headlines (earned + owned)
  - priority sectors + objections to pre-empt
- Content format & distribution strategy:
  - Recommend formats: white paper; blog; opinion piece; webinar; press release; media interview proposal
  - PLUS propose at least 3 new formats not on the list
  - Map formats to demand creation vs demand capture goals
- Media outreach plan (from Top 15 outlets only):
  outlet–journalist pairs, tailored hooks, suggested timing (linked to observed peaks), and sector-specific messaging
- Cross-country whitespace/opportunity table:
  One table with one row per country in {COUNTRIES_IN_SCOPE}, comparing:
  - top narratives
  - high-salience gaps
  - under-owned angles
  - best outlets/journalists (from Top 15)
  - recommended formats and distribution channels
  - key objections/risks and how to address them

DELIVERABLE 2 — Evidence Pack + CSV (internal)
1) Quote log (for every quote used in Deliverable 1):
   - quote (original language)
   - translation ({OUTPUT_LANGUAGE})
   - canonical_url
   - publisher
   - publication_date
2) Methodology & QA: sentiment scale; framing taxonomy; theme taxonomy; dedup rules; second-pass agreement %; limitations.
3) CSV DATA EXPORT — Output this as a fenced code block with ```csv marker.
   - Use the column headers from the first row of the uploaded Excel file as your CSV columns.
   - Add any additional coded columns from your analysis (e.g., sentiment, framing, themes) as extra columns after the original ones.
   - One row per analyzed article/source.
   - Use comma as delimiter, quote fields containing commas.

H) SEO & Content Optimisation Suggestions (for each narrative territory)
For each recommended narrative territory or content opportunity identified in section G, provide:

1. SEO CONTENT DIAGNOSIS
   - Primary search intent: informational / commercial investigation / transactional
   - Content type recommendation: regulatory explainer / industry insight / thought leadership / product-led education / use-case article / comparison / trend analysis
   - Likely funnel stage: awareness / consideration / decision

2. KEYWORD & SEARCH LANDSCAPE
   - Recommended primary keyword for the content piece
   - 5–10 supporting keyword themes or long-tail opportunities
   - "People Also Ask" or recurring question themes relevant to the narrative
   - Featured snippet opportunity, if any

3. SEO ASSET RECOMMENDATIONS
   A. Metadata suggestions:
      - 2 meta title options (optimised, not clickbait)
      - 1 meta description option
      - 1 recommended URL slug
   B. Structural guidance:
      - Recommended heading hierarchy (H1, H2s, H3s)
      - FAQ questions (3–5) that could be added to the content
      - Schema recommendations (e.g. Article, FAQPage)
   C. Internal linking opportunities:
      - 3–5 relevant Signicat pages/topics to link to (based on product pillars)
      - Suggested anchor text for each
   D. Visual opportunities:
      - 1 featured image concept
      - 1–2 in-article visual ideas (comparison table, framework, flowchart)

4. CONTENT BRIEF (for each top-3 narrative territory)
   - Recommended title
   - Target word count
   - Key sections to include
   - Tone guidance: authoritative but accessible; educates rather than hypes; strong European regulatory lens
   - CTA recommendation and placement

NOTE: These SEO suggestions are strategic recommendations, not corpus-evidenced findings. Label them clearly as "Recommended" to distinguish from corpus analysis.

IMPORTANT CONSTRAINTS
- Do not invent facts, quotes, events, reach numbers, or bylines.
- If something is unknown due to paywall/inaccessibility, mark Unknown and confidence Low.
- Separate clearly: what is evidenced in corpus vs what is recommended as strategy.
- If any requested analysis cannot be supported by the corpus, state: "Insufficient data in corpus."

Please provide your response in {OUTPUT_LANGUAGE}.

3 Review prompt

The generated prompt reflects the first selected region as a sample; queued jobs receive their own region, source URLs and context.

Select your parameters above, then let's get started!

4 Run research

Job Queue 0

Your queue is empty. Ready when you are!

📡 Live Research Stream

🔮

Ready to discover insights

Add jobs to your queue and we'll handle the rest

🔮

Here to help — let's find the insights you need

Total Jobs: 0

Completed: 0

Avg Time: --

Elapsed: --

5 Results & exports

Your research results will appear here — quality insights, delivered with care

Media Analysis

Media Analysis

Mission Control

Next best action

Ready when everything is green

📅 Quarter & Settings

1 Scope & sources

2 Competitor watchlist (optional)

🏛️ Product Pillars Include

📤 Output Settings

3 Review prompt

4 Run research

Job Queue 0

📡 Live Research Stream

5 Results & exports