Skip to content
Guides/Mentions & Monitoring
Product guide

Keywords Monitoring & Setup: The Complete Guide

Master brand antennas, competitor benchmarks, Boolean query rules, real-time AI diagnostic audits, and media channel footprint inspection.

Updated Nov 19, 2026 · 23 min read · ElixBrand Product Team

Monitored keywords form the foundational sensory antennas of your ElixBrand workspace. Every brand mention, executive quote, crisis alert, and competitive insight ingested across web, press, broadcast, podcast, and social channels begins with a configured keyword rule. This comprehensive guide details the complete operational lifecycle of monitored keywords: navigating the Keywords Overview dashboard, establishing primary and competitor benchmarks, configuring precision Boolean and phrase-matching parameters in the unified setup wizard (/keywords/add), evaluating signal quality through the 4-dimension ambient AI diagnostic engine, and analyzing media distribution with the Channel Footprint drawer.

Foundations & Ingestion Architecture

In ElixBrand, a monitored keyword is an intelligent search antenna that continuously scans our live media feed for matching content. Rather than relying on rigid, one-size-fits-all string matching, ElixBrand evaluates each keyword rule against a high-velocity stream encompassing social networks, editorial news publications, regional web portals, industry blogs, community forums, customer review platforms, podcast transcripts, and generative AI search responses.

When an active keyword rule is saved or updated, the system compiles its parameters—including primary terms, aliases, match modes, required context anchors, negative exclusions, platform selections, languages, and geographic borders—into a unified project search filter. Background ingestion crawlers and streaming pipelines evaluate incoming documents against this compiled filter, attributing verified matches to your project feed.

Every keyword maintains independent telemetry, calculating 7-day volume trends, rolling audience reach, net sentiment distributions, toxicity rates, and channel concentration ratios. This enables marketing, PR, and risk teams to isolate trends for specific product lines, track individual executive visibility, and benchmark rival brands without data bleed.

  • Continuous live ingestion across social channels, editorial news, blogs, forums, podcasts, and AI engines.
  • Unified project query compilation guarantees high-speed multi-keyword filtering without duplicate ingestion overhead.
  • Granular per-keyword telemetry tracks isolated volume growth, rolling sentiment, reach exposure, and channel footprints.
  • Full support for multi-script characters including Latin, Arabic, Cyrillic, Greek, and CJK alphabets.

Keywords Overview Dashboard & Quota Telemetry

The Keywords Overview page (/dashboard/projects/[id]/keywords) serves as the mission control center for all tracking antennas configured in your workspace. At the top of the interface, the header displays your workspace's real-time keyword quota utilization ({activeCount}/{allowedKeywords} Quotas Used), reflecting the number of actively monitored terms against your plan allowance.

If any monitored keywords are paused—whether voluntarily to preserve monthly volume or automatically following a plan downgrade—an amber reactive alert banner surfaces immediately below the header. The banner highlights how many keywords are currently paused, clarifies that paused terms do not ingest live mentions, and provides a direct one-click link to upgrade keyword quotas or adjust billing.

For workspaces establishing new tracking antennas, the Competitor Onboarding Prompt evaluates your keyword roster. If your project lacks a designated primary brand, it prompts you to crown your main brand anchor to unlock Share of Voice calculations and executive benchmarks. If a primary brand exists but no competitor terms are tracked, it suggests adding 1 to 4 rival brands to activate competitive positioning metrics.

A responsive data toolbar above the table provides instant debounced search across all keyword terms, alongside faceted multi-select dropdowns to filter rows by monitoring status (Active vs. Paused), brand role (My Brand, Competitor, Neutral), and entity category (Brand, Person, Geographical, Other).

  • Live quota counter ({activeCount}/{allowedKeywords}) displays active monitoring capacity in real time.
  • Reactive paused alert banner flags dormant keywords with one-click quota expansion links.
  • Competitor Onboarding Prompt guides teams to set primary brand anchors and rival benchmarks.
  • Faceted data table toolbar filters rows dynamically by keyword name, operational status, brand role, and entity category.

Keywords Table Columns & Live Intelligence

The monitored keywords data table displays comprehensive live metrics for each tracking antenna. Clicking any row navigates directly to the live Mentions Feed filtered by that specific keyword, allowing rapid inspection of raw mentions.

The table organizes information across 8 specialized columns designed for executive scanning and operational triage:

  • Target Keyword & Status: Glowing emerald status dot for active monitoring (zinc when paused), keyword text with hover copy button, and metadata badges indicating Brand Role (MY BRAND, COMPETITOR), Target Type (BRAND, PERSON, GEOGRAPHICAL, OTHER), and Match Method (EXACT PHRASE, ALL WORDS, ANY WORD).
  • Mentions & 7-Day Velocity: Total ingested mentions count with hover tooltips displaying exact numbers, unread mentions indicator pill, and a 7-day trend badge showing positive or negative percentage velocity compared to the preceding 7-day period.
  • Reach & Share of Voice (SOV): Total potential audience impressions summed from verified publisher metrics, paired with a gradient micro-progress bar displaying the keyword's mathematical Share of Voice as a percentage of overall project conversation.
  • Net Sentiment Score & Tri-Tone Capsule: Calculated as ((Positive - Negative) / Total) * 100 on a -100 to +100 scale. Features an inline tri-tone multi-segment capsule bar (emerald for positive, zinc for neutral, rose for negative) and hover cards detailing exact category counts and AI Brand Safety ratings (100 - toxicityRate)% Safe.
  • Social Presence Channels: Proportional breakdown of chatter originating from social networks versus editorial media, featuring automated weak social warning triggers.
  • Websites Count: Aggregated count of unique publisher domains, news outlets, and portals discussing the keyword.
  • Updated Smart Date: Relative timestamp of the latest recorded mention, displaying an amber warning icon if data has not refreshed within 24 hours.
  • Row Actions Menu: Contextual options to browse mentions, open Query Lab, launch the Channel Footprint drawer, edit keyword parameters, toggle active status, adjust brand roles, or permanently delete the keyword.

Social Presence Cell & Weak Social Detection

The Social Presence column evaluates how consumer discussions and editorial coverage are distributed for each monitored term. In modern reputation management, brand chatter rarely distributes evenly across media types: some enterprise B2B brands dominate trade news while remaining virtually absent from consumer social feeds, whereas direct-to-consumer lifestyle brands often experience the inverse.

ElixBrand continuously computes the proportion of total mentions originating from supported social platforms (X/Twitter, Reddit, YouTube, LinkedIn, Instagram, Threads, Bluesky, Mastodon, TikTok) versus traditional editorial web publications.

If a monitored keyword exhibits dormant consumer chatter—specifically triggered when social discussions represent 8% or less of volume across 5 or more total mentions, or when social presence is entirely dormant—the cell displays a distinct amber warning badge labeled 'Weak Social' with an AlertTriangle icon.

Hovering over the warning reveals a diagnostic tooltip detailing the exact balance between social networks and editorial publications. Clicking the cell dispatches an event that instantly slides open the Channel Footprint Drawer for comprehensive media distribution analysis.

  • Automated 'Weak Social' warning triggers when social mentions comprise 8% or less of total volume.
  • Distinguishes between viral consumer buzz and institutional press coverage.
  • Interactive hover cards reveal exact platform mention splits across Facebook, X, LinkedIn, and YouTube.
  • Single-click trigger directly launches the slide-over Channel Footprint Drawer.

Channel Footprint Drawer & Media Archetypes

The Channel Footprint Drawer is a slide-over analytical sheet accessible from the table's social cell or row actions menu. It delivers an architectural breakdown of how your brand or keyword resonates across global media channels.

The header presents high-level volume metrics—Total Mentions, Potential Reach, and Share of Voice—before classifying the keyword into one of four intelligent media archetypes:

  • Institutional / Press-Centric Footprint: Assigned to public agencies, B2B enterprises, industrial organizations, or legal entities where consumer social chatter is negligible and coverage is concentrated in national press, trade publications, and official gazettes.
  • Social-Media Heavy Footprint: Assigned to consumer apps, lifestyle products, creator brands, or viral campaigns where discourse thrives on X, Reddit, TikTok, and YouTube with minimal formal press releases.
  • Multi-Channel Balanced Footprint: The optimal corporate profile, reflecting healthy, concurrent presence across editorial journalism and active community discussions.
  • Awaiting Mentions Ingestion: A transient diagnostic state displayed for newly created keywords while initial crawler streams collect and index baseline content.
  • Proportional Media Distribution: A 4-segment comparative bar breaking down volume across News (blue), Web Portals & Blogs (sky), Social Networks (purple), and Other Channels (zinc).
  • Strategic Recommendations: Contextual guidance advising next steps—such as configuring verified brand handles as Custom Sources when public consumer chatter is low, or applying Boolean homonym filters when social volume skews generic.

Brand Roles & Competitive Benchmarking

Every monitored keyword is assigned a functional Brand Role that dictates how ElixBrand calculates comparative market intelligence and executive benchmarks. Rather than treating all keywords as an undifferentiated list of search terms, assigning roles establishes a structured hierarchy between your own brand, your competitors, and broader industry themes.

The three available Brand Roles are:

  • My Brand (primary): Identifies the primary corporate or brand identity of your project. Each workspace is allowed exactly one primary brand keyword. Designating a primary brand serves as the anchor for Share of Voice calculations, head-to-head sentiment gap analysis, and executive briefing summaries.
  • Competitor (competitor): Identifies direct or indirect market rivals. Tracking competitor keywords allows marketing and PR teams to benchmark visibility, monitor competitor crisis moments, track product announcements, and compare net sentiment scores side-by-side.
  • General Topic (neutral): The default role assigned to industry themes, regulatory terms, event hashtags, customer pain points, or general market trends where ownership attribution is unnecessary.
  • Keyword Target Types: In addition to brand roles, keywords are classified into four Target Types: Brand / Org (companies and commercial entities), Person / Figure (CEOs, executives, creators, politicians), Geographical (countries, cities, regions), and General / Other (hashtags, campaigns, products).

Keyword Creation Wizard (/keywords/add) & Unified Architecture

Creating a new monitored keyword is handled through the unified Keyword Details workspace (/dashboard/projects/[id]/keywords/add). This streamlined interface eliminates disjointed multi-step modals in favor of a single, responsive configuration center equipped with live AI diagnostics, real-time volume estimates, and search query validation.

The setup interface features two distinct operating modes selected via the mode tabs in the header:

  • Rule Builder (simple): The recommended visual interface combining intuitive form controls with real-time AI guidance. It provides direct access to primary terms, match modes, aliases, required context anchors, negative exclusions, platform filters, languages, and geographic targeting in a single organized canvas.
  • AI Assistant (ai): A natural-language prompt interface available on Pro, Business, and Enterprise plans. It allows teams to describe their tracking requirements in plain language and automatically converts those instructions into calibrated keyword parameters.
  • URL Parameter Auto-Configuration: The creation page automatically parses incoming URL parameters (?keyword=...&type=...&brandRole=...), making it seamless to create new keywords directly from entity clouds, mentions cards, or onboarding prompts.
  • Live Search Query Preview: A dedicated Query Preview button (<Code2 /> icon) in the header enables technical users and admins to inspect the exact compiled search request payload before committing changes.

Natural Language AI Assistant Mode

For teams managing complex enterprise monitoring requirements, the AI Assistant mode allows you to generate robust keyword configurations using conversational prompts. Instead of manually assembling dozens of inclusion, exclusion, and platform rules, you simply describe what you wish to track and what noise to avoid.

For example, entering a prompt such as 'Track Acme Cloud and products like Acme Server, exclude financial earnings reports and investor filings, and restrict monitoring to English and French on Twitter and TechCrunch' triggers an instant AI translation process.

The assistant analyzes the prompt, resolves named entities, identifies industry categories, constructs contextual keyword groups, and maps out appropriate platform and language filters. The generated parameters are automatically injected into the form fields, allowing you to review the visual configuration in the Rule Builder or activate monitoring immediately with a single click.

  • Prompt-to-Configuration Engine: Converts natural language descriptions into complete, production-ready keyword parameters.
  • Automated Parameter Extraction: Automatically maps primary terms, aliases, match types, negative exclusions, platforms, and languages.
  • Pre-Configured Prompt Templates: Provides starter prompts for executive tracking, brand defense, competitor intelligence, and crisis management.
  • Plan Gated: Accessible to Pro, Business, and Enterprise workspaces, featuring a distinct PRO badge in the interface.

Match Modes & Search Matching Logic

The accuracy of your brand antenna depends heavily on selecting the correct Matching Mode. ElixBrand offers four precision matching configurations that dictate how our search engine tokenizes and evaluates incoming text:

  • Smart Matching (smart): The default, balanced configuration. Employs intelligent multi-match algorithms with automatic fuzziness (fuzziness: 'AUTO'). It captures natural linguistic inflections, common typos, and grammatical variations while maintaining strict relevance boundaries.
  • Exact Phrase Only (exact_phrase): Enforces zero fuzziness (fuzziness: '0') and requires words to appear in the exact, uninterrupted sequence specified. Ideal for multi-word brand names, legal entities, or distinct marketing slogans where word order is non-negotiable.
  • All Words (all_words): Requires every word in your keyword phrase to appear within the document, regardless of order or proximity. Useful for capturing references where verbs, prepositions, or adjectives separate your core brand terms.
  • Any Words (any_words): A broad catch-all configuration that triggers a match if any single word from the phrase is detected. Recommended primarily for broad topic monitoring or exploratory research, and should be paired with strong negative exclusions to prevent noise.
  • Word Proximity Slop: For phrase matching, word proximity controls how many intermediate words can appear between your target terms while still qualifying as a verified match.

Precision Context (AND) & Exclusion (NOT) Rules

Brand names that double as common dictionary nouns, colors, geographical locations, or popular surnames face high risks of irrelevant mentions. If your brand is named 'Orange', 'Apple', 'Target', or 'Mercury', tracking the bare term alone will flood your workspace with consumer recipes, retail catalogs, and astronomy articles.

ElixBrand resolves this challenge through structured Boolean context gates and negative exclusion boundaries:

  • Required Context Words (Boolean AND): Words that MUST appear in the article or post alongside your primary keyword for a mention to be captured. When multiple context words are configured (e.g. 'telecom', '5G', 'fiber', 'recharge'), ElixBrand groups them into a contextual gate requiring at least one anchor term to be present. This cleanly eliminates dictionary false positives.
  • Excluded Words (Boolean NOT): Words, phrases, or topics that immediately disqualify a mention from being ingested. Any post containing an excluded term is dropped at the ingestion gateway before entering your feed or consuming monthly quotas.
  • Multi-Word Exclusion Safeguards: Excluded terms containing multiple words or hyphens (e.g. 'fruit juice', 'apple pie') are automatically compiled as exact phrase matches, ensuring that ordinary single-word occurrences of 'fruit' or 'pie' do not accidentally suppress genuine brand mentions.
  • Zero-Noise Guarantee: Properly pairing primary brand terms with 3 to 5 required context anchors delivers near-100% signal precision without missing authentic customer feedback.

Aliases, Synonyms & Multi-Script Coverage

Customers, journalists, and social media users frequently refer to organizations using colloquial nicknames, stock tickers, social handles, or native-script transliterations. To ensure total visibility, ElixBrand allows you to register comprehensive Aliases under a single unified keyword antenna.

Aliases are evaluated under a Boolean OR relationship with the primary keyword. If a document matches the primary term OR any configured alias (while respecting required context anchors and negative exclusions), it is captured into your feed.

For global and multilingual brands, aliases provide essential cross-alphabet coverage. For example, a Moroccan enterprise can configure 'Maroc Telecom' as its primary term, while adding 'اتصالات المغرب' (Arabic script), 'IAM' (abbreviation), and '@maroctelecom' (social handle) as active aliases.

The 'Allow Typos' toggle enables intelligent character transposition tolerance, capturing common misspellings without requiring you to manually list dozens of typo variations.

  • Unified Alias Architecture: Consolidates official names, stock symbols, product nicknames, and social handles into one tracking antenna.
  • Multi-Script Ingestion: Seamlessly combines Latin, Arabic, Cyrillic, Greek, Hebrew, and CJK alphabets within a single keyword rule.
  • Quota Preservation: Aliases do not consume additional keyword quotas; all configured variations count as a single monitored antenna.
  • Typo Tolerance Control: Toggle between fuzzy matching (capturing misspellings automatically) and strict character validation.

Geographic & Language Targeting

Monitoring global discourse often requires geographic and linguistic precision. A regional retailer operating exclusively in North America has little interest in homonym mentions originating from Europe or Southeast Asia, while a luxury fashion house may require worldwide multi-language coverage.

ElixBrand provides granular targeting controls integrated directly into the keyword configuration workspace:

  • Language Filtering: Select from 11 core international languages (English, Spanish, French, German, Italian, Portuguese, Russian, Chinese, Japanese, Korean, Arabic) or specify any of 240+ international ISO language codes. Leaving the language selector empty monitors all languages globally.
  • Country / Territorial Targeting: Filter mentions across 240+ countries and territories using localized names and two-letter ISO country codes. Ingestion filters cross-reference publisher top-level domains, server IP geography, and author metadata to verify geographic origin.
  • Global Default Behavior: If no specific languages or countries are selected, ElixBrand operates in an unrestricted global monitoring state, capturing worldwide mentions as long as keyword matching conditions are satisfied.

Media Source & Channel Routing

Not all tracking antennas serve the same business purpose. While a PR communications team may focus strictly on editorial newspapers, trade magazines, and wire services, a social customer care team may focus entirely on Reddit, X, and community forums.

ElixBrand splits media sources into two parallel selector matrices, allowing independent filtering across Web Publications and Social Networks:

  • Web & Editorial Media: Toggle independent coverage for News Media (journalistic publications and syndicated wires), Blogs & Articles (industry blogs and Substack newsletters), Discussion Forums (community boards), General Web Content (corporate sites and portals), Customer Reviews (consumer review platforms), Podcasts & Audio (transcribed audio streams), and AI Search Engines (generative engine citations).
  • Social Media Networks: Toggle independent coverage for X / Twitter, Reddit, Bluesky, YouTube, LinkedIn, Mastodon, Threads, and TikTok.
  • Effective Routing Compilation: Selecting 'All' or leaving either matrix unselected defaults to unrestricted monitoring across that category. Selecting specific platforms restricts ingestion exclusively to verified content from those channels.

Sentiment Focus & Priority Alerting

In high-volume workspaces, capturing every neutral press release or minor brand reference can overwhelm operational teams. The Sentiment Focus selector enables you to establish specialized, tone-gated tracking antennas.

Available sentiment focus options include:

  • All Mentions (all): The standard monitoring setting. Ingests positive advocacy, objective reporting, factual updates, and negative criticism indiscriminately.
  • Positive Only (positive): Restricts ingestion strictly to praise, recommendations, advocacy, and positive sentiment. Ideal for influencer discovery, customer testimonial curation, and brand champion tracking.
  • Negative Only (negative): Restricts ingestion exclusively to complaints, product friction, negative reviews, and crisis discussions. Essential for PR defense, risk monitoring, and automated escalation webhooks.
  • Neutral Only (neutral): Captures factual news reporting, industry roundups, financial filings, and objective media mentions without emotional bias.
  • Automated Label Expansion: When filtering by positive sentiment, the system automatically includes both 'positive' and 'strongly positive' classifications; filtering by negative sentiment includes both 'negative' and 'strongly negative' classifications.

Ambient AI Diagnostics & 4-Dimension Quality Scoring

As you enter or modify your primary keyword in the configuration workspace, ElixBrand's real-time AI diagnostic engine activates in the background with an 1800ms debounce. Grounded in live 30-day search index data and actual matching snippets from our live intelligence feed, the diagnostic engine calculates an objective Signal Quality Score (0 to 100).

The scoring engine evaluates your rule across a calibrated 4-Dimension Rubric (0 to 25 points each):

  • Multilingual Coverage (0–25 points): Assesses multi-script completeness, international transliterations, and native alphabet coverage (e.g. verifying Arabic transliterations for Middle Eastern brands).
  • Disambiguation (0–25 points): Measures protection against dictionary homonyms, cross-industry collisions, and generic word confusion.
  • Noise Boundaries (0–25 points): Evaluates the strength and specificity of negative exclusions in filtering spam clusters and non-brand meanings.
  • Recall Safety (0–25 points): Ensures required context terms and match modes are not overly restrictive, protecting against false negatives and missed mentions.
  • Signal Quality Tiers: Scores of 80–100 achieve the Optimal Signal tier (emerald badge with ShieldCheck), scores of 60–79 fall into the Moderate tier (amber badge with AlertTriangle), and scores below 60 indicate an Ambiguous Query (rose badge with ShieldAlert) requiring immediate context refinement.
  • Ambient Entity Badge: Renders directly beneath the primary keyword input, displaying the detected entity category, industry vertical, regional market, 30-day volume estimate, and ambiguity alerts.

Inline AI Recommendations & Query Preview Modal

In addition to scoring signal quality, the ambient AI engine generates actionable, one-click recommendations directly inside the configuration form. Below the Aliases, Required Context Words, Excluded Words, Languages, Countries, and Sources fields, interactive recommendation pills surface suggested additions tailored to your brand.

You can click individual suggestion pills to add specific terms or click 'Add All' to accept comprehensive recommendations instantly. Dynamic Recommendation Chips also propose strategic changes—such as advising you to switch from Smart Matching to Exact Phrase when tracking multi-word trademarks, or recommending that you assign a Competitor role when a rival brand is detected.

For developers, security officers, and advanced PR analysts, the Query Preview Modal (<Code2 /> icon) provides full visibility into the compiled search DSL request payload before saving. You can inspect the exact boolean query, verify tokenization logic, test rule execution against recent media, and copy the formatted JSON request body.

  • Interactive Suggestion Pills: One-click addition of recommended aliases, required context anchors, and noise exclusions.
  • Strategic Recommendation Chips: Context-aware advice proposing matching mode refinements and brand role assignments.
  • Search DSL Inspection: Real-time Query Preview modal displaying the full JSON search request payload.
  • Zero-Risk Verification: Test and inspect query mechanics before committing rules or consuming monthly ingestion quotas.

Keyword Lifecycle: Pause, Resume, Edit & Delete

Managing keyword antennas involves active lifecycle governance as marketing campaigns launch, evolve, and conclude. In ElixBrand, keywords can be updated, temporarily paused, resumed, or permanently deleted depending on operational needs.

Pausing a keyword is non-destructive: all previously ingested mentions, historical analytics, sentiment distributions, and Share of Voice metrics remain fully preserved in your workspace. Pausing simply suspends live crawler polling, ensuring dormant campaigns do not consume monthly mention quotas.

When resuming a paused keyword, the system executes an automated quota check (assertCanActivate). If your workspace has reached its plan keyword limit, the resume action is blocked with an informative notification advising you to pause another keyword or upgrade your plan.

If a project subscription is downgraded or expires, ElixBrand enforces Zero Data Loss: excess keywords beyond the new plan tier are automatically paused in order of creation date (oldest keywords remain active), ensuring your historical data is never deleted.

Deleting a keyword permanently removes the tracking rule along with its associated mention records and analytics. This action requires confirmation and is restricted to project Owners, Admins, and Editors.

  • Non-Destructive Pausing: Halts live ingestion while preserving 100% of historical mentions, metrics, and report archives.
  • Quota-Gated Resumption: Validates available keyword quotas before reactivating paused antennas.
  • Zero Data Loss Downgrades: Excess keywords are gracefully paused rather than destroyed when plan limits decrease.
  • Permanent Deletion Protection: Deletion requires explicit confirmation and automatically cleans up associated database references.

Historical Backfill & Ingestion Pipeline Mechanics

When a new keyword is registered or an existing keyword is resumed, ElixBrand activates both a real-time listening stream and an automated historical backfill process. Understanding how these pipelines operate ensures accurate data interpretation during onboarding.

The historical backfill depth is determined strictly by your workspace subscription tier:

  • Individual Plan ($29/mo): Backfills and retains mentions for the past 30 days.
  • Pro Plan ($99/mo): Backfills and retains mentions for the past 90 days (~3 months).
  • Business Plan ($299/mo): Backfills and retains mentions for the past 365 days (1 full year).
  • Enterprise Plan (Custom): Backfills and retains mentions for up to 730 days (2 full years).
  • Dual Pipeline Activation: Keyword creation emits an immediate live fetch job to query external social networks and news APIs, alongside an AI scan that searches existing indexed media to enrich matching content.
  • Alert Storm Prevention: To prevent webhook and notification floods when backfilling months of historical data, automated instant alerts enforce a strict 7-day lookback limit (published_at >= now - 7d). Mentions published more than 7 days ago will populate your feed and reports but will not trigger real-time alert notifications.

Role Permissions & Team Collaboration

Workspace security and governance in ElixBrand follow strict Role-Based Access Control (RBAC). Permissions to view, create, edit, pause, and delete monitored keywords are enforced consistently across both API endpoints and the user interface.

Project roles are organized into a strict hierarchical permission ladder:

  • Owner (Rank 100) & Admin (Rank 80): Full administrative authority. Can create, edit, pause, resume, reassign brand roles, and delete any monitored keyword.
  • Editor (Rank 60): Full operational authority. Can create new keywords, edit search parameters, pause/resume monitoring, and delete keywords.
  • Analyst (Rank 40): Triage authority. Can view the keywords overview, inspect metrics, launch the Channel Footprint drawer, and browse mentions. Cannot create, edit, pause, or delete keywords (form controls are disabled).
  • Viewer (Rank 20) & Client (Rank 10): Read-only access. Can inspect keyword performance, view charts, and browse mentions. All modification, creation, and deletion actions are hidden in the UI and rejected with HTTP 403 Forbidden at the API layer.
  • Quota Attribution Integrity: Keyword quotas are always evaluated against the workspace Owner's subscription allowance, preventing quota leakage across invited team members.

Audited Metrics & Mathematical Formulas

To ensure radical transparency and analytical precision, every metric displayed across the Keywords Overview table and Channel Footprint drawer is governed by standardized mathematical formulas. PR strategists and marketing leaders can rely on these definitions when presenting data to executive stakeholders:

Active Keyword Quota Utilization0 to Plan Limit (e.g. 3, 15, 50, 500+)

The count of actively monitored keywords (isActive: true) owned by the workspace compared to the maximum allowed keyword quota of the subscription tier.

Pro tip: Pause seasonal campaign keywords when their active window ends to free up quota capacity for ongoing brand monitoring.

7-Day Volume Trend VelocityPercentage (+X% or -X%, or flat)

Compares total mention volume ingested over the preceding 7 days against the baseline volume of the 7-day period immediately before that (now-7d to now vs. now-14d to now-7d).

Pro tip: A sudden velocity surge (>+50%) paired with dropping net sentiment is the primary early warning indicator of an emerging PR crisis.

Share of Voice (SOV %)0.0% to 100.0%

The mathematical proportion of total project mentions generated by a specific keyword, calculated as: (Keyword Mentions / Total Project Mentions) * 100.

Pro tip: Track SOV trends between your Primary Brand keyword and Competitor keywords to quantify market presence gains over quarterly campaigns.

Net Sentiment Score-100 to +100

The standardized balance of customer and media sentiment, calculated as: ((Positive Mentions - Negative Mentions) / Total Mentions) * 100.

Pro tip: Scores above +20 indicate strong brand equity; scores falling below -10 warrant immediate operational review and narrative triage.

AI Brand Safety Index0.0% to 100.0% Safe

Quantifies the absence of toxic or hostile content, calculated as: Math.max(0, 100 - toxicityRate)%, where toxicityRate is the percentage of mentions flagged by AI toxicity classifiers.

Pro tip: Enterprise brands should aim for a Brand Safety Index of 98.0% or higher. Drops below 95.0% indicate hostile comment brigade activity.

Social Presence Ratio0% to 100%

The percentage of total mentions originating from supported consumer social networks (X, Reddit, YouTube, LinkedIn, Instagram, TikTok, Threads, Bluesky, Mastodon) versus editorial press.

Pro tip: If your Social Presence Ratio falls to 8% or lower, ElixBrand triggers an amber 'Weak Social' alert advising you to track verified accounts as Custom Sources.

Publisher Outlets Diversity1 to Unlimited (integer)

The number of unique web domains and news publication sites that have published articles referencing the keyword, computed via cardinality aggregation on source domains.

Pro tip: High mention volume concentrated across very few unique websites indicates syndication duplication rather than broad media interest.

Signal Quality Score0 to 100 points

The composite score evaluated by the ambient AI diagnostic engine, summing four 25-point dimensions: Multilingual Coverage + Disambiguation + Noise Boundaries + Recall Safety.

Pro tip: Always aim for an Optimal Signal score (80+) before activating a keyword to ensure clean data ingestion and avoid wasting mention quotas.

Homonym Collision RiskLow, Medium, or High

AI diagnostic rating assessing whether a brand name conflicts with dictionary nouns, geographical locations, or popular personal names.

Pro tip: High-risk terms must always be configured with at least 2 to 3 Required Context Words (AND rules) to filter out dictionary homonyms.

Word Proximity Slop1 to 20 words (default: 10)

The maximum number of intervening words permitted between terms in an Exact Phrase or phrase-matching query while still registering as a valid hit.

Pro tip: Use slop 10 to allow natural journalistic prose (e.g. 'Acme, the global leader in cloud infrastructure, Corp') while preventing false connections across distant paragraphs.

Operational Playbooks & Real-World Frameworks

Deploying and managing monitored keywords requires disciplined execution. The following four playbooks offer battle-tested frameworks for setting up primary brands, benchmarking competitors, tracking executives, and remediating high-noise tracking rules.

1

Setting Up a Bulletproof Primary Brand Antenna (Under 10 Minutes)

Navigate to /keywords/add and enter your official brand name as the Primary Term in Rule Builder mode. Assign the Brand Role to 'My Brand' (primary) and verify the Target Type is 'Brand / Org'. Add native script transliterations, colloquial abbreviations, and official social handles into the Aliases field. Review the Ambient Entity Badge: if homonym collision risk is Medium or High, add 3 to 5 core industry anchors (e.g. 'software', 'cloud', 'security') into Required Context Words. Inspect the Signal Quality Score to confirm it reaches 80+ (Optimal), then click 'Start Monitoring'.

Deliverable:A calibrated primary brand antenna delivering zero-noise mentions, accurate Share of Voice calculations, and executive sentiment benchmarks.
2

Configuring Competitor Benchmark Keywords for Share of Voice

Identify 2 to 4 key market competitors. For each competitor, create a new keyword setting Brand Role to 'Competitor' and adding their official corporate website in Competitor Details. Select Smart Matching or Exact Phrase depending on name complexity, and configure their major product line names as Aliases. Keep language and geographic targeting identical to your Primary Brand antenna to ensure statistically valid, apples-to-apples Share of Voice and sentiment comparisons.

Deliverable:Automated head-to-head competitive benchmarking across media volume, audience reach, net sentiment gap, and channel footprint.
3

Setting Up a Targeted Executive & Spokesperson Antenna

Create a new keyword using the executive's full name as the Primary Term. Set Target Type to 'Person / Figure' and Brand Role to 'My Brand' or 'Neutral'. Select Exact Phrase matching to prevent false matches on individual first or last names. In Required Context Words, add your company name and executive titles (e.g. 'Acme', 'CEO', 'Founder', 'CTO'). In Excluded Words, add unrelated namesakes, sports figures, or politicians sharing the same name.

Deliverable:High-precision thought leadership tracking and PR visibility monitoring for key corporate leadership.
4

Remediating High-Noise or Dormant Monitoring Rules

Inspect the Keywords Overview table for keywords with unusually high volume but low sentiment relevance, or those flagged with 'Weak Social' warnings. Click row actions and open Query Lab or Channel Footprint. If volume is inflated by dictionary homonyms, edit the keyword and add 2 to 4 mandatory context anchors into Required Context Words. If consumer social is dormant, verify whether your target audience discusses the brand on specialized forums or LinkedIn, and configure verified corporate channels as Custom Sources.

Deliverable:Restored signal purity, eliminated false-positive noise, and optimized monthly mention quota consumption.

Frequently Asked Questions

What happens when my workspace reaches its keyword quota? When active keywords reach your plan limit (e.g. 3 on Individual, 15 on Pro, 50 on Business), creating new keywords or resuming paused keywords is blocked. You can pause an existing keyword or upgrade your subscription from Billing to expand capacity immediately.

Does pausing a keyword delete its captured mentions or historical charts? No. Pausing is completely non-destructive: all previously ingested mentions, sentiment analytics, and report metrics remain fully preserved. Only live crawler polling is suspended.

How far back does historical backfill reach when I create a new keyword? Historical reach matches your plan's data retention window: 30 days on Individual, 90 days on Pro, 365 days (1 full year) on Business, and 730 days (2 full years) on Enterprise.

Will creating a keyword trigger an alert flood for older mentions? No. Instant alerts enforce a strict 7-day lookback limit (published_at >= now - 7d). Backfilled historical content older than 7 days will populate your feed and reports but will never trigger notification webhooks or alert emails.

Can I track brand names in non-Latin scripts like Arabic, Cyrillic, or Japanese? Yes. ElixBrand fully supports multi-script Unicode characters across primary terms, aliases, and context anchors without character corruption.

What is the difference between Smart Matching and Exact Phrase Only? Smart Matching allows fuzzy matching to capture typos and natural inflectional variations. Exact Phrase enforces zero fuzziness and requires words to appear in the exact specified order.

Why does my brand have a 'Weak Social' warning in the table? The alert triggers when social network mentions account for 8% or less of total volume across 5 or more mentions. It indicates discourse is concentrated in editorial news rather than consumer social channels.

Who on my team can create, edit, pause, or delete keywords? Project Owners, Admins, and Editors have full management rights. Analysts, Viewers, and Clients have read-only access and cannot modify or delete keywords.

What happens to my keywords if I downgrade my plan? ElixBrand enforces Zero Data Loss. Keywords beyond your new plan quota are automatically paused in order of creation date (oldest remain active) rather than deleted.

How many aliases can I add to a single keyword? You can add up to 30 aliases per keyword. All aliases are evaluated under a Boolean OR relationship and count as a single monitored antenna against your plan quota.

Can I test my search query before saving it? Yes. Clicking the Query Preview button (<Code2 /> icon) in the header opens a modal displaying the exact compiled search request payload and percolator matching structure.

Can I restrict a keyword to capture only negative mentions? Yes. In the Sentiment Focus selector, select 'Negative Only' to restrict ingestion strictly to complaints, friction, and crisis discussions.

Track your brand and competitors with precision antennas

Eliminate dictionary homonyms with smart context rules, benchmark market share against rivals, audit signal quality with real-time AI diagnostics, and monitor web, press, and social channels seamlessly.