Our Methodology

The methodology that makes
the output defensible.

Every claim is sourced. Every estimate is labeled. Every deliverable is built so your team can validate it with confidence.

Source-Tier Hierarchy

Not all sources are equal. We rank them.

Every data point in a Soundcheck deliverable is tagged with its source tier. Higher tiers are always preferred. Lower tiers are used only when higher-tier data is unavailable, and they are explicitly flagged.

T1

Government and official sources (preferred)

USDA, U.S. Census Bureau, BLS, SEC EDGAR, FDA, Eurostat, national statistical offices, central bank publications.

T2

Syndicated research

Euromonitor, IBISWorld, Statista, SimilarWeb, Nielsen, Mintel, GlobalData.

T3

Industry associations and regulatory filings

Trade bodies, industry associations, regulatory filings, patent databases, professional surveys.

T4

Web and media

News outlets, press releases, analyst commentary, web scraping, social-media intelligence.

Used only when T1 through T3 are unavailable. Always explicitly flagged in the deliverable.

Claim Typing and Confidence

Every claim is tagged. Nothing is presented without context.

Every statement in a Soundcheck deliverable carries two labels: a claim type (what kind of statement it is) and a confidence level (how strongly the evidence supports it). The reader always knows exactly what they are looking at.

HIGH
MEDIUM
LOW
FACT
U.S. sparkling-water retail volume grew 11.4% in FY 2024 (Euromonitor).
Industry association reports category CAGR at approximately 8% over the past three years.
A single trade-press article cites a regional growth figure without primary sourcing.
ESTIMATE
Total addressable market estimated at 14.2B liters based on Census and BLS consumption data.
Market share estimate derived from syndicated panel data with a 90-day lag.
Projected share based on proxy categories where direct data is unavailable.
INTERPRETATION
Strategic recommendation supported by converging T1 and T2 sources.
Competitive positioning inference grounded in public filings and press coverage.
Hypothesis requiring further validation through primary research.
Senior Validation

AI generates the research base. Your team makes it decision-grade.

The AI layer accelerates data collection, synthesis, and pattern recognition. Your team validates source quality, adjusts confidence labels, adds strategic context, and decides when the output meets the standard they would present to leadership.

01

AI research base

Data collection, source ranking, claim typing, and initial synthesis across hundreds of sources.

02

Team review

Your team validates sources, adjusts confidence labels, shapes the narrative, and layers on strategic interpretation.

03

Decision-grade output

A deliverable your team owns and can present at a board review, not an AI output that needs a disclaimer.

Proprietary Scores

Quantified readiness. Not gut feel.

Two proprietary scoring models convert qualitative research into a single, defensible number your team can track over time.

CVE Score

Coefficient of Viability of Expansion

0 to 100

Market Opportunity40%
Competitive Position35%
Internal Readiness25%

Used in: GoGlobal Viability Analysis

SRS Score

Scale-Readiness Score

0 to 100

Structural Foundation40%
Market Position35%
Scale Capacity25%
Verdict Bands
Scale-ReadyScale-ConditionalGrowth-OnlyRestructure

Used in: Scale Assessment

Focus Group Methodology

Structured qualitative rigor, not open-ended chatbot output.

Virtual Focus Groups follow established focus group research methodology adapted for synthetic audiences. Every session is moderated, structured, and analyzed with the same discipline your team would apply to a live panel.

Funnel-approach discussion

Sessions follow a structured funnel: engagement (category relationship), exploration (perceptions and unmet needs), concept probing (per-element reactions), comparison and prioritization, and exit reflections.

Neutral moderation principles

Open-ended, unbiased questions. Dominant voices are managed, quiet participants drawn out. The moderator never leads or shows agreement. Every surface reaction is probed for the “why.”

Group dynamics simulation

Personas interact as they would in a real group. Social influence, challenges, and validation shifts are simulated realistically and flagged as [GROUP DYNAMIC] in the analysis.

Information typing

Every reaction is tagged: [PERSONA-CONSISTENT] (traceable to profile traits), [INFERRED] (reasonable extrapolation), or [GROUP DYNAMIC] (emerged from interaction). Your team always knows what drove each finding.

Qualitative analysis

Thematic coding by frequency, intensity, and breadth. Cross-segment comparison between current buyers and aspirational targets. Sentiment mapping. Every finding paired with a strategic implication.

Structured report output

Detailed per-persona reactions, cross-concept insights, traceable preference votes (math matches actual panel size), attributed participant quotes, and actionable recommendations with confidence levels.

Every Virtual Focus Group report explicitly states that this is an AI simulation with synthetic personas. Findings are directional and should be validated with real consumer research before major investment decisions.

Assumptions and Gaps

Transparency is the credibility signal.

Every proxy, assumption, and data limitation in a Soundcheck deliverable is logged explicitly. Your team sees exactly where the evidence is strong and where it is not.

Source

Where the proxy or assumption originated.

Rationale

Why this proxy was selected over alternatives.

Confidence

How strongly the evidence supports the assumption.

Impact

What changes in the output if the assumption is wrong.

See the methodology on your category.

Talk to Sales →