Methodology & honesty

How a Crowdproof study works, and where to doubt it.

Crowdproof produces simulated evidence. This page is the contract for what that evidence is, how it is made, and what it can and cannot tell you. Inspect cited evidence separately from modeled findings, and use the stated assumptions and limitations when comparing options. Simulations are not validated predictions of real-market outcomes.

simulation scenarios · explicit limitations
Provenance

Every artifact is labeled with what it is.

Nothing in Crowdproof pretends to be a human. Provenance labels are printed on each artifact rather than hidden in a footnote.

simulatedsimulated Interviews, modeled audience responses, and scenario outcomes are generated by AI. Detailed simulated conversations remain available in the interview workspace.
cited evidencecited External market context linked to its public source. Simulated conclusions remain distinct from observed evidence.
livelive A run in progress. Coral is reserved for running simulations; once a run completes, its outputs revert to the labels above.
The society

An audience scoped to the approved study.

The audience is modeled. Its country scope and segments follow the approved research design and the current source capabilities. Demographic provenance and scenario assumptions remain separate.

sampling capabilityThe current sampling source documents country, age-range, and gender filters. Other desired attributes cannot be presented as verified sampling constraints without supporting capability and execution evidence.
cohortsReview the proposed countries, segment definitions, supported filters, and scenario assumptions before approving a study. Audience size follows the approved design and available source capacity; an unsupported requirement is disclosed rather than silently treated as satisfied.
diffusionFor launch and channel-mix questions, propagation scenarios model reach across simulated social channels. Their audience scale and channel assumptions describe a scenario, not the number of recruited participants or verified future customers.
limitsA modeled audience is not a recruited panel or proof of representative real-market behavior. Treat segment-level findings as modeled hypotheses and check them against observed behavior in your market.
Evidence

Facts and simulations stay separate.

Crowdproof combines cited public market evidence with modeled audience research, structured simulated interviews, and scenario analysis. Reports label each evidence class so a simulation cannot masquerade as observed customer behavior.

market factsPublic market and industry claims retain their sources so readers can inspect the evidence behind the report. A citation does not certify publisher accuracy, recency, independence, or relevance to the study's market.
simulated evidenceAudience reactions and interviews are explicitly modeled outputs, not observations from recruited participants.
decision boundaryThe final report names limitations, uncertainty, counter-signals, and the real post-launch metrics that should calibrate or overturn its recommendation.
Interviews

Stratified, analyzed, decision-ready.

Synthetic interviews are distributed across the available audience segments so loud cohorts do not drown out quiet ones. The report presents cross-interview themes, tensions, and decision implications instead of burying the user in raw transcripts.

samplingStratified across the materialized audience segments. A study can interview up to 50 representatives, but the count is a maximum rather than a quota: the system stops at the available decision-relevant segments.
transcriptsThe management report contains the interview analysis, not question-by-question answers or raw quotes. Detailed simulated conversations remain available in the interview workspace when a reviewer needs to inspect one persona.
churn reasonsNeeds, barriers, and choice drivers are ranked by segment and translated into product, pricing, positioning, and promotion implications. Check those rankings against observed conversion, retention, and business outcomes after launch.
Seeds & uncertainty

The range is the result.

Propagation scenarios execute as multiple stochastic cascades. P10–P90 shows the outcome spread within the model's assumptions; it is not a validated confidence interval for real-market behavior.

multi-seedIndependent cascades expose variation across timing, network paths, and initial conditions. They do not establish the model's predictive accuracy or cover every source of real-world uncertainty.
reading a band“Retained 18.0% (P10 14.2 – P90 22.4)” is an assumption-bound scenario band. Choose the best-supported option, then plan safeguards against the downside edge rather than treating the midpoint as a promise.
wide rangesA wide range calls for closer review of assumptions and a smaller, measurable next step. Compare the recommendation with pilot results or observed business outcomes before expanding the rollout.
Illustrative simulation example · retained % · 9 seeds
10%18%26%

reported: 18.0% (P10 14.2 – P90 22.4)

Calibration

Compare scenarios, then check real outcomes.

Crowdproof's simulations are not validated predictions of real-market outcomes. No public calibration report is linked on this site, so no backtest accuracy or quarterly publication record is claimed here.

what a scenario supportsA simulation helps compare options under stated assumptions. Funnel shapes, leak points, and segment rankings remain modeled hypotheses until checked against observed outcomes.
how to check a recommendationCompare the study's recommendation with measured pilot or post-launch conversion, retention, cost, and revenue. Record where the result differs and revisit the assumptions behind the decision.
study contextInspect the report's sources, audience composition, simulation assumptions, and limitations. A report label or methodology description is not a certificate of predictive accuracy.
What we don't claim

The honest edges of the instrument.

no person-level predictionNo agent predicts what a specific human will do. The unit of meaning is the funnel and the segment, never the individual.
no guaranteesA favorable simulation is evidence, not a promise of success. Markets contain competitors, timing, and luck that neither simulated nor human panels can hold still.
no hidden uncertaintyRead modeled ranges together with their assumptions and limitations. Qualitative themes do not establish population prevalence, and a simulation range does not capture every source of market uncertainty.
no human cosplaySimulated interviews are labeled simulated. Their themes describe modeled responses, not observed reactions from recruited participants or proof of real-market demand.
Confidentiality

Your unreleased work stays yours.

Crowdproof provides conversation deletion and isolates tenant research data. Read the current privacy policy for the exact retention, subprocessors, model-training, and deletion boundaries.

questions about the method · simulation assumptions · evidence limits