Research rigor

For UX Researchers

Ship research at product speed while keeping the instrument, valid base, quality decisions, uncertainty, and reviewer context intact.

Study design

Usability evidence

Research quality

Create a research study Compare research modes

Illustrative product view—not a customer result

Available

A relevant product path

Design and check the respondent experience

For ux researchers, start with preference testing: Use MaxDiff or conjoint when the decision requires explicit trade-offs rather than independent ratings.

Open the matched decision guide

Problems this playbook solves

Each path starts with a concrete decision, names the method, and links to the deeper problem guide.

Preference testing

Problem

Rating grids often make every idea look important.

How to solve it

Use MaxDiff or conjoint when the decision requires explicit trade-offs rather than independent ratings.

MaxDiff

Conjoint

Open the decision guide

Product discovery

Problem

Surveys alone can miss the context behind behavior and unmet needs.

How to solve it

Pair structured evidence with human-led follow-up while preserving provenance and review notes.

Structured survey

Human follow-up

Open the decision guide

Research quality

Problem

Low-quality responses and thin segment bases can create false confidence.

How to solve it

Review readiness, response-quality signals, sample bases, and caveats before publishing findings.

Quality review

Base-size checks

Open the decision guide

Survey examples

Three surveys for the decisions ahead

Start with the decision, inspect the evidence available today, and keep the limits attached to every handoff.

Visual concept preference test

Compare visual directions consistently while retaining the reasons and contradictions behind preference.

Desktop seeded visual-comparison workflow. — Analysis · Captured
Synthetic example · not a customer result.

Public: Beta

Boundary: Stated preference does not prove usability.

Evidence gap: Local authenticated synthetic decision views are captured. Production deployment, customer outcomes, claim substantiation, and method-specific live certification remain unverified.

Preview preference test Read study evidence and limitations

How it works and evidence limits

Use this when

Design directions are compared inconsistently or without rationale.

What you do

Randomize visual stimuli and collect preference plus qualitative follow-up.

Randomized visual stimuli
Qualitative follow-up
Contradiction-aware comparison

Evidence returned

Distribution, segment differences, contradictions, and reasons.

Product: Beta

Evidence: Captured Synthetic Product Evidence

Illustrative product view using synthetic data. It is not a customer result or outcome claim.

Capabilities: survey-ab-testing, review-assisted-text-analysis

Activation-friction pulse

Catch onboarding friction at the moment it occurs and route consent-safe follow-up.

Public: Available

Boundary: Feedback does not establish causal activation lift.

Evidence gap: Local authenticated synthetic decision views are captured. Production deployment, customer outcomes, claim substantiation, and method-specific live certification remain unverified.

Preview friction pulse Read study evidence and limitations

How it works and evidence limits

Use this when

Onboarding pain appears in anecdotes after it has already cost users.

What you do

Collect low-friction moment feedback and route follow-up.

One-tap moment feedback
Driver and follow-up routing
Touchpoint trend context

Evidence returned

Touchpoint score, drivers, open text, and trend.

Product: Live

Evidence: Captured Synthetic Product Evidence

Illustrative product view using synthetic data. It is not a customer result or outcome claim.

Capture: Public Template Setup. Live execution is not certified.

Capabilities: basic-survey-logic-runtime, presentation-and-share-links

Response-quality review

Review questionable responses before they become findings or disappear without a trace.

Public response-quality workflow capture. — Analysis · Captured
Synthetic example · not a customer result.

Public: Available

Boundary: Quality signals reduce risk; they do not guarantee perfect detection.

Evidence gap: Authenticated response-level flag and disposition captures are missing.

Use quality template Read study evidence and limitations

How it works and evidence limits

Use this when

Invalid, duplicate, implausible, or thin responses become findings.

What you do

Configure checks and review flags before interpreting results.

Reviewable quality signals
Explicit response disposition
Valid-base caveat handoff

Evidence returned

Response-level signals, disposition, valid bases, and report caveats.

Product: Live

Evidence: Template Ready

Illustrative product view using synthetic data. It is not a customer result or outcome claim.

Capabilities: presentation-and-share-links, csv-excel-json-exports

Role-specific workflow

Before launch: document the decision, instrument, sample, stimuli, and quality thresholds.

Before sharing: connect every finding to its method, base, caveat, and recommended next step.

Decision-ready handoffs

Study brief

Survey instrument

Quality log

Analysis notes

Research report

Important context: Structured studies and reviewed analysis are live. Native human calling and autonomous AI moderation are not presented as generally available.