Eight years of mixed-methods research. The methods I use, the frameworks I rely on, and the operational practice behind every study.
There is no best research method — only the right method for a specific question, audience, constraint, and decision the team is about to make. My job as a researcher is to pick the method that produces the most defensible evidence for the decision in front of us, within the time and budget we actually have.
That means I move comfortably between qualitative and quantitative approaches, between exploratory and evaluative work, between research that drives tactical product decisions and research that shapes multi-year strategy. The skill isn't knowing every method — it's knowing which one to choose and why.
Good research is useful research.
It earns its place in the roadmap.
When the question is about behavior, motivation, mental models, or unmet needs — depth matters more than scale.
The workhorse of my qualitative practice. Discussion guides tailored to the strategic question; semi-structured to allow follow-up; recorded and tagged for synthesis. Used at Honda (40+ SMEs), Intuit (10 small business owners), EDF (17 stakeholders across NA).
In-context studies where the question is about workflow — what people actually do, not just what they say they do. Particularly useful in B2B and enterprise contexts where workflow specificity is the whole story.
Capturing user experience over time. Onboarding → mid-trial → end-of-trial → post-conversion / churn touchpoints (Intuit Expert 365 Business). Reveals what changes — and what doesn't — across the customer journey.
Participants narrate decisions and errors while completing tasks. Surfaces the why behind performance data. Used in the WhatsApp study to diagnose "low predictability" as the through-line across six tasks.
When the question is about group dynamics, social acceptability, or generative ideation. Used selectively — not as a default, but when group conversation reveals something 1:1s can't.
Going to the user's environment when their environment is the whole context — homes, workplaces, retail spaces. Lower scale, much deeper insight.
When the question is about magnitude, distribution, or whether a change actually moved the needle.
Value proposition surveys, segmentation surveys, satisfaction tracking. Designed to be statistically defensible: appropriate sample sizes, balanced scales, careful attention to leading questions.
Task completion time, error rate, subjective satisfaction (1–10 scales), level of difficulty. Used in the WhatsApp study with randomized task order to control for learning effects.
When comparing two conditions (original vs. redesigned WhatsApp tasks), I run paired t-tests and report p-values. The discipline matters — a "significant improvement" without a p-value is just a feeling.
Partnering with data science to triangulate self-reported behavior against actual behavior. Identifying drop-off points, "aha" moments, and where stated preferences diverge from revealed preferences.
Designing tests with enough statistical power, defining success metrics up front, interpreting results without p-hacking. Particularly important for conversion-focused work like the Intuit landing page validation.
For emerging-tech research where traditional methods don't apply. Used at Honda for spatial computing research — capturing physiological responses to spatial interfaces.
Most strategic research questions need both depth and scale. The discipline is sequencing them right.
Quant surveys first, so qualitative discussion guides can be tailored to each respondent's context. Used at EDF to make 17 interviews far more efficient than a flat qual sample of the same size.
Pairing self-reported data (interviews, diary studies) with behavioral data (analytics, performance metrics). Reveals why users behave the way they do, not just what they do.
Concept → user feedback → refined concept → testing → refined concept. Used in the WhatsApp study (sketches → Figma prototypes → t-test validation) and Intuit (landing page variants → A/B testing roadmap).
Combining moderated interviews, diary check-ins, behavioral analytics, and expert observation across the same participants over time. Captures the customer journey end-to-end (Intuit Phase 2).
The research that informs multi-year decisions — and the workshop facilitation that turns it into action.
For future-state work where users don't yet exist (Honda 2030–2040), domain experts are the bridge. 40+ SME interviews across gaming, sci-fi, neurotech, and futurism — partnering with Meta, Marvel, Apple Vision Pro, USC, and Stanford.
Mapping what exists in adjacent categories to give the team a vocabulary for what they're building. Used to ground spatial computing research and validate decisions in tools migration work.
Synthesizing research findings into three time-horizons. Intuit's deliverable included Current → Target State (Feb 2026) → Future State (3-Year), giving stakeholders a shared model for both quarterly and long-term decisions.
Department-specific personas when business units have genuinely different needs (EDF: Asset Optimization, Corporate Finance, Grid-Scale Power). Personas as decision-making tools, not posters.
Multi-day workshops that turn research into prioritized roadmaps. Two-day strategic alignment sessions with Product, Design, Marketing, and Market Research (Intuit Phase 3) — including How Might We reframing and Target/Future state mapping.
Packaging findings for the audience that will act on them. Executive presentations to VP of Product / Marketing leadership; research reports calibrated to whether the reader is a designer, PM, engineer, or exec.
Modern UXR is more than moderation and synthesis. The operational craft is what separates research that ships from research that stalls.
Designing screeners to hit segmentation targets. Recruiting through UserInterviews, internal panels, and direct outreach. Coupon distribution & tracking.
Informed consent flows, NDA management, participant compensation, IRB awareness from academic work.
Multi-stakeholder timelines, cross-functional coordination, dependency management. Have led research alongside parallel design and engineering tracks.
Cross-system audits (EDF Tableau inventory; Lamps Plus replatform data governance across Syndigo, IBMi, SuperCMS).
Thematic analysis with color-coding systems. Tagging in Reduct. FigJam synthesis workshops. Affinity diagramming.
Word reports, PowerPoint decks, Excel workbooks, Figma-anchored design artifacts, workshop playbooks, executive briefings.
Frameworks are scaffolding — useful when they help you think, harmful when they replace thinking. Here are the ones I actually reach for.
When framing user needs as outcomes rather than features. Particularly useful for early-stage discovery work.
For reframing problems before solutioning. Used at Intuit (Reframe Value / Offer Flexibility / Prove the Magic) to open up the solution space.
For strategic work that needs both quarterly and multi-year alignment. Three time-horizons keep stakeholders on the same page.
Business criticality × technical complexity (EDF). Impact × effort. Reach × confidence. Simple but defensible.
For heuristic reviews and structured usability evaluation. The "low predictability" diagnosis in the WhatsApp study mapped to Nielsen's consistency / visibility heuristics.
For segmenting users by need rather than demographics. Used at Intuit to surface 5 distinct BGS needs across the small business owner segment.