October 2020 ARVIC

Across nine sessions, the October 2020 ARVIC conference addressed a consistent set of pressures facing the insights function: organizations moving faster than traditional research can follow, AI reshaping how data is processed, and a global pandemic forcing teams to rebuild plans from scratch. Speakers responded not with calls to abandon established methods but with practical frameworks for making research more rigorous, more integrated, and more clearly tied to decisions. The talks span brand tracking, UX research, longitudinal consumer panels, hiring, interviewing craft, multi-method program design, and AI strategy, but the through-line is the same: research teams that operate with intention and communicate with discipline earn a seat at the table.

Running the Research Function as a Discipline

Several sessions turned the lens inward, treating the insights team itself as an object of study. One team applied the same outcome-driven framework they use for customers to their own operations, covering six phases of the research process and a set of cross-cutting enablers. The result was a shared language that scaled across org structures and immediately generated three workstreams: research quality measurement, a structured interview question bank, and a team skills self-assessment. The self-assessment, covering competence, motivation, and confidence, surfaced quantitative methods as the clearest gap and forced an explicit choice about whether to build that capability internally or lean on adjacent teams.

The hiring sessions reinforced that operational discipline extends to talent acquisition. Speakers from Zillow and TIAA both emphasized behavioral interviewing, structured question banks, and non-traditional sourcing channels including contract workers, internal partners, and vendor referrals. Hard-to-teach attributes, empathy, curiosity, storytelling ability, and negotiation instinct, were consistently ranked above product knowledge, which can be trained.

• Build frameworks with full team input but assign one owner to the initial draft to maintain momentum.
• Identify concrete use cases before finalizing any operational framework so owners can be assigned immediately.
• Treat negotiation aptitude as a financial skill with measurable organizational value, not a soft competency.
• Behavioral interview questions should also set honest expectations about job pace and stress.

Craft at the Individual Researcher Level

Two sessions made the case that research quality ultimately depends on what happens in the room between a researcher and a participant. Drawing on clinical psychology and therapeutic interviewing, one speaker presented five fundamentals for in-depth interviews and argued that the same three conditions that predict therapy success, rapport, unconditional positive regard, and therapeutic alliance, also predict the quality of research sessions and cross-functional partnerships. Practical guidance included coding participant statements line by line to surface mental models, building discussion guides collaboratively with stakeholders before fieldwork, delegating operational tasks so the lead researcher stays fully present, and debriefing immediately after each session before findings blur.

A separate session on a long-form documentary research program offered a candid counterpoint: running eight rounds of quantitative studies across duplicated samples exceeded what was necessary. When results align across multiple studies, that is a signal to stop testing, not to add more waves. The clearest missed opportunity named by the speaker was the absence of even one qualitative phase, despite an early quantitative signal that warranted follow-up.

Tracking and Longitudinal Research in Volatile Conditions

Three sessions dealt directly with research conducted under pandemic conditions, and each reinforced the value of continuity and curation over volume. A seven-month consumer tracker launched one week into the U.S. shutdown followed the same 100 participants across nine waves, making wave-over-wave sentiment shifts meaningful rather than anecdotal. A general-population sample was augmented with brand-relevant segments, and findings were organized into clusters of three: what consumers felt, what they did, and what they needed. Distribution evolved away from long decks toward a one-page newsletter, an internal research library, and a planned podcast to reach stakeholders who prefer listening over reading.

A seven-year brand measurement evolution at a financial services firm demonstrated that iterative change outperforms methodological overhauls. Early wins came from simple adjustments before any structural changes were made. Consolidating two brand surveys into one freed budget that funded a six-vendor specialist ecosystem. Daily fielding revealed market events that quarterly point-in-time tracking would have missed entirely. The consistent principle across both talks: preserve trend continuity by routing new monitoring needs through always-on or syndicated tools rather than altering core trackers mid-flight.

Adapting When the Plan No Longer Holds

The NASCAR session offered one of the conference's most direct accounts of research under pressure. When the 2020 plan was invalidated, the team simplified immediately: identify the handful of KPIs still measurable and orient all activity around them. Third-party research aggregation, synthesizing external findings into a weekly executive brief, kept the insights function visible and useful when custom research was neither feasible nor tactful. A competitive calendar of other leagues' return timelines, built entirely from public sources, gave leadership a clear broadcast landscape view and helped the organization move first. The team was also explicit about the limits of research in values-driven decisions: research did not drive the Confederate flag ban, but it quantified likely impact and supplied post-announcement reassurance.

• Avoid asking panels about spending or return intentions before conditions make the question appropriate.
• Curating and synthesizing external findings is a legitimate, high-value deliverable when custom research is constrained.
• Being first back to live programming produced measurable audience gains because there was no competition for live content.

Multi-Method Program Design and Activation

A session on complex, long-term strategic questions made the case for replacing single- or two-phase projects with structured multi-method research programs. The core discipline is mapping every key research question to a specific method or information source before fieldwork begins, which reveals gaps, prevents redundancy, and creates stakeholder alignment upfront. Internal subject-matter experts, regional managers, and customer care staff should be treated as information sources on par with formal methods and brought in at the right stage rather than consulted only at the end. The recommendation to hold the full story until data across all or most methods is in hand guards against partial findings misleading stakeholders before the picture is complete.

AI Integration as a Strategic Responsibility

The conference's most forward-looking session argued that AI is a tool for insights teams today, not a future consideration. With over 80 percent of data produced today being unstructured, traditional methods alone cannot keep pace at scale. The speaker presented two practical roles for insights teams: embedding AI tools to accelerate insight generation, and using traditional research methods to audit, pressure-test, and refine AI and machine learning solutions. Insights teams can add concrete value to data science partnerships by building upfront hypotheses, verifying data authenticity, flagging imbalances in training data, and gathering ongoing stakeholder feedback after deployment.

Human intelligence is not replaceable by AI. AI mimics human reasoning but does not replicate consciousness, emotional nuance, or creativity.

The call for diverse perspectives in AI development was not abstract. The session grounded it in the history of computing itself, noting that contributions from outside the mainstream shaped the field in foundational ways. For insights teams building AI partnerships today, that history is a practical argument for inclusive input into how models are trained and where they are applied.