Frequently Asked Questions

Electric Insights is a senior-led methodology firm. It is the analytical engine that turns research output into decision-ready tools — producing the same headline-explained-and-simulatable deliverable across commercial research, public-opinion polling, and operational or behavioral data. The deliverable turns the report’s headline number into a tool the brand or insights team can interrogate after delivery.

The site shows the approach working in three live case studies, and describes how to engage the firm to apply the same workflow to your data.

No. The case studies cross deliberately different data shapes — a 1963 public-opinion survey, 2014–15 NBA shot-tracking data, and a craft-beer concept test — to show that the same workflow applies across domains.

The platform is data-agnostic. It works wherever there is a binary outcome that matters, in three broad buyer worlds: commercial research (concept tests, conjoint, MaxDiff, pricing, brand tracking, copy testing), public-opinion polling (approval and election research, issue and policy work, mixed-mode survey programs), and operational or behavioral data (transactional records, panel data, event-level logs, sports analytics). See the three buyer worlds on the Approach page for what the engine produces in each.

If you want the big picture, start with Approach. If you want to see the method working immediately, go to Case Studies and pick the case study whose domain matches your interest.

If you already know which domain interests you, go directly to JFK, NBA 3-Point Shots, or the Concept Test.

The approach has four main steps: define a meaningful outcome, model the measured factors associated with it, translate the results into outcome-level terms that non-specialists can understand, and pair the model with tools that let users test scenarios for themselves.

That is the common structure behind all three live case studies on the site.

Many reports stop at the headline result: approval, purchase intent, make rate, conversion, support, or some other topline number. That tells you where the outcome stands, but not what is driving it, how much those drivers matter, or what happens under alternative conditions.

The Electric Insights approach is designed to make that missing explanatory layer visible and testable.

Descriptive summaries are useful, but they do not tell you what survives once multiple predictors are evaluated together. A factor can look prominent in a descriptive table and still contribute little to movement in the headline once other related factors are considered jointly.

That gap between descriptive prominence and explanatory leverage is one of the main reasons the site emphasizes modeling rather than description alone.

Because a model becomes more useful when people can test it. Interactive tools let users inspect assumptions, compare scenarios, see how much the outcome moves, and understand how the same model behaves under different conditions.

That is why the site is organized around live case studies rather than around static prose alone.

The JFK case study uses a 1963 Harris/Newsweek survey to show how explanatory modeling can be paired with public-opinion data. The NBA case study uses 2014–15 SportVU shot logs to show the same logic in sports analytics. The concept test uses a 2024 craft-beer concept test (n=803 consumers) to show the same logic in commercial research, modeling top-2-box purchase intent.

Different domains, same structure: a domain-specific landing page, a set of tools, contextual background, source-data framing, and scenario exploration.

• All-or-Nothing Simulators apply a single condition across the full dataset to show the effect of a maximally uniform scenario.
• Fine-Tuning Simulators redistribute conditions more gradually to mimic more realistic shifts.
• Explorers help users inspect frequencies, cross-tabs, distributions, and related descriptive patterns.
• Model Builders let users build, compare, and inspect explanatory models directly.
• Non-Response Stress Tests let users specify how respondents who didn’t answer might have answered, and see whether the headline survives. Applies to sample-based studies (JFK, concept test); not applicable to census-like data (NBA).

The All-or-Nothing version imposes a single condition across the full dataset. That makes it useful for seeing the effect of an extreme, uniform scenario.

The Fine-Tuning version redistributes conditions gradually instead, which is often a better way to represent realistic change. Both are useful; they just answer different kinds of what-if questions.

Not necessarily. The tools are similar in type, but they are adapted to the structure of the data and the needs of each domain.

That is part of the point of the site: the approach is shared, but the content and interface of each tool can be domain-specific.

Binary outcomes are especially natural: top-2 purchase intent, top-box brand fit, chose-our-offer, recalled-the-ad, renewed, repurchased, approve vs. disapprove, make vs. miss.

Those outcomes are common across all three of the buyer worlds Electric Insights serves — commercial research, public-opinion polling, and operational or behavioral data — which is one reason the same engine travels well across them. See the three buyer worlds on the Approach page for the methods and data sources inside each.

No. The site is designed to make scenario implications visible under a disclosed model. Those implications can be informative and decision-relevant without automatically becoming causal claims.

The right reading is conditional: given the measured predictors, the model specification, and the recorded data structure, what does the outcome look like under the stated scenario?

Electric Insights is led by George Terhanian, Ph.D. — three-decade methodologist, former president at Harris Interactive and at NPD, credited with bringing propensity scoring into market research.

Full background, career history, and credentials are on the About page.

Three shapes: a short Methodology Diagnostic (Tier 1, 2-4 weeks), a full Applied Case Study (Tier 2, 6-10 weeks), or an ongoing Methodology Partnership (Tier 3, 12-month commitment).

Most engagements start with the diagnostic and expand from there if the results warrant it. Full details — scope, timeline, deliverables — are on the Services page.

Building on a concept test or brand study? Brand and insights teams may prefer the dedicated overview for their audience, which uses brand-team vocabulary instead of methodology-buyer terms.

Fixed-fee for scoped work, retainer for ongoing partnerships. No hourly meters; scope changes are agreed in writing before new work begins.

Tier amounts and full pricing structure are on the Services page.