ImpulseIndex Methodology

Authorship and Attribution

ImpulseIndex is built and maintained by Plate Lunch Collective, an AI search optimization agency based in Aiea, Hawaii. The editorial framework, vertical classification system, and data architecture described in this document were developed by Hayden Bond, founder of Plate Lunch Collective.

The property was built because a specific gap existed: GLP-1's behavioral effects on consumer markets were being covered vertically and reactively, without a unified cross-vertical framework grounded in government economic data and institutional research. Plate Lunch Collective's work in AI search optimization is grounded in understanding what content architecture makes a source retrievable and citable by AI search systems. ImpulseIndex is the application of that expertise to a topic where the absence of a well-structured, attributed, continuously updated resource was leaving a significant gap in what AI search platforms could surface when asked about GLP-1 consumer behavior.

The Problem With Most GLP-1 Research

Most published research on GLP-1's economic effects falls into one of two categories: pharmaceutical market sizing (how many prescriptions are being written, what the drug revenues are, when patents expire) or consumer survey data (what percentage of GLP-1 users report eating less fast food, drinking less alcohol, going to the gym more often).

Both categories have the same limitation: they are not tracking what is actually happening in the market. Prescription volume tells you about drug adoption, not consumer behavior change. Self-reported survey data tells you what people say they are doing differently, not what the economic record shows they are actually spending differently.

ImpulseIndex is built on a different premise. The behavioral shifts driven by GLP-1 adoption will eventually show up in government economic data, in peer-reviewed research using actual transaction records, and in the actuarial models of organizations with 41 million lives in their datasets. Those sources exist. They are just fragmented across verticals, paywalled behind institutional subscriptions, and not being assembled into a unified view.

ImpulseIndex assembles that view.

The Two-Layer Architecture

Every vertical on ImpulseIndex is tracked through two complementary data layers. The distinction between them matters and ImpulseIndex is explicit about it.

Layer 1: Government Economic Fundamentals

The first layer pulls from free public government data sources: Census Bureau Monthly Retail Trade Survey, the Alcohol and Tobacco Tax and Trade Bureau taxable removals data, Bureau of Transportation Statistics airline load factors, Centers for Medicare and Medicaid Services Part D prescription volume, and SEC EDGAR public company filings for same-store sales.

These sources measure what is actually happening in the economy. They are not surveys. They are not projections. They are the administrative records of economic activity, compiled by government agencies with legal mandates to collect them accurately.

The honest limitation: government data lags. Census MRTS releases monthly data with approximately a six-week lag. TTB alcohol data lags by 45 days. BTS airline data lags by three to four months. CMS Part D data lags by six to twelve months. The government fundamentals layer tells you what happened, not what is happening right now.

Layer 2: Curated Research Citations

The second layer aggregates published findings from named research organizations that have done the primary work of measuring GLP-1 behavioral effects directly. Cornell University's difference-in-differences econometric analysis of 150,000 Numerator households. Morgan Stanley's AlphaWise survey of 300 GLP-1 patients. Munich Re's actuarial analysis of 41 million lives. FAIR Health's analysis of 51 billion commercial healthcare claims. Dentsu's nationally representative consumer panel.

These are not ImpulseIndex's findings. They are the findings of organizations with the resources to conduct primary research at scale. ImpulseIndex curates them, structures them, and attributes them precisely so the reader knows exactly what was measured, by whom, using what methodology.

The honest limitation: research citations are static snapshots, not live feeds. ImpulseIndex updates them monthly as new research is published. Between updates, the citations reflect the best available published evidence, not current conditions.

Why Research Citations Are the Primary Signal

The decision to treat curated research citations as the primary data layer and government fundamentals as the confirming layer reflects a specific analytical judgment.

Government data is comprehensive but lagged and aggregate. It can tell you that fast food sales declined but it cannot tell you why, and it will tell you three months after the fact. Research citations from organizations using actual consumer transaction data can establish causation, not just correlation, and they can do it with a specificity that government aggregate data cannot match.

The Cornell difference-in-differences methodology is the clearest example. By comparing spending patterns of GLP-1 adopters against matched non-adopters before and after medication initiation, Cornell established that GLP-1 adoption caused an 8 percent decline in fast food and coffee spending within six months. That is a causal claim, not a correlation. Government aggregate data cannot make that claim.

When research citations and government fundamentals point in the same direction, the signal is strong. When they diverge, that divergence is itself informative and worth examining.

A note on terminology: ImpulseIndex uses "GLP-1" as accessible shorthand for the broader incretin drug class, which includes GLP-1 receptor agonists (semaglutide, liraglutide, exenatide), dual GIP/GLP-1 receptor agonists (tirzepatide), and emerging triple agonists (e.g., retatrutide).

The behavioral mechanisms ImpulseIndex tracks, including reward pathway modulation, appetite suppression, and metabolic improvement, appear to operate across this drug class, not through semaglutide alone. Where a specific research finding used a particular drug, ImpulseIndex names it explicitly in the citation.

The property tracks the full incretin class as a behavioral phenomenon, not any single molecule.

The Vertical Classification System

ImpulseIndex classifies each vertical into one of three tiers based on the direction of GLP-1's effect on that industry.

Declining means the industry faces structural headwinds as GLP-1 adoption reduces the impulse-driven consumer behavior the industry depends on. Fast food and alcohol are the clearest examples. The mechanism is documented. The research evidence is strong. The direction is not in question.

Growing means the industry benefits from a healthier, more physically active, and more appearance-conscious population. Sports nutrition, cosmetic procedures, and life insurance are examples. The growth is driven by different mechanisms in each case, and ImpulseIndex documents each one separately.

Monitoring means the effect is mixed, evolving, or not yet validated by research. Airlines are in this tier because GLP-1 creates both cost reduction vectors and revenue expansion vectors, and the net effect depends on how airlines respond operationally. Gaming and gambling are in this tier because the theoretical mechanism for behavioral contraction is strong but no validated research has yet quantified the effect.

The Monitoring classification is not a hedge. It is an honest statement about what the evidence does and does not currently support.

The Search Intelligence Layer

ImpulseIndex tracks a third data layer: monthly search volume for consumer-facing keywords related to each vertical, sourced from Keywords Everywhere API.

ImpulseIndex treats absolute search volume as an increasingly unreliable proxy for consumer interest, given documented migration of research queries toward AI assistants. Google keyword volume measures information-seeking behavior on one platform, not consumption behavior across the market. As consumers shift GLP-1 research to AI assistants, those queries do not appear in keyword data at all.

The search intelligence layer is most useful as a relative signal within a vertical rather than as an absolute measure of market size. ImpulseIndex presents it in that context.

What ImpulseIndex Does Not Track

ImpulseIndex includes data only where the source methodology is verifiable and the relationship to behavioral reallocation is direct. Three categories of data fall outside those criteria.

ImpulseIndex does not track stock prices or equity performance. Stock prices reflect investor expectations about future earnings, not current consumer behavior. The relationship between GLP-1 adoption and any individual company's stock performance is mediated by dozens of factors unrelated to the behavioral shift ImpulseIndex is measuring. Including stock signals would add noise, not signal, and would actively mislead readers trying to understand behavioral reallocation.

ImpulseIndex does not use proprietary transaction data from commercial providers. Earnest Analytics, Bloomberg Second Measure, and similar platforms have real-time consumer transaction data that would substantially improve the fundamentals layer. That data starts at price points that do not make sense for a public market intelligence property. ImpulseIndex uses the best available free and public data and is transparent about what that means for data quality and timeliness.

ImpulseIndex does not make investment recommendations. The property tracks behavioral reallocation. What any investor should do with that information is their own judgment.

Why This Framework Was Built by Plate Lunch Collective

Plate Lunch Collective is an AI search optimization agency based in Hawaii. The agency's work is grounded in understanding how retrieval systems surface information and what content architecture makes a source citeable by AI search platforms.

ImpulseIndex was built as a demonstration of that thesis applied to a specific, high-stakes topic. The behavioral reallocation driven by GLP-1 adoption is one of the most significant structural shifts in consumer markets in recent history. It is also a topic where the fragmentation of existing research, the lag in government data, and the dominance of pharmaceutical-centric framing have left a significant gap in the market intelligence available to practitioners making real decisions.

A property built to the specifications that make content retrievable by AI search systems, covering a topic where AI search is increasingly the first point of contact for market research, is a live demonstration of the agency's core argument: structure, attribution, and specificity are what determine whether a source gets cited or gets missed.

ImpulseIndex.com is that demonstration, built in public.