Earnings, revenue, and KPI forecasts on US equities

What is ExtractAlpha?

ExtractAlpha is a quantitative data provider whose Estimize platform crowdsources forward looking financial forecasts from a broad community of contributors. Estimize collects estimates from independent analysts, buy side and sell side professionals, private investors, and academics, and turns them into a consensus that is structurally different from feeds drawn only from brokers. The result is a more representative view of expectations on US equities than any single contributor pool can produce on its own.

‍

What academic researchers should know about ExtractAlpha Estimize data

Coverage runs from 2012 to the present across US equity earnings per share and revenue estimates, US equity KPIs, and economic indicators. Each observation captures a single forecast from a single contributor at a specific point in time, with daily refreshes. Because the contributor base extends well beyond brokerage analysts, the data is less prone to the herding and selection biases documented in traditional sell side feeds, which is a feature when the research question is about analyst behavior, consensus formation, or how new information sources reshape expectations.

‍

Why academic researchers choose ExtractAlpha on Dewey

Most academic finance research on earnings expectations has historically run through consensus feeds drawn only from brokerage analysts, which carry well documented biases tied to incentive structures inside sell side firms. Estimize provides a structurally different consensus, with contributor identifiers and timestamps that make it straightforward to study who is forecasting, when, and how their estimates evolve. That makes it a magnet for accounting and finance scholars studying analyst behavior, market reactions, and the role of crowdsourced expectations. It slots in alongside traditional consensus feeds when researchers want to test whether the crowdsourced view adds information over the existing baseline.

‍

ExtractAlpha academic research ideas and use cases

Forecaster nudges and herding behavior. Can a simple platform design choice make consensus forecasts more accurate? Researchers from Temple, University at Buffalo, and University of Saskatchewan ran a social norm nudge directly on the Estimize platform and showed that nudging contributors away from the existing consensus reduced herding, lowered bias, and produced a consensus that better predicted actual earnings. This is fertile ground for behavioral economics work on professional forecasters and for further field experiments inside online forecasting communities.

-Effects of crowdsourced forecasting on firm behavior. Does the arrival of a crowdsourced consensus change how firms communicate or how their stock prices behave? A researcher at George Washington University used Estimize to study exactly that question, documenting shifts in firm disclosure and information environment once a firm became part of the platform. Researchers studying corporate disclosure, market microstructure, or the spillover effects of additional information sources have a clean natural experiment to work with.

Event studies and earnings surprise. With per-contributor estimates and timestamps, Estimize lets researchers reconstruct how expectations move into earnings announcements and how the market reacts to surprises measured against the crowdsourced consensus rather than the traditional one. This is a natural setup for comparing surprise definitions, testing the drift in returns after earnings announcements, or studying how different consensus measures predict abnormal returns.

Crowdsourced macro and KPI forecasts. Beyond earnings and revenue, the dataset includes US equity KPIs and economic indicators, which open research questions in macroeconomic forecasting and operational metric prediction. Researchers can compare the accuracy of crowdsourced macro forecasts against professional consensus, study which contributor types do best on which kinds of economic series, and test whether KPI level estimates carry information beyond top-line forecasts.

‍

Dive deeper with Dewey documentation

Detailed information on onboarding with Dewey, data partner details, and technical documentation on data access.

View documentation

‍