Nanos Research

Nanos Research is a Canadian market and public opinion research firm founded in 1987 by Nik Nanos.^[1][2] The company specializes in quantitative and qualitative research services, delivering strategic intelligence to corporations, government agencies, and advocacy associations through tools like political tracking, consumer sentiment analysis, and high-velocity data insights.^[3][1] Headquartered in Canada with operations extending to the United States, Nanos Research employs advanced methodologies, including machine learning, to generate reliable public mood assessments and predictive analytics.^[1][4] As the official pollster for CTV News and producer of the Bloomberg Nanos Canadian Confidence Index, the firm has established itself as a key provider of election forecasting and economic indicators.^[1] Nanos Research is particularly noted for its election polling accuracy, having predicted national vote shares within margins as narrow as one-tenth of a percentage point in past federal elections, earning it recognition as a benchmark for reliability in Canadian political research.^[5][6]

Overview

Founding and Rebranding

Nanos Research traces its origins to 1987, when Nik Nanos established SES Research in Kingston, Ontario, Canada, while he was an undergraduate student at Queen's University.^{[6] [7]} The firm initially operated as a part-time public opinion polling operation alongside Nanos's studies, laying the groundwork for its expansion into market research services.^[6] SES Research grew in its early years by securing contracts from corporations, government agencies, and advocacy groups to gauge public sentiment through survey-based insights.^[1] This foundation emphasized understanding consumer and voter moods, establishing the company's reputation in North American research circles.^[7] The company later rebranded to Nanos Research, adopting the founder's surname to align with its evolving identity in public opinion and strategic polling.^[7] This transition marked a shift from the original SES acronym, though specific timing remains undocumented in primary corporate records.^[1]

Leadership and Key Personnel

Nik Nanos serves as the founder and chief data scientist of Nanos Research, directing its strategic focus on public opinion polling and market analysis.^[8] With expertise in political science and business strategy, he oversees the design and interpretation of proprietary research methodologies employed by the firm.^[9] In recognition of his advancements in market and public opinion research, Nanos was appointed a Member of the Order of Canada (CM) on December 28, 2023.^[10] Nanos holds academic affiliations that bolster his analytical credentials, including a tenure as research associate professor in the Department of Political Science at the University at Buffalo (State University of New York) from 2008 to 2021, where he led longitudinal studies on Canada-U.S. relations.^[11] He also serves as a Global Fellow at the Wilson Center's Canada Institute, contributing insights on North American policy and public attitudes.^[12] These roles underscore his integration of empirical data analysis with cross-border strategic intelligence. Publicly available details on other Nanos Research executives remain sparse, highlighting Nanos's predominant, hands-on leadership in core functions such as poll instrumentation and result synthesis, with family member John Nanos noted as a senior vice president.^[6] This centralized structure aligns with the firm's emphasis on Nanos's proprietary approaches to data-driven decision-making.^[1]

Operations and Scope

Nanos Research maintains its headquarters at 56 Sparks Street in Ottawa, Ontario, with an additional office at 90 Eglinton Avenue East in Toronto, Ontario, supporting operations primarily across North America and extending to global engagements.^[13][1] As part of the broader Nanos Group—which encompasses Ethic Strategies for tailored strategic advisory and Nanos dimap analytika for behavioral and sentiment data targeting—the firm structures itself as an integrated provider of research-driven intelligence.^[1] The company caters to a wide array of clients, including government agencies, corporations, advocacy associations, and media entities such as CTV News and Bloomberg News Service, delivering scalable services from standard tracking to bespoke strategic assessments.^[1][14] Its annual output features consistent high-frequency deliverables, notably the weekly Bloomberg Nanos Canadian Confidence Index, alongside customized reports that inform market shaping and policy decisions.^[1] Employing between 51 and 200 staff, Nanos Research positions itself as a full-service strategic intelligence firm, leveraging proprietary metrics like the multi-dimensional components of its confidence index—encompassing economic, personal, and governance sentiment tracking—to enable nuanced, action-oriented opinion analysis without reliance on singular indicators.^[7][1] This scope underscores its role in high-velocity data provision for decision-makers across sectors.^[15]

Methodology

Data Collection Techniques

Nanos Research primarily utilizes a hybrid data collection methodology that combines probability-based random digit dialing (RDD) across landline and cell phone frames with online survey completion to maintain sample representativeness in the context of falling landline usage rates.^[4][16] This dual-frame RDD approach recruits respondents randomly via live telephone agents, who then administer surveys either telephonically or direct participants to online platforms, ensuring broader coverage of mobile-only households.^[17][18] National polls typically draw from sample sizes of 1,000 or more respondents, achieving margins of error of ±3.0 to ±3.1 percentage points, 19 times out of 20.^[19][20] For tracking surveys, data aggregates over rolling periods, such as four-week averages of 250 interviews or three-night averages of 400, to enhance stability while preserving probabilistic foundations.^[4][21] All methodologies adhere to standards set by the Marketing Research and Intelligence Association (MRIA), including full disclosure of weighting procedures to adjust for demographics such as age, gender, region, and education, thereby correcting for potential non-response biases.^[22][23] Weighting variables are calibrated against census benchmarks to reflect the Canadian adult population aged 18 and over.^[24]

Analytical Approaches

Nanos Research processes collected data through proprietary multi-dimensional frameworks that emphasize directional indicators of voter behavior, including shifts in ballot support, preferred prime ministers, and unprompted issue salience, derived from aggregated telephone survey responses.^[21] These approaches prioritize rolling averages—such as four-week aggregates for stability—and first-ranked unprompted responses to capture authentic preference dynamics without priming effects from prompted questions.^[25][26] A core element involves unprompted issue tracking, where respondents freely identify top concerns, revealing empirical priorities like inflation, housing, and economic stability as dominant in recent waves, independent of media framing.^[27] This method contrasts with prompted surveys, enabling causal inference on sentiment drivers by correlating issue rankings with observable economic conditions rather than narrative-driven interpretations.^[28] In consumer confidence indexing, such as the Bloomberg Nanos Canadian Confidence Index, post-collection analysis integrates survey metrics on personal finances, economic expectations, and purchase intentions with macroeconomic data to model sentiment trajectories, using statistical weighting for demographic balance via dual-frame telephone sampling.^[29] This avoids assumptions of exogenous media influence, focusing instead on direct linkages between indicators like job market perceptions and overall indices.^[30] Nanos integrates traditional statistical techniques, including margin-of-error calculations and probabilistic sampling from large RDD (random digit dial) panels, with scaled data volumes from nightly polls of 300-400 respondents to enhance precision over smaller or non-representative samples.^[21] The firm critiques over-reliance on self-reported online panels, which often inflate educated, urban, or activist demographics due to opt-in biases, favoring live-agent telephone methods for verifiable random selection and lower non-response distortion.^[4][31] This hybrid of rigorous stats and volume-driven tracking supports robust derivation of causal patterns in public opinion shifts.^[32]

Services and Products

Political and Election Tracking

Nanos Research conducts ongoing political tracking services focused on Canadian federal elections, including daily and nightly ballot support measurements and preferred prime minister metrics. These efforts encompass the 2025 Canadian federal election, with tracking initiated well in advance of voting day on October 20, 2025.^[33][34] The firm has provided such election monitoring since the introduction of daily tracking polls during the 2004 federal election, marking the first instance of continuous polling in Canadian electoral history.^[35] The tracking includes weekly political and issue monitoring, capturing voter concerns such as jobs and the economy, which have frequently ranked as top priorities in recent aggregates.^[36][37] Data collection relies on random-digit dialing (RDD) hybrid telephone surveys, typically involving rolling samples of 1,000 respondents over four weeks.^[4] Partnerships with media outlets like CTV News and The Globe and Mail enable nightly updates during election campaigns, offering real-time insights into shifting voter intentions.^[21] Regional and demographic breakdowns are integral to the service, allowing for granular analysis by province, age, and gender, though full access requires subscription.^[38][33] These metrics support strategic guidance for political campaigns and media outlets by highlighting trends in key battlegrounds and issue salience. Publicly available data is limited to national-level aggregates for ballot support and leader preferences, with comprehensive datasets, including historical trends from prior elections, reserved for paid subscribers via the Nanos Federal Political Tracking Portal.^[33][39][40]

Market and Consumer Research

Nanos Research offers bespoke market research services tailored for corporations, encompassing analyses of brand perception, emerging market trends, and competitive positioning through empirical examination of consumer behaviors and preferences. These custom engagements support strategic decision-making by delivering actionable insights derived from targeted data collection on purchasing patterns and market dynamics, with a emphasis on North American contexts.^[33][1] A key component of its consumer research portfolio is the Bloomberg Nanos Canadian Confidence Index (BNCCI), a proprietary weekly tracker developed in partnership with Bloomberg News to gauge Canadian sentiment on economic conditions, job security, and discretionary spending intentions. The index, scored on a 100-point diffusion scale where readings above 50 indicate positive outlook, aggregates responses from random telephone surveys of roughly 1,000 adults using a four-week rolling average for stability.^[25][29] Recent BNCCI data have underscored persistent challenges in consumer sentiment, including downward pressures from inflation concerns and housing market stagnation; for example, the index reached a five-month low in October 2025 amid broader economic uncertainty, with sub-indices on personal finances and future direction showing notable declines.^[29][41] These metrics provide corporations with timely indicators of sectoral shifts, such as reduced optimism in retail and real estate sectors, informing adjustments in marketing and investment strategies.^[42][43] In addition to syndicated indices, Nanos Research conducts specialized studies for consumer goods sectors, incorporating tools like mystery shopping to evaluate retail execution and customer experiences, thereby offering verifiable benchmarks for competitive intelligence.^[44] This focus yields data on verifiable consumer responses to pricing, product innovation, and brand loyalty, distinct from broader economic polling.^[45]

Custom Strategic Intelligence

Nanos Research offers bespoke strategic intelligence services that extend beyond routine polling to deliver customized advisory reports for long-term planning, targeting nonprofits, governments, and businesses seeking insights into policy consequences, risk evaluations, and potential future scenarios.^[3] These tailored analyses incorporate proprietary methodologies to assess organizational challenges, including public sentiment alignment and market positioning, enabling clients to develop mission-oriented strategies.^[46] The firm's approach integrates quantitative polling data with qualitative assessments, employing machine learning algorithms to forecast causal relationships, such as public reactions to proposed fiscal or regulatory changes.^[1] For instance, strategic counsel may evaluate voter or consumer backlash risks by modeling sentiment shifts against policy variables, providing decision-makers with probabilistic outcomes rather than mere descriptive statistics.^[47] In practice, Nanos has assisted advocacy organizations with data-driven recommendations on engaging stakeholders, exemplified by advisory work for a global nonprofit aimed at influencing public and decision-maker opinions without presupposing policy merits.^[46] Such engagements emphasize empirical evidence from hybrid research designs, yielding actionable intelligence on reputational risks and opportunity identification, while maintaining methodological transparency through adherence to standards like those of the Canadian Research Insights Council.^[1]

Historical Track Record

Early Polling Accuracy (1987–2005)

Nanos Research, originally operating as SES Research, was established in 1987 by Nik Nanos in Kingston, Ontario, where initial efforts in political polling faced sustainability challenges in a small market, prompting a primary focus on market and consumer research.^[6] The firm nonetheless developed capabilities in public opinion surveys, entering political polling more substantively in the 1990s with emphasis on provincial elections, where its limited but reliable forecasts established an early baseline of competence amid peers reliant on traditional landline methods.^[5] By the early 2000s, as mobile phone penetration began eroding landline sample representativeness, Nanos expanded federal election tracking, including during the 2000 contest, by integrating hybrid telephone approaches to maintain coverage and reduce nonresponse biases. Its methodology prioritized unprompted, open-ended questions to gauge top-of-mind voter priorities, such as economic concerns, enabling more authentic detection of shifts without the priming effects of prompted alternatives that plagued competitors' surveys. This approach yielded verifiable outperformance; for instance, in the 2004 federal election, Nanos's final poll (n=977, margin of error ±3.1%) forecasted Liberal support at 34.2% (actual 36.7%), Conservatives at 30.4% (actual 29.6%), NDP at 19.5% (actual 15.7%), and Bloc Québécois at 12.2% (actual 12.4%), positioning it as the most accurate aggregator according to La Presse analysis.^[5] Such precision underscored Nanos's early edge in causal inference from raw voter sentiment over stylized questioning.

Performance in Federal Elections (2006–Present)

In the 2006 Canadian federal election held on January 23, Nanos Research's tracking polls accurately anticipated a Conservative minority government, projecting the party's vote share within 1.5 percentage points of the final result of 36.3% and correctly identifying the narrow path to 124 seats, in contrast to Ipsos Reid's overestimation of Conservative seats at 143-147.^[48][5] This performance underscored Nanos' early reliance on hybrid telephone and online methodologies to capture late-deciding voters in a fragmented race.^[5] Nanos demonstrated consistent accuracy in subsequent elections by closely tracking the dominant Liberal-Conservative contest. In the 2015 election on October 19, its final tracking aligned precisely with the NDP's 19.7% vote share (to one decimal place) and Liberal surge to 39.5%, contributing to projections of a majority government that materialized with 184 seats.^[49] For the 2019 election on October 21, Nanos' aggregates underestimated Conservative support by under 2 points (actual 34.4%) but captured the Liberal minority outcome at 33.1%, with regional breakdowns aiding seat models amid a three-way split.^[50][5] The 2021 snap election on September 20 saw Nanos within 1-2 points on major parties, reflecting stable Liberal-Conservative dynamics at around 32% each, though turnout volatility slightly widened errors; its nightly tracking highlighted voter fatigue without overemphasizing minor party gains.^[51][5] As of October 2025, amid tracking for the ongoing federal campaign, Nanos reports a statistical dead-heat between Liberals and Conservatives, with ballot support hovering at 38-40% for each in early October aggregates, underscoring deepened partisan polarization and minimal third-party traction beyond regional bases.^[52][53] This neutral depiction contrasts with some media interpretations amplifying volatility, as Nanos' methodology emphasizes committed voter intentions over undecideds.^[21]

Reception and Criticisms

Achievements and Recognition

Nanos Research has earned recognition as Canada's leading polling firm through partnerships with major media outlets, including CTV News and The Globe and Mail for election tracking, and Bloomberg News for the weekly Bloomberg Nanos Canadian Confidence Index, which it has produced as the firm of record for over a decade.^[54][55][6] These collaborations underscore its methodological rigor in a competitive landscape where many pollsters prioritize rapid surveys over hybrid telephone-online approaches, often resulting in higher variance in predictions.^[21] The firm's track record includes notably accurate federal election forecasts, such as setting a record for precision in the 2006 election and maintaining low average errors in subsequent cycles, earning descriptions of its numbers as the "gold standard" in Canadian political polling amid critiques of less consistent rivals.^[5] Independent analyses, like those post-2021 election, have placed Nanos among top performers relative to aggregates such as the CBC Poll Tracker.^[51] Founder Nik Nanos has received professional acknowledgment through frequent speaking engagements at conferences and as a commentator on data trends, leveraging the firm's insights into public sentiment dynamics.^[56][2] This recognition highlights Nanos Research's edge in delivering verifiable, low-error intelligence over flashier but empirically weaker competitors in the polling industry.^[5]

Methodological Debates and Accuracy Disputes

Nanos Research employs a hybrid methodology combining random digit dialing (RDD) dual-frame telephone surveys (landline and cellular) with online supplementation, emphasizing telephone interviews to achieve broader demographic reach and mitigate non-response biases inherent in online-only panels.^[4] This approach has sparked debates in Canadian polling circles regarding mode effects, where online surveys often overestimate support for left-leaning parties due to overrepresentation of urban, educated, and progressive respondents who are more likely to participate in opt-in panels, while telephone polling better captures rural and conservative-leaning voters less inclined toward digital self-selection.^[57] Empirical comparisons from the 2015 federal election, for instance, revealed systematic differences in vote intention estimates between telephone and online modes, with telephone methods yielding results closer to validated turnout patterns by reducing coverage errors among lower-propensity voters.^[58] Nanos defends its telephone-heavy hybrid as empirically superior for directional stability, arguing that pure online methods amplify house effects—consistent biases tied to polling firm practices—that distort aggregate trends toward urban progressive overestimation. Critics have noted occasional delays in Nanos tracking capturing abrupt opinion shifts, such as short-term "bounces" following leadership resignations or scandals, attributable to the multi-day fieldwork cycles required for RDD telephone sampling, which introduce a lag of 3-5 days in rolling averages.^[51] However, Nanos demonstrates stronger long-term directional accuracy compared to online-reliant firms like Léger, as evidenced by its third-place ranking in absolute error for the 2021 federal election (behind Léger but ahead of most others), with consistent underestimation of major parties offset by precise trend forecasting over election cycles.^[51] Léger, employing a mix but with heavier online components, exhibited smaller absolute errors in that contest (4.5 points total across parties) yet higher volatility in interim shifts, underscoring Nanos's advantage in stable weighting schemes that prioritize propensity-to-vote adjustments over raw sample volume.^[59] Nanos counters claims of systemic underpolling of conservatives—often leveled amid discrepancies with online aggregates—by highlighting its transparent weighting protocols, which disclose adjustments for age, gender, region, and education derived from census benchmarks and validated against historical turnout data, resulting in empirically lower house effects than peers.^[4] Analyses of Canadian house effects confirm Nanos's deviations from final results average under 2 points nationally, attributing inter-firm variances to methodological artifacts like mode-specific non-response rather than intentional bias, with telephone integration empirically reducing overreliance on self-selected online cohorts that skew leftward.^[60] This transparency enables independent verification, debunking narratives of conservative suppression as artifacts of comparator firms' online dominance, where causal turnout dynamics favor telephone's probabilistic sampling for realism in vote projection.^[61]

Specific Controversies

In 2013, Nanos Research conducted a poll for Sun News Network that prompted complaints of potential bias, particularly regarding question wording alleged to favor conservative framings on public accountability issues such as Senate expenses. The Marketing Research and Intelligence Association (MRIA) initiated a panel review, concluding on October 2, 2013, that Nanos did not violate professional standards or ethical guidelines in its methodology or execution.^[62] Critics, including labor advocacy groups, argued the poll's structure influenced responses toward support for transparency measures aligned with conservative policy positions, though the MRIA found no evidence of intentional manipulation or deviation from industry norms.^[63] During the lead-up to the 2025 Canadian federal election, Nanos faced online critiques on platforms like Reddit and Facebook for underpredicting the extent of the NDP's support collapse, with tracking showing the party at around 7-8% in late April while some competitors registered lower figures earlier.^[64] Nanos responded that such variances stemmed from sample volatility inherent to telephone-based RDD surveys, emphasizing their data's consistency in capturing conservative voter strength without evidence of systemic overestimation of NDP backing.^[65] In the context of broader polling disputes, Nanos avoided the overconfidence pitfalls evident among competitors during the 2006 federal election, where firms like Ipsos Reid projected a Conservative majority (143-147 seats) shortly before voting, contrasting the actual minority outcome of 124 seats.^[48] Nanos's approach incorporated conservative error margins and regional adjustments, aligning final national vote estimates within 0.5 percentage points of results, a precision that insulated the firm from similar accuracy challenges faced industry-wide.^[5]

Impact and Influence

Role in Canadian Politics

Nanos Research's tracking polls have exerted causal influence on Canadian political strategies by illuminating voter issue priorities, particularly the persistent dominance of economic concerns over social or identity-based issues in the 2020s. Weekly surveys conducted ahead of the 2025 federal election consistently ranked jobs and the economy as the top unprompted concern, with data from October 2025 showing it as the clear leading issue amid a tight Liberal-Conservative race.^[66] This empirical pattern, captured through stratified telephone sampling of over 1,000 respondents, has driven campaign pivots, as parties recalibrate platforms to address economic salience—evident in narrowed partisan gaps when focus shifts from external factors like U.S. trade tensions to domestic job security.^[67] Such data counters assumptions of voter preoccupation with identity politics, grounding strategy in first-hand respondent priorities rather than elite-driven narratives. The firm's regional breakdowns further inform tactical decisions, revealing conservative support's stability in suburban and rural ridings despite national volatility, as seen in 2025 tracking where Conservatives maintained leads in non-metropolitan areas even as urban Liberal gains accelerated.^[34] This resilience, derived from Nanos' dual-frame RDD methodology ensuring proportional representation across provinces and demographics, challenges urban-biased projections common in some academic and media analyses, compelling parties to allocate resources toward winnable peripheral seats.^[4] As a non-partisan pollster, Nanos data fosters cross-party accountability by offering verifiable, publicly accessible trends that all major parties—Liberals, Conservatives, NDP, and others—monitor for strategic calibration, independent of proprietary partisan polling.^[33] For instance, the dramatic 2025 shift from Conservative dominance (47% support early in the year) to a statistical tie has prompted universal reassessment of platform viability, emphasizing empirical trends over ideological priors.^[34] This utility enhances causal realism in politics, as leaders adjust to data-driven signals of voter intent rather than insulated assumptions.

Media and Business Partnerships

Nanos Research maintains exclusive partnerships with major media outlets to deliver proprietary tracking data on public opinion and consumer sentiment. Since at least 2006, it has provided CTV News with regular political tracking polls, including ballot support and issue prioritization, often co-branded with the Globe and Mail for weekly national and regional insights.^[54] Similarly, Bloomberg News has engaged Nanos for the Bloomberg Nanos Canadian Confidence Index (BNCCI), a weekly survey of consumer sentiment launched in the early 2010s, with results streamed to global Bloomberg terminals every Monday to inform economic analysis.^[1][47] To broaden empirical access, Nanos operates subscriber-based live data portals offering real-time visualizations and downloadable datasets from its tracking programs, such as political ballot tracking and economic indicators, while restricting full regional and demographic breakdowns to paid users to preserve data integrity and prevent misuse.^[68] Partial free access is provided for national-level election tracking, enabling public verification of headline results without compromising proprietary methodologies.^[33] In 2024, Nanos expanded transparency efforts through new data-sharing initiatives, including enhanced access to aggregated polling datasets via collaborations with academic platforms like the Political Analysis and Negotiation Lab (PANL) at Carleton University, allowing researchers to analyze trends in political and economic sentiment under controlled terms.^[69] Beyond media, Nanos provides business advisory services to corporations, integrating public opinion metrics with policy risk assessments to guide corporate positioning and communications strategies. These engagements emphasize linking consumer sentiment data—such as reactions to regulatory changes or market disruptions—to potential business impacts, delivering action-oriented recommendations derived from hybrid telephone-online surveys without prescribing political or interventional advocacy.^[46][1] Clients receive tailored insights on public acceptance, branding resonance, and issue framing, drawn from Nanos's proprietary panels and high-velocity tracking, to inform evidence-based decision-making in volatile environments.^[3]

Broader Contributions to Public Opinion Research

Nanos Research has advanced polling methodologies by developing hybrid telephone and online survey approaches, which integrate random-digit-dialing (RDD) phone sampling with online panels to mitigate selection biases inherent in low-response-rate digital-only methods.^[20] This hybrid framework addresses the decline in telephone response rates—often below 10% in recent decades—while avoiding the non-representative opt-in biases of pure online panels, such as overrepresentation of urban or tech-savvy demographics.^[19] By weighting responses from both modes against census benchmarks, the method enhances representativeness and reduces house effects compared to single-mode alternatives, contributing to more reliable longitudinal tracking of public sentiment amid technological shifts in data collection.^[20] In analyzing polarization trends, Nanos Research emphasizes empirical patterns over narrative-driven interpretations, highlighting data showing limited divides on issues like climate policy and immigration rather than assuming deepening societal rifts aligned with partisan media framings.^[70] This approach prioritizes observable voter stability and regional variations—such as consistent "supermajorities" on economic priorities across demographics—over projections of irreversible progressive shifts, underscoring causal factors like economic pressures in shaping opinions rather than ideological inevitability.^[71] Such interpretations challenge overreliance on correlational snapshots, promoting causal assessments grounded in repeated surveys that reveal persistent public pragmatism amid polarized rhetoric.^[70] The firm has further contributed through transparent publications of methodological track records and public engagements, including detailed accuracy reports and lectures that expose polling's empirical constraints, such as margins of error and non-response impacts.^[5] These resources, available via Nanos's data portals and speaking series, enable researchers and the public to evaluate long-term performance metrics—like sub-indice tracking of consumer confidence—fostering scrutiny of biases in response modes and question wording.^[55] Nik Nanos's recognition in the Order of Canada in 2023 underscores these efforts in elevating standards for verifiable, data-centric public opinion analysis.^[72]