Savanta (formerly ComRes) is one of the UK's leading polling firms, known for polling for The Daily Mail and i newspaper. They conduct weekly voting intention surveys using online panel methodology and are a full member of the British Polling Council.

What methodology does Savanta use?

Savanta uses an online panel methodology with quota sampling by age, gender, region, and social grade. Surveys are weighted by 2024 General Election vote recall. Typical sample sizes are 1,500-2,000 respondents.

What did Savanta show at the 2024 general election?

Savanta performed broadly in line with the industry at the 2024 General Election. Their Reform UK figures were slightly below the final result, a common issue across most pollsters. Since 2024, their tracking has shown Reform UK rising consistently to around 28-30% in May 2026.

What is Savanta's current voting intention for UK parties?

Savanta's most recent voting intention (May 2026) shows Reform UK at approximately 27%, Labour at 20%, Conservatives at 19%, Liberal Democrats at 21%, and Greens at 9%. Savanta uses telephone polling methodology, which produces slightly higher Labour and lower Reform figures than online panels. They publish polls for the Daily Telegraph and other outlets. Their results are broadly consistent with the industry average with the typical phone-panel adjustment.

Has Savanta changed its methodology after 2024?

Savanta made methodological adjustments after the 2024 election to address the pattern of Conservative underestimation common across phone polls. They increased the weight given to 2019 and 2024 past vote recall in their weighting scheme, which has produced slightly higher Conservative and slightly lower Labour figures in 2025-2026 polling compared to their 2023-2024 baseline. These adjustments are consistent with industry-wide changes in response to the 2024 accuracy review. Savanta's new methodology has not been tested against a general election but is considered more robust by independent methodologists.

Who commissions most of Savanta's political polling?

Savanta publishes most of its UK political polling for The Daily Telegraph. The Telegraph's audience is traditionally Conservative-leaning, and there has been academic debate about whether polling commissioned by partisan outlets systematically differs from polls commissioned by more neutral outlets. Analysis of Savanta's results over time shows they are within the normal range of industry variation, suggesting no significant client-direction effect. Savanta also conducts polling for the Scottish Parliament and Welsh Senedd, making them one of the more active firms in devolved political research.

Pollster Profile

Savanta: UK Voting Intention, Methodology & Accuracy

Last updated: May 2026 — BritPolls Editorial

Savanta is one of the UK most active polling firms, formed by the rebrand of ComRes in 2022. ComRes had a long and distinguished pedigree in British political research, having published voting intention data continuously since the early 2000s and holding well-known contracts with ITV News and the Daily Mail. The Savanta rebrand aligned the firm with the broader Savanta group brand while preserving the methodological continuity of the ComRes operation.

ComRes was particularly associated with constituency polling — the firm conducted large-scale marginal seat surveys for ITV News during the 2015, 2017, and 2019 general elections. That heritage of sub-national polling continues under the Savanta brand.

Savanta methodology

Savanta conducts voting intention polling exclusively via online panel. The firm operates its own proprietary panel, supplemented by third-party panel providers when needed. The standard voting intention survey process:

Sample design: Quota-based sampling targets age, gender, region, and social grade proportions drawn from ONS population estimates.
Fieldwork: Surveys run for 24–48 hours, enabling fast turnaround when news events require rapid polling. Standard sample size is 2,000 completed interviews.
Screening: Respondents are screened for Great Britain residency and age (18+). Attention checks remove low-quality responses.
Weighting: Post-fieldwork weights are applied for age, gender, region, social grade, and 2024 General Election vote recall. Savanta adjusts for differential recall bias.
Undecided treatment: Undecided respondents are not forced to a choice. Published headline figures exclude “don’t know” and “would not vote” responses from the denominator.

Savanta May 2026 voting intention

Labour

26%

Reform UK

27%

Conservatives

22%

Lib Dems

13%

Greens

Savanta, fieldwork 6–7 May 2026. n=2,012.

Recent Savanta polls: 2025–2026

Date	Lab	Con	Reform	LD	Grn	n
7 May 2026	26%	22%	27%	13%	7%	2,012
23 Apr 2026	27%	22%	26%	13%	7%	2,008
9 Apr 2026	27%	23%	25%	13%	6%	2,003
26 Mar 2026	28%	23%	24%	13%	7%	2,016
12 Mar 2026	28%	23%	24%	13%	6%	2,019
26 Feb 2026	29%	24%	23%	13%	6%	2,011

Savanta accuracy record

Election	Final Savanta/ComRes poll	Actual result	Assessment
2024	Lab 37%, Con 20%, Ref 15%, LD 13%	Lab 33.7%, Con 23.7%, Ref 14.3%, LD 12.2%	MAE 3.3 pts; overstated Lab, understated Con
2019	Con 42%, Lab 36%	Con 44%, Lab 33%	Good on Con margin; Lab slightly overstated
2017	Con 44%, Lab 38%	Con 42%, Lab 40%	Slightly overstated Con lead
2015	Con 33%, Lab 35%	Con 37%, Lab 31%	Industry-wide failure; significant Con underestimate

Savanta house effects: Reform UK and Labour

Savanta sits close to the centre of the UK polling distribution. The firm does not exhibit a particularly strong Reform-high or Reform-low bias. In 2026, Savanta Reform figures are typically within 1 point of YouGov and within 2 points of Techne. For Labour, Savanta tends to show figures approximately 1 point below YouGov. Where Savanta shows a consistent pattern is slightly lower Lib Dem figures than most rivals — typically 1–2 points below the polling average. For the full house-effects analysis, see the house effects page.

How to read Savanta numbers

Savanta publishes less frequently than YouGov or Redfield & Wilton — typically fortnightly rather than weekly. When reading Savanta data, account for the fact that the survey fieldwork runs over 48 hours, which means results can be affected by major news events that break during fieldwork. The firm’s standard 2,000-person sample gives a margin of error of approximately plus or minus 2.2 percentage points at 95% confidence. To track the full cross-firm average, see our voting intention tracker. All Savanta data tables are deposited with the British Polling Council archive on publication.

Savanta and the ComRes legacy

ComRes was founded in 2003 and became one of the most prominent names in UK political polling, particularly through its long-running ITV News and Daily Mail polls. The firm was known for its attention to constituency-level variation and for introducing innovations in online panel management. The rebrand to Savanta in 2022 merged the political research operation into the broader Savanta group — a global market research business — while preserving methodological continuity. Former ComRes staff continue to lead the UK political polling operation. Clients have included ITV News, the Daily Mail, The Scotsman, and various political parties and campaign groups. The ComRes datasets, covering over two decades of British political opinion, are archived and available via the UK Data Service.

Savanta constituency polling heritage

One of Savanta distinctive strengths, inherited from ComRes, is constituency-level polling. The firm has conducted large-sample surveys in specific Westminster constituencies, providing granular data on individual seat contests that national polls cannot provide. During the 2019 general election campaign, ComRes fielded MRP-adjacent constituency surveys for ITV News. Under the Savanta brand, this constituency capability has been extended, with the firm conducting targeted seat surveys for various media clients in 2024 and 2025. For constituency-level projections, see the MRP tracker.