First-round voting intentions: what the polling firms publish

Elabe for BFMTV, La Tribune DimancheFieldwork 14–16 October 2026Sample of 1,701 people, of whom 1,553 registered on the electoral rolls (intentions expressed by registered voters intending to vote, likelihood rated 8 to 10)Internet survey; representative sample of residents of metropolitan France aged 18 and over, quota method (sex, age, socio-professional category, region of residence and type of urban area); adjusted on the quotas, housing tenure, the 2024 legislative vote (1st round) and the 2022 presidential vote (1st and 2nd rounds)Margin of error: between 1.2 and 2.8 percentage pointsPolling firm’s document · Notice (Commission des sondages)

Owned by:Elabe: its partners, through the holding company LUZ (Bernard Sananès and Marie-Rose Beslin, co-managers); breakdown of the share capital not publishedBFMTV: CMA CGM (Saadé family), since 2024La Tribune Dimanche: CMA CGM (Saadé family), since it bought La Tribune in 2023

Scenario 1: M. Tondelier as Green candidate / E. Philippe as central bloc candidate / D. de Villepin as candidate: Express no voting intention: 13

Scenario 2: M. Tondelier as Green candidate / G. Attal as central bloc candidate / D. de Villepin as candidate: Express no voting intention: 13

Scenario 3: M. Tondelier as Green candidate / E. Philippe as central bloc candidate / D. Lisnard as candidate: Express no voting intention: 12

Scenario 4: F. Ruffin as candidate / G. Attal as central bloc candidate / D. Lisnard as candidate: Express no voting intention: 12; Tondelier not tested

Scenario 5: M. Tondelier as Green candidate / G. Attal and E. Philippe as central bloc candidates: Express no voting intention: 11

Scenario 6: M. Tondelier as Green candidate / N. Polony as candidate / E. Philippe as central bloc candidate: Express no voting intention: 11

Recorded: 2026-10-19: Commission notice 10324 (PDF, 9 pp.) read with pdftotext -layout, pp. 1-8; « Publié » (published) tables on pp. 7-8 rendered as images and reread; Elabe report (49 pp., « Les Français et l'élection présidentielle 2027 », linked from the elabe.fr page published on 17/10/2026): first-round summary (p. 7) rendered as an image, scenario labels on pp. 8 to 19; identical figures in the notice and the report. Six scenarios shown in random order, order used = the polling firm's numbering

YouGov for Le HuffPostFieldwork 12–15 October 2026Sample of 1,094 people, of whom 725 decided voters: registered voters who said they intend to vote and stated a voting intention for one of the listed candidates (raw base hypothesis 1: 725; weighted base: 689); 976 registered votersonline poll on YouGov's proprietary panel; weighting (gender, age, region, CSP (socio-professional category), education, interest in politics, area of residence, 2022 presidential vote), plus weighting on age, whether or not people voted in 2022 and the intention to go and vote (0-10) for voting intentionsMargin of error: between plus or minus 1.4 and 3.1 points depending on the score (overall results)Polling firm’s document (notice filed with the Commission des sondages)

Owned by:YouGov: YouGov plc, listed in London with no controlling shareholder; largest shareholders are fund managers (Artemis 9.2%, Liontrust 8.9%, Jupiter 7.2%)Le HuffPost: Le Monde group (about 85%), Matthieu Pigasse (15%)

Hypothesis 1: Nathalie Arthaud: “0%” published in the weighted column (1% raw), recorded as “less than 1%”; raw base 725, weighted 689

Hypothesis 2 – G. Attal: without Édouard Philippe; hypotheses 2, 3 and 4 shown in random order; Arthaud “0%” published (weighted), recorded as “less than 1%”; raw base 713, weighted 680; published total 101 (rounding)

Hypothesis 3 – E. Philippe: without Gabriel Attal; Arthaud “0%” published (weighted), recorded as “less than 1%”; raw base 721, weighted 683; published total 99 (rounding)

Hypothesis 4 – F. Hollande: François Hollande (PS) instead of Raphaël Glucksmann, without Gabriel Attal; raw base 708, weighted 666; published total 101 (rounding)

Hypothesis 5: 14 candidates; Arthaud “0%” published (weighted), recorded as “less than 1%”; raw base 734, weighted 696; published total 99 (rounding)

Recorded: 2026-10-19: Commission notice 10323 (PDF, 29 pp.) read with pdftotext -layout, pp. 4-5 and 10-14 (questionnaire); pages 20 to 24 (first-round results) rendered as images and reread, consistent; « Pondérés » (weighted) column used. Report not found on yougov.com/fr-fr as at 19 October: url = notice

OpinionWay for CNewsFieldwork 13–14 October 2026Sample of 1,043 people, of whom 731 people registered on the electoral rolls certain to vote (score 9 or 10), out of 978 registered votersself-administered online questionnaire on a CAWI system, quota method (sex, age, socio-professional category, type of urban area, region of residence)Margin of error: from 0.7 to 3.5 points at mostPolling firm’s document · Notice (Commission des sondages)

Owned by:OpinionWay: Les Echos-Le Parisien group (LVMH, Bernard Arnault), about 75% since 2022, with its managers (about 25%); sale to the managers announced in August 2026, not completed to our knowledgeCNews: Canal+ group, whose largest shareholder is the Bolloré group (30.4%)

Single hypothesis: Express no voting intention: 15%; same fieldwork and same sample as the OpinionWay PrésiTrack for Les Echos and Radio Classique, separate questionnaire; that PrésiTrack is not included (polling firm controlled by the group that owns Les Echos), this poll is (CNews has no shareholder in common with OpinionWay)

Recorded: 2026-10-19: Commission notice 10319 (PDF, 16 pp.) read with pdftotext -layout, pp. 2-5 and 12; page 12 rendered as an image and reread; « Publié » (published) column = socio-demographic + 2022 presidential + 2024 legislative, among those certain to vote; cross-checked against the OpinionWay report for CNews (29 pp., linked from the survey's opinion-way.com page): chart on p. 8 rendered as an image and breakdowns on p. 9, identical. The notice's questionnaire (p. 5) names François Hollande, but the results (notice and report) cover Raphaël Glucksmann. The report states « 1019 personnes » [“1,019 people”] on its breakdowns (p. 9); the notice, 978 registered voters of whom 731 are certain to vote. Margin cited = statement in the report (the notice gives 1.6 to 3.7 points for 731 people). Same fieldwork and same sample as the OpinionWay PrésiTrack for Les Echos (notice 10318)

Toluna Harris Interactive for M6 and RTLFieldwork 5–8 October 2026Sample of 2,348 people, of whom 2,004 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Hypothesis Raphaël Glucksmann, Gabriel Attal and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,285 people

Hypothesis Raphaël Glucksmann and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,274 people

Hypothesis Raphaël Glucksmann and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,264 people

Hypothesis François Hollande and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,267 people

Hypothesis François Hollande and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,251 people

Recorded: 2026-10-14: Commission notice 10305 (PDF, 69 pp.) read with pdftotext -layout, pp. 3 and 33-46, summary on p. 33 reread as an image; report on tolunacorporate.com (wave 7, page published on 12/10/2026) checked as accessible (HTTP 200)

Odoxa for Backbone, Public Sénat, the regional press and Le Nouvel ObsFieldwork 24 September – 7 October 2026Sample of 12,005 people, of whom 7,599 registered voters intending to vote and having expressed a first-round voting intention (7,599 in the Philippe hypothesis, 7,374 in the Attal hypothesis); 11,103 registered voterssurvey conducted via the Internet; purposive sample of around 1,000 people in each of the 12 metropolitan regions, quota method (sex, age, occupation, département, type of urban area), each region weighted by its actual weight; adjustment on the quota variables and on the vote in the first and second rounds of the 2022 presidential electionMargin of error: between plus or minus 0.6 and 1.2 points depending on the score (report, p. 2); « entre plus ou moins 0,6 et 1,1 point » [“between plus or minus 0.6 and 1.1 points”] in the notice filed with the CommissionPolling firm’s document · Notice (Commission des sondages)

Owned by:Odoxa: its founders Gaël Sliman and Céline Bracq; no known outside shareholder, breakdown of the share capital not publishedBackbone: reputation and opinion consultancy, merged with Enderby in 2025 (twelve partners); breakdown of the share capital not publishedPublic Sénat: public parliamentary channel, funded by the Senatela presse régionale: about twenty titles from several regional press groupsLe Nouvel Obs: Le Nouvel Observateur du Monde, indirectly controlled by the Fonds pour l'indépendance de la presse (Le Monde group) since 2024

Hypothesis: Edouard Philippe as candidate: base: 7,599 registered voters intending to vote and having expressed a voting intention; did not express a voting intention: 28%; fieldwork completed before the PS and Place Publique primary

Hypothesis: Gabriel Attal as candidate: base: 7,374 registered voters intending to vote and having expressed a voting intention; did not express a voting intention: 29%

Recorded: 2026-10-15: Odoxa report (40 pp., « Intentions de vote des Français à l'élection présidentielle 2027, vague d'octobre 2026 », linked from the odoxa.fr page published on 14/10/2026, embargo lifted on 14 October at 6 pm) read with pdftotext -layout, pp. 1-2, 12, 14 and 20; pages 12 and 20 rendered as images and reread; « Total » row of the detailed results (p. 14) and Odoxa summary (left 33% / 35.5%, right 50% / 53.5%) consistent; Commission notice 10310 (6 pp.) read: method, questions and bases, no scores

Toluna Harris Interactive for M6 and RTLFieldwork 22–24 September 2026Sample of 2,334 people, of whom 1,974 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Hypothesis Raphaël Glucksmann, Gabriel Attal and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,242 people

Hypothesis Raphaël Glucksmann and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,227 people

Hypothesis Raphaël Glucksmann and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,216 people

Hypothesis François Hollande and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,219 people

Hypothesis François Hollande and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,223 people

Recorded: 2026-10-14: Commission notice 10277 (PDF, 72 pp.) read with pdftotext -layout, pp. 3 and 34-47, summary on p. 34 reread as an image; report on tolunacorporate.com (wave 6, page published on 28/09/2026) checked as accessible (HTTP 200)

YouGov for Le HuffPostFieldwork 17–21 September 2026Sample of 1,103 people, of whom 719 decided voters: registered voters who said they intend to vote and stated a voting intention for one of the listed candidates (raw base hypothesis 1: 719; weighted base: 646); 1,008 registered votersonline poll on YouGov's proprietary panel; weighting (gender, age, region, CSP (socio-professional category), education, interest in politics, area of residence, 2022 presidential vote), plus weighting on age, whether or not people voted in 2022 and the intention to go and vote (0-10) for voting intentionsMargin of error: between plus or minus 1.4 and 3.1 points depending on the score (overall results)Polling firm’s document (notice filed with the Commission des sondages)

Owned by:YouGov: YouGov plc, listed in London with no controlling shareholder; largest shareholders are fund managers (Artemis 9.2%, Liontrust 8.9%, Jupiter 7.2%)Le HuffPost: Le Monde group (about 85%), Matthieu Pigasse (15%)

Hypothesis 1: raw base 719, weighted base 646; published total 102 (rounding)

Hypothesis 2 – B. Le Maire: hypotheses 2, 3 and 4 presented in random order; raw base 729, weighted 653

Hypothesis 2 – G. Attal: without Édouard Philippe; raw base 719, weighted 641

Hypothesis 3 – E. Philippe: without Gabriel Attal or Raphaël Glucksmann; raw base 707, weighted 636

Hypothesis 4 – O. Faure: Olivier Faure (PS) instead of Raphaël Glucksmann; raw base 733, weighted 658

Hypothesis 5: 16 candidates; raw base 739, weighted 669

Recorded: 2026-10-14: Commission notice 10270 (PDF, 31 pp.) read with pdftotext -layout, pages 21 to 26 (results) rendered as images; « Pondérés » (weighted) column used. Report not found on yougov.com (fr.yougov.com redirects to yougov.com/fr-fr, politics sections return 404, 1 web search with no result): url = notice

Cluster17 for PoliticoFieldwork 15–16 September 2026Sample of 2,045 people, of whom 1,733 registered respondents having expressed a voting intention for first-round scenario 1 (1,733; 1,723 for scenario 2; 1,729 for scenario 3); effective sample after adjustment and weighting by likelihood to vote: 1,337, 1,342 and 1,341; 1,943 registered votersself-administered online questionnaire, quota method (sex, age, CSP (socio-professional category), size of municipality, region); socio-demographic and political adjustment (2022 presidential 1st round, 2024 European elections), weighting by probability of votingMargin of error: published table: for the usable sample sizes of the 1st-round scenarios (1,337 to 1,342 people), from 1.2 points (score of 5%) to 2.7 points (scores of 40 to 60%)Polling firm’s document · Notice (Commission des sondages)

Owned by:Cluster17: its founder Jean-Yves Dormagen, chairman; breakdown of the share capital not publishedPolitico: Axel Springer (Friede Springer and Mathias Döpfner, 95%)

Hypothesis 2: base: 1,723 respondents having expressed a voting intention; effective sample 1,342

Hypothesis 3: without Éric Zemmour or Raphaël Glucksmann; base: 1,729 respondents having expressed a voting intention; effective sample 1,341

Recorded: 2026-10-14: Cluster17 report (21 pp., linked from cluster17.com, article of 24 Sept. 2026) pages 5, 10 and 15 rendered as images; scores = published values rounded to the nearest half-point; cross-checked against Commission notice 10273 (19 pp., « Redressé » (adjusted) column to one decimal place, pages 14-16 rendered as images): consistent

Toluna Harris Interactive for M6 and RTLFieldwork 8–10 September 2026Sample of 2,358 people, of whom 2,052 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Hypothesis Raphaël Glucksmann, Gabriel Attal and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,339 people

Hypothesis Raphaël Glucksmann and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,317 people

Hypothesis Raphaël Glucksmann and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,304 people

Hypothesis François Hollande and Édouard Philippe: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,309 people

Hypothesis François Hollande and Gabriel Attal: base: registered voters certain to vote and having expressed a voting intention, i.e. 1,312 people

Recorded: 2026-10-14: Commission notice 10263 (PDF, 72 pp.) read with pdftotext -layout, pp. 3 and 33-46, summary on p. 33 reread as an image; report on tolunacorporate.com (wave 5, page published on 14/09/2026) checked as accessible (HTTP 200)

OpinionWay for CNewsFieldwork 9–10 September 2026Sample of 1,001 people, of whom 690 people registered on the electoral rolls certain to vote (score 9 or 10), out of 935 registered votersself-administered online questionnaire on a CAWI system, quota method (sex, age, socio-professional category, type of urban area, region of residence)Margin of error: from 0.7 to 3.5 points at mostPolling firm’s document · Notice (Commission des sondages)

Owned by:OpinionWay: Les Echos-Le Parisien group (LVMH, Bernard Arnault), about 75% since 2022, with its managers (about 25%); sale to the managers announced in August 2026, not completed to our knowledgeCNews: Canal+ group, whose largest shareholder is the Bolloré group (30.4%)

Hypothesis 1: François Hollande as Socialist candidate: Express no voting intention: 14%

Hypothesis 2: Ségolène Royal as Socialist candidate: Express no voting intention: 15%

Hypothesis 3: Raphaël Glucksmann as Socialist candidate: Express no voting intention: 15%

Recorded: 2026-10-14: Commission notice 10260 (PDF, 17 pp.) pp. 2-5 and 12-14 rendered as images and read; « Publié » (published) column = socio-demographic + 2022 presidential + 2024 legislative, among those certain to vote; cross-checked against the OpinionWay report for CNews (34 pp., pp. 9, 13, 17), identical. Margin cited = statement in the report under each scenario (the notice gives 1.6 to 3.7 points for 690 people).

Ipsos bva for CEVIPOF, Le Monde, Fondation Jean Jaurès, Institut MontaigneFieldwork 3–9 September 2026Sample of 13,060 people, of whom 9,445 registered on the electoral rolls, saying they are certain to vote, expressing a choice (hypothesis 1; « dont 8 984 à 9 445 certaines d'aller voter et exprimant une intention de vote selon l'hypothèse » [“of whom 8,984 to 9,445 certain to vote and expressing a voting intention, depending on the hypothesis”])online, via Ipsos's Access Panel; 13,060 registered voters, representative sample of the French population aged 18 and over, quota method (sex, age, occupation, region, type of urban area)Margin of error: published by score for each scenario, from +/- 0.1 point (0.5%) to +/- 1 point (33.5-35%); for the base of 9,445 people, +/- 0.8 point for a score of 20%Polling firm’s document · Notice (Commission des sondages)

Owned by:Ipsos bva: Ipsos, a listed company with no controlling shareholder; largest shareholder DT & Partners, Didier Truchot's holding company (11.4% of the capital, 19.3% of the voting rights), followed by Bpifrance (7.1%)CEVIPOF: Sciences Po research centre, no shareholderLe Monde: Fonds pour l'indépendance de la presse (72.5%, through Le Monde libre) and the Pôle d'indépendance of staff, readers and founders (25.4%)Fondation Jean Jaurès: foundation recognised as being of public utility, no shareholderInstitut Montaigne: association funded by more than 200 member companies (each under 1.2% of the budget), no shareholder

Hypothesis: R. Glucksmann, G. Attal, E. Philippe: base 9,445; no choice expressed among those certain to vote: 8%

Hypothesis: R. Glucksmann, E. Philippe: base 9,327; no choice expressed among those certain to vote: 9%

Hypothesis: R. Glucksmann, G. Attal: base 9,233; no choice expressed among those certain to vote: 10%

Hypothesis: O. Faure, G. Attal, E. Philippe: base 9,290; no choice expressed among those certain to vote: 9%

Hypothesis: O. Faure, E. Philippe: base 9,142; no choice expressed among those certain to vote: 11%

Hypothesis: O. Faure, G. Attal: base 8,984; no choice expressed among those certain to vote: 12%

Recorded: 2026-10-14: Commission notice (PDF, 69 pp.) read with pdftotext -layout, pp. 22 to 42 (one page of scores per scenario) cross-checked against the summary on p. 21 rendered as an image; full report found on ipsos.com/fr-fr (page « Présidentielle 2027 : le point à 7 mois du scrutin »). Enquête électorale française 2026, wave 3

OpinionWay for Le Journal du DimancheFieldwork 2–3 September 2026Sample of 1,003 people, of whom 680 people registered on the electoral rolls certain to vote (score 9 or 10), out of 929 registered votersself-administered online questionnaire on a CAWI system, quota method (sex, age, socio-professional category, type of urban area, region of residence)Margin of error: from 1.6 to 3.8 points for 680 people (from 1.4 to 3.2 points for 930 people)Polling firm’s document (notice filed with the Commission des sondages)

Owned by:OpinionWay: Les Echos-Le Parisien group (LVMH, Bernard Arnault), about 75% since 2022, with its managers (about 25%); sale to the managers announced in August 2026, not completed to our knowledgeLe Journal du Dimanche: Lagardère, controlled by Louis Hachette Group, whose largest shareholder is the Bolloré group (30.4%)

Hypothesis 1: Gabriel Attal as centre candidate: Express no voting intention: 16%

Hypothesis 2: Édouard Philippe as centre candidate: Express no voting intention: 15%

Hypothesis 3: Gabriel Attal and Édouard Philippe as candidates: Express no voting intention: 13%

Recorded: 2026-10-14: Commission notice 10252 (PDF, 18 pp.) pp. 2-7 and 12-14 rendered as images and read; « Publié » (published) column = socio-demographic + 2022 presidential, among those certain to vote, cross-checked column by column. url = notice: the opinion-way.com page for the survey (publications/...-2027-2-2026-24181, 7 Sept.) links to a PDF « ...Septembre-2026-1.pdf » which in fact contains the June report (fieldwork 10-11 June).

Ipsos bva for Le ParisienFieldwork 31 August – 2 September 2026Sample of 1,500 people, of whom 1,020 registered on the electoral rolls, saying they are certain to vote, expressing a choice (hypothesis 1; base specific to each hypothesis, from 979 to 1,020)online; 1,500 registered voters, nationally representative sample of the French population aged 18 and over, quota method (sex, age, occupation, type of urban area, region)Margin of error: published by score for each scenario, from +/- 0.4 point (0.5%) to +/- 3 points (34.5-36%); for the base of 1,020 people, +/- 2.5 points for a score of 20%Polling firm’s document · Notice (Commission des sondages)

Owned by:Ipsos bva: Ipsos, a listed company with no controlling shareholder; largest shareholder DT & Partners, Didier Truchot's holding company (11.4% of the capital, 19.3% of the voting rights), followed by Bpifrance (7.1%)Le Parisien: LVMH (Bernard Arnault), through the Les Echos-Le Parisien group

Hypothesis: R. Glucksmann, G. Attal and E. Philippe: base 1,020; no choice expressed among those certain to vote: 7%

Hypothesis: R. Glucksmann and E. Philippe: base 1,005; no choice expressed among those certain to vote: 9%

Hypothesis: R. Glucksmann and G. Attal: base 999; no choice expressed among those certain to vote: 9%

Hypothesis: O. Faure, G. Attal and E. Philippe: base 1,004; no choice expressed among those certain to vote: 9%

Hypothesis: O. Faure and E. Philippe: base 992; no choice expressed among those certain to vote: 10%

Hypothesis: O. Faure and G. Attal: base 979; no choice expressed among those certain to vote: 11%

Recorded: 2026-10-14: Commission notice (PDF, 32 pp.) read with pdftotext -layout, pp. 13 to 23 (one page per scenario) cross-checked against the summary on p. 12 rendered as an image; full report found on ipsos.com/fr-fr (page « Présidentielle 2027 : Marine Le Pen conserve une avance nette dans les intentions de vote »). Notice dated 4 Sept., release scheduled for 5 Sept. at 6 pm

Cluster17 for Le PointFieldwork 31 August – 1 September 2026Sample of 1,711 people, of whom 1,459 registered respondents having expressed a voting intention for first-round scenario 1 (1,459; 1,451 for scenario 2; 1,491 for scenario 3); effective sample after adjustment and weighting by likelihood to vote: 1,150, 1,131 and 1,211; 1,628 registered votersself-administered online questionnaire, quota method (sex, age, CSP (socio-professional category), size of municipality, region); socio-demographic and political adjustment (2022 presidential 1st round, 2024 European elections), weighting by probability of votingMargin of error: published table: for the usable sample sizes of the 1st-round scenarios (1,131 to 1,211 people), from 1.2 points (score of 5%) to 3 points (score of 50%)Polling firm’s document · Notice (Commission des sondages)

Owned by:Cluster17: its founder Jean-Yves Dormagen, chairman; breakdown of the share capital not publishedLe Point: Artémis (Pinault family)

Hypothesis 2: base: 1,451 respondents having expressed a voting intention; effective sample 1,131

Hypothesis 3: without Marine Tondelier; base: 1,491 respondents having expressed a voting intention; effective sample 1,211; Commission notice: Zemmour 3.3%

Recorded: 2026-10-14: Cluster17 report (23 pp., linked from cluster17.com, article of 7 Sept. 2026) pages 6, 11 and 16 rendered as images; scores = published values rounded to the nearest half-point; cross-checked against Commission notice 10251 (41 pp., « Redressé » (adjusted) column to one decimal place, pages 19-21 rendered as images): consistent, except Zemmour scen. 3 (notice 3.3%, report 3%)

Elabe for BFMTV, La Tribune DimancheFieldwork 26–28 August 2026Sample of 1,501 people, of whom 1,378 registered on the electoral rolls (« seules les personnes inscrites sur les listes électorales et ayant l'intention d'aller voter sont prises en compte » [“only people registered on the electoral rolls and intending to vote are taken into account”])Internet survey; representative sample of residents of metropolitan France aged 18 and over, quota method (sex, age, socio-professional category, region of residence and type of urban area)Margin of error: between 1.4 and 3.1 percentage pointsPolling firm’s document · Notice (Commission des sondages)

Owned by:Elabe: its partners, through the holding company LUZ (Bernard Sananès and Marie-Rose Beslin, co-managers); breakdown of the share capital not publishedBFMTV: CMA CGM (Saadé family), since 2024La Tribune Dimanche: CMA CGM (Saadé family), since it bought La Tribune in 2023

Hypothesis 1: Abstention, blank or spoilt vote: 14 (published in the same table, outside the total of votes cast)

Hypothesis 2: Express no voting intention: 14

Hypothesis 3: Express no voting intention: 15

Hypothesis 4: Express no voting intention: 16

Hypothesis 5: Express no voting intention: 15

Hypothesis 6: Express no voting intention: 13; Tondelier not tested

Recorded: 2026-10-14: Commission notice (PDF, 9 pp.); text extracted in jumbled columns, figures taken from pp. 7-8 rendered as images and then cross-checked against the order of the values in the text; elabe.fr page consulted (text consistent)

Toluna Harris Interactive for M6 and RTLFieldwork 18–19 August 2026Sample of 2,114 people, of whom 1,764 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Recorded: 2026-10-14: Commission notice 10239 (PDF, 71 pp.) read with pdftotext -layout, pp. 3 and 33-44; column text jumbled for hypotheses 1, 4 and 5: pages 33, 34, 42 and 44 re-read as images; cover dated « Août 2026 », publication = date of the tolunacorporate.com page (24/08/2026) and of the Commission's list; report checked as accessible (HTTP 200)

Cluster17 for Le Grand ContinentFieldwork 22–24 July 2026Sample of 1,506 people, of whom 1,207 registered respondents having expressed a voting intention for first-round scenario 1 (1,207; 1,209 for scenario 2); effective sample after adjustment and weighting by likelihood to vote: 1,032 (scenario 1), 1,052 (scenario 2); 1,404 registered votersself-administered online questionnaire, quota method (sex, age, CSP (socio-professional category), size of municipality, region); socio-demographic and political adjustment (2022 presidential 1st round, 2024 European elections), weighting by probability of votingMargin of error: published table: for the usable sample sizes of the 1st-round scenarios (1,032 and 1,052 people), from 1.4 points (score of 5%) to 3 points (score of 50%)Polling firm’s document (notice filed with the Commission des sondages)

Owned by:Cluster17: its founder Jean-Yves Dormagen, chairman; breakdown of the share capital not publishedLe Grand Continent: Groupe d'études géopolitiques, a research centre run as an association and based at the ENS, no shareholder

Hypothesis 2: base: 1,209 respondents having expressed a voting intention; effective sample 1,052

Recorded: 2026-10-14: Commission notice 10246 (PDF, 26 pp.) read with pdftotext -layout, pages 18-19 (results) and 4 (margins) rendered as images; « Redressé » (weighted) column used. No results report on cluster17.com (only an analysis page referring to Le Grand Continent): url = notice

Elabe for BFMTV, La Tribune DimancheFieldwork 9–10 July 2026Sample of 1,503 people, of whom 1,390 registered on the electoral rolls (« seules les personnes inscrites sur les listes électorales et ayant l'intention d'aller voter sont prises en compte » [“only people registered on the electoral rolls and intending to vote are taken into account”])Internet survey; representative sample of residents of metropolitan France aged 18 and over, quota method (sex, age, socio-professional category, region of residence and type of urban area)Margin of error: between 1.4 and 3.1 percentage pointsPolling firm’s document · Notice (Commission des sondages)

Owned by:Elabe: its partners, through the holding company LUZ (Bernard Sananès and Marie-Rose Beslin, co-managers); breakdown of the share capital not publishedBFMTV: CMA CGM (Saadé family), since 2024La Tribune Dimanche: CMA CGM (Saadé family), since it bought La Tribune in 2023

Hypothesis 1: Abstention, blank or spoilt vote: 14 (published in the same table, outside the total of votes cast)

Hypothesis 2: Abstention, blank or spoilt vote: 14

Hypothesis 3: Abstention, blank or spoilt vote: 14

Hypothesis 4: Abstention, blank or spoilt vote: 15

Hypothesis 5: Abstention, blank or spoilt vote: 13

Hypothesis 6: Abstention, blank or spoilt vote: 14

Recorded: 2026-10-14: Commission notice (PDF, 9 pp.); extracted text offset by one line, figures taken from pp. 7-8 rendered as images; elabe.fr page consulted (matching text: Le Pen 34 to 35.5%)

Verian for L'HémicycleFieldwork 8–10 July 2026Sample of 1,047 people, of whom 818 people having expressed a voting intention, out of 987 people registered on the electoral rollsself-administered quantitative online questionnaire on a panel, quota method (sex, age, occupation of the individual, region); socio-demographic and political adjustment (2022 presidential 1st round, 2024 European elections)Margin of error: from 1.5 to 3.4 points for 818 respondents (table on p. 3 of the notice); the table on p. 21 of the same notice gives 1.6 to 3.8 points for the same sample sizePolling firm’s document (notice filed with the Commission des sondages)

Owned by:Verian: Trilantic Europe fund, since it bought Kantar Public in 2022L'Hémicycle: quarterly review edited by Éric Revel; publisher and share capital not published in its legal notice

Hypothesis 1: A single list published, of 8 candidates (no Arthaud, Tondelier or Dupont-Aignan), whereas the notice states that « les hypothèses d'intentions de vote ont été posées dans un ordre aléatoire » [“the voting-intention hypotheses were asked in random order”]. Did not express a voting intention: 23%

Recorded: 2026-10-14: Commission notice 10225 (PDF, 22 pp.) read with pdftotext -layout; p. 17 (chart, « en % des votes exprimés » [“as a % of votes cast”]) and p. 18 (table, Total row) rendered as images, identical. url = notice: the only « intention de vote présidentielle 2027 » page on the veriangroup.com/fr site (site map) is dated 15 Dec. 2025 (a different survey).

Toluna Harris Interactive for M6 and RTLFieldwork 7–8 July 2026Sample of 1,837 people, of whom 1,592 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel) from 7 July (after the announcement of the verdict in Marine Le Pen's appeal trial and her declaration of candidacy on the TF1 evening news) to 8 July 2026, quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Recorded: 2026-10-14: Commission notice 10223 (PDF, 64 pp.) read with pdftotext -layout, pp. 3 and 33-42; column text jumbled for hypotheses 2 and 3: pages 33, 36 and 38 re-read as images; cover dated « Juillet 2026 », publication = date of the tolunacorporate.com page (08/07/2026) and of the Commission's list; report checked as accessible (HTTP 200)

Ifop for LCI and Le FigaroFieldwork 7–8 July 2026Sample of 1,075 people, of whom 984 people registered on the electoral rollsself-administered online questionnaire, after Marine Le Pen's declaration of candidacy on the TF1 8 pm news; quota method (sex, age, occupation of the respondent) after stratification by region and type of urban area; socio-demographic adjustment (INSEE 2020) and political adjustment (reconstructed past votes: 2022 presidential 1st round, 2024 legislative 1st round)Margin of error not published by the polling firmPolling firm’s document · Notice (Commission des sondages)

Owned by:Ifop: LFPI fund (majority since 2022), with the management team; its election surveys for Sud Radio are co-signed “Ifop-Fiducial” with Fiducial (Christian Latouche), owner of Sud Radio, with no known stake of Fiducial in its capitalLCI: TF1 group, controlled by Bouygues (about 44%)Le Figaro: Dassault group (Dassault family), since 2004

Hypothesis E. Philippe: 745 registered voters expressed a voting intention (624 among registered voters certain to vote); undecided: 25.7%

Hypothesis G. Attal: 751 registered voters expressed a voting intention (625 among registered voters certain to vote); undecided: 24.2%

Recorded: 2026-10-14: Commission notice (PDF, 13 pp., « Les intentions de vote à l'élection présidentielle 2027 et l'opinion des Français après la déclaration de candidature de Marine Le Pen ») read with PyMuPDF (word positions) and pdftotext -layout, pp. 8 to 9, the two readings agree (pp. 10 to 12 = second rounds, not included); Ifop report (ifopgroup.com, article of 8 July 2026) downloaded

OpinionWay for Le Journal du DimancheFieldwork 10–11 June 2026Sample of 1,001 people, of whom 683 people registered on the electoral rolls certain to vote (score 9 or 10), out of 901 registered votersself-administered online questionnaire, quota method (sex, age, socio-professional category, area of residence)Margin of error: from 1.8 to 4.0 points at mostPolling firm’s document · Notice (Commission des sondages)

Owned by:OpinionWay: Les Echos-Le Parisien group (LVMH, Bernard Arnault), about 75% since 2022, with its managers (about 25%); sale to the managers announced in August 2026, not completed to our knowledgeLe Journal du Dimanche: Lagardère, controlled by Louis Hachette Group, whose largest shareholder is the Bolloré group (30.4%)

Hypothesis 1: Gabriel Attal as centre candidate: Option offered: « Le candidat du Rassemblement national, Marine Le Pen ou Jordan Bardella » [“The Rassemblement national candidate, Marine Le Pen or Jordan Bardella”]. Express no voting intention: 16%

Hypothesis 2: Édouard Philippe as centre candidate: Option offered: « Le candidat du Rassemblement national, Marine Le Pen ou Jordan Bardella » [“The Rassemblement national candidate, Marine Le Pen or Jordan Bardella”]. Express no voting intention: 15%

Hypothesis 3: Édouard Philippe as centre candidate, no Communist or Green candidate: Option offered: « Le candidat du Rassemblement national, Marine Le Pen ou Jordan Bardella » [“The Rassemblement national candidate, Marine Le Pen or Jordan Bardella”]. Express no voting intention: 16%

Recorded: 2026-10-14: Commission notice 10205 (PDF, 15 pp.) pp. 3-5 and 12-14 rendered as images and read; « Publié » (published) column = socio-demographics + 2022 presidential among those certain to vote; cross-checked against the OpinionWay report (19 pp., pp. 6, 10, 14), identical. Margin cited = statement in the report under each hypothesis (the notice gives 1.6 to 3.7 points for 683 people).

Ipsos bva for Le ParisienFieldwork 27–28 May 2026Sample of 1,500 people, of whom 1,079 registered on the electoral rolls, saying they are certain to vote, expressing a choice (hypothesis 1; base specific to each hypothesis, from 1,037 to 1,079)online, via the Ipsos Access panel; 1,500 registered voters, nationally representative sample of the French population aged 18 and over, quota method (sex, age, occupation, type of urban area, region)Margin of error: published by score for each scenario, from +/- 0.6 point (1%) to +/- 2.9 points (34-36%); for the base of 1,079 people, +/- 2.4 points for a score of 20%Polling firm’s document · Notice (Commission des sondages)

Owned by:Ipsos bva: Ipsos, a listed company with no controlling shareholder; largest shareholder DT & Partners, Didier Truchot's holding company (11.4% of the capital, 19.3% of the voting rights), followed by Bpifrance (7.1%)Le Parisien: LVMH (Bernard Arnault), through the Les Echos-Le Parisien group

Hypothesis: R. Glucksmann, G. Attal, E. Philippe and J. Bardella: base 1,079; people certain to vote who did not express a voting intention: 7%

Hypothesis: R. Glucksmann, G. Attal, E. Philippe and M. Le Pen: base 1,073; no choice expressed among those certain to vote: 7%

Hypothesis: F. Hollande, G. Attal, E. Philippe and J. Bardella: base 1,060; no choice expressed among those certain to vote: 9%

Hypothesis: F. Hollande, G. Attal, E. Philippe and M. Le Pen: base 1,065; no choice expressed among those certain to vote: 8%

Hypothesis: R. Glucksmann, G. Attal and J. Bardella: base 1,050; no choice expressed among those certain to vote: 9%

Hypothesis: R. Glucksmann, E. Philippe and J. Bardella: base 1,057; no choice expressed among those certain to vote: 9%

Hypothesis: F. Hollande, G. Attal and J. Bardella: base 1,037; no choice expressed among those certain to vote: 11%

Hypothesis: F. Hollande, E. Philippe and J. Bardella: base 1,041; no choice expressed among those certain to vote: 11%

Recorded: 2026-10-14: Commission notice (PDF, 22 pp.) read with pdftotext -layout, pp. 10 to 17 (one page per hypothesis) cross-checked against the summary on p. 9 rendered as an image; full report found on ipsos.com/fr-fr (page « Présidentielle 2027 : les candidats du Rassemblement National largement en tête au premier tour »). Survey « Ipsos bva Cesi école d'ingénieurs pour Le Parisien »

Toluna Harris Interactive for M6 and RTLFieldwork 25–27 May 2026Sample of 2,068 people, of whom 1,744 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Recorded: 2026-10-14: Commission notice 10192 (PDF, 69 pp.) read with pdftotext -layout, pp. 1, 3 and 35-44; column text jumbled for hypotheses 2 to 5: pages 35, 38, 40, 42 and 44 re-read as images; cover « Etude réalisée pour M6 & RTL, Mai 2026 », publication = date of the tolunacorporate.com page (29/05/2026) and of the Commission's list; report checked as accessible (HTTP 200)

Odoxa for Mascaret, Public Sénat and 20 regional newspapersFieldwork 20–21 May 2026Sample of 1,005 people, of whom 681 people registered on the electoral rolls intending to vote and having expressed a voting intention (681 individuals in the first round); 928 registered votersFrench people surveyed via the Internet; representative sample (quota method: sex, age, occupation, after stratification by region and type of urban area)Margin of error: between plus or minus 1.8 and 4.1 points depending on the scorePolling firm’s document · Notice (Commission des sondages)

Owned by:Odoxa: its founders Gaël Sliman and Céline Bracq; no known outside shareholder, breakdown of the share capital not publishedMascaret: consultancy bought from Dentsu in 2021 by its managers; Entrepreneur Invest fund a minority shareholder since 2023Public Sénat: public parliamentary channel, funded by the Senatela presse régionale: about twenty titles from several regional press groups

Hypothesis 1: single hypothesis; did not express a voting intention: 23%

Recorded: 2026-10-14: Odoxa report (Baromètre politique May 2026, 42 pp., linked from the odoxa.fr page « Philippe s'effondre, Mélenchon décolle, Bardella s'impose ») read with pdftotext -layout and page 20 rendered as an image; Commission notice 10189 read

Toluna Harris Interactive for M6 and RTLFieldwork 28–30 April 2026Sample of 1,989 people, of whom 1,725 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel), quota method and adjustment (sex, age, CSP (socio-professional category), region, past voting behaviour)Margin of error: between plus or minus 1.4 and 3.1 points depending on the score (methodology); « entre plus ou moins 1,4 et 2,9 points » [“between plus or minus 1.4 and 2.9 points”] on the voting intention pagesPolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Recorded: 2026-10-14: Commission notice 10180 (PDF, 44 pp.) read with pdftotext -layout, pp. 1, 3 and 27-34; column text jumbled for hypotheses 1 and 2: pages 27, 28 and 30 re-read as images; cover « Avril 2026 », publication = date of the tolunacorporate.com page (04/05/2026) and of the Commission's list; report (wave 1) checked as accessible (HTTP 200)

Elabe for BFMTV, La Tribune DimancheFieldwork 25–27 March 2026Sample of 1,504 people, of whom 1,422 registered on the electoral rolls (« seules les personnes inscrites sur les listes électorales et ayant l'intention d'aller voter sont prises en compte » [“only people registered on the electoral rolls and intending to vote are taken into account”])Internet survey; representative sample of residents of metropolitan France aged 18 and over, quota method (sex, age, socio-professional category, region of residence and type of urban area)Margin of error: between 1.4 and 3.1 percentage pointsPolling firm’s document · Notice (Commission des sondages)

Owned by:Elabe: its partners, through the holding company LUZ (Bernard Sananès and Marie-Rose Beslin, co-managers); breakdown of the share capital not publishedBFMTV: CMA CGM (Saadé family), since 2024La Tribune Dimanche: CMA CGM (Saadé family), since it bought La Tribune in 2023

Hypothesis 1: Blank or spoilt vote, abstention: 15 (published in the same table, outside the total of votes cast)

Hypothesis 2: Blank or spoilt vote, abstention: 16

Hypothesis 3: Blank or spoilt vote, abstention: 15; neither Retailleau nor Zemmour tested

Hypothesis 4: Blank or spoilt vote, abstention: 19

Hypothesis 5: Blank or spoilt vote, abstention: 19

Hypothesis 6: Blank or spoilt vote, abstention: 18

Recorded: 2026-10-14: Commission notice (PDF, 8 pp.) read with pdftotext -layout, p. 7 (tables « Intentions de vote – 1er tour ») checked as an image; elabe.fr page consulted (matching text)

Odoxa for Mascaret, Public Sénat and 20 regional newspapersFieldwork 25–26 March 2026Sample of 1,299 people, of whom 853 people registered on the electoral rolls intending to vote and having expressed a voting intention (853 individuals in the first round for hypothesis 1, 847 for hypothesis 2); 1,206 registered votersFrench people surveyed via the Internet; representative sample (quota method: sex, age, occupation, after stratification by region and type of urban area)Margin of error: between plus or minus 1.5 and 3.5 points depending on the scorePolling firm’s document · Notice (Commission des sondages)

Owned by:Odoxa: its founders Gaël Sliman and Céline Bracq; no known outside shareholder, breakdown of the share capital not publishedMascaret: consultancy bought from Dentsu in 2021 by its managers; Entrepreneur Invest fund a minority shareholder since 2023Public Sénat: public parliamentary channel, funded by the Senatela presse régionale: about twenty titles from several regional press groups

Hypothesis 1: candidacies of Edouard Philippe and Bruno Retailleau: base: 853 individuals; did not express a voting intention: 25%

Hypothesis 2: Edouard Philippe as sole candidate for the centre and the right: base: 847 individuals; did not express a voting intention: 25%

Recorded: 2026-10-14: Odoxa report (Baromètre politique March 2026, 45 pp., linked from the odoxa.fr page of the article « Spectaculaire come-back d'Édouard Philippe ») read with pdftotext -layout and pages 25 and 27 rendered as images; Commission notice 10167 read (questions, bases, margins)

Toluna Harris Interactive for M6 and RTLFieldwork 22 March 2026Sample of 1,062 people, of whom 1,000 people registered on the electoral rollssurvey conducted online (Toluna-Harris Interactive proprietary panel) on 22 March 2026 from 8.10 pm, just after the announcement of the first results of the 2nd round of the municipal elections; quota method and adjustment (sex, age, CSP (socio-professional category), region, size of municipality, past voting behaviour)Margin of error: between plus or minus 1.8 and 3.8 points depending on the scorePolling firm’s document · Notice (Commission des sondages)

Owned by:Toluna Harris Interactive: Toluna, a group with no shareholder above 25%; largest shareholder Verlinvest (Belgian holding company of the founding families of AB InBev), followed by Macromill (17.4%) and its founder Frédéric-Charles PetitM6: M6 group, controlled by RTL Group (Bertelsmann, about 48%); CMA CGM (Saadé family) about 11%; sale relaunched by Bertelsmann in 2026, not completedRTL: M6 group since 2017, controlled by RTL Group (Bertelsmann)

Recorded: 2026-10-14: Commission notice 10165 (PDF, 31 pp.) read with pdftotext -layout, pp. 1, 3 and 20-23; column text jumbled: pages 20, 21 and 23 re-read as images; cover « Etude réalisée pour M6 & RTL, Mars 2026 »; tolunacorporate.com page « Municipales 2026 : Intentions de vote à la présidentielle 2027 » published on 22/03/2026, report checked as accessible (HTTP 200)