Campaigning Without Paid Social: The EU experiment

In brief

What fills the space paid social advertising left is reach that cannot be bought. Public broadcasters must give parties free airtime, allocated equally but scheduled mostly outside prime time, so its reach is limited. News coverage and debate invitations go to parties with seats or strong polling, not to whoever tries hardest. The rest is unpaid posting, which costs staff time rather than media budget. The money was the smaller half. In Czechia the entire platform market came to 25.8% of one contestant's legal spending cap. Nobody has measured what any of this does to a vote.

How to use this

Before promising to campaign mainly through social networks, budget for the cost of production rather than placement, meaning the staff, video and design time that make something worth sharing, plus a check on what actually reached anyone. Ask who sets the polling threshold that decides debate invitations, and on whose polls. When someone quotes a per-party airtime figure, check whether it assumes every list took an equal share, and remember the blocks mostly air outside prime time. Treat every claim about the channel mix as a national rule or an observer's impression, because no budget-level measurement exists. Do not describe any of it as an effect on votes, because none has been estimated.

What the story is about

Czechia voted on 3 and 4 October 2025, days before the EU's political advertising rules applied and before Meta stopped selling political ads in Europe (European Union, 2024). That timing makes it the last election in this article where parties could still buy space on the big platforms, and the only one where the whole chain sits on the record in an observers' report. Every party the observers met said it relied on paid online advertising, and over the three months before election day the main competitors spent an estimated CZK 23.2 million on Google and Meta. One contestant in that election was legally capped at CZK 90 million. The entire seven-party platform market, in other words, came to 25.8% of a single contestant's legal ceiling. Paid broadcast was never part of that market anyway. In Czechia, paid political advertising is allowed only in print and online media (OSCE Office for Democratic Institutions and Human Rights, 2026a).

What replaced it was free, equal, and limited in reach. Between 17 September and 1 October, public television and public radio each broadcast 14 hours of free airtime, divided equally among the 26 registered candidate lists (OSCE Office for Democratic Institutions and Human Rights, 2026a). On those two figures, each list received about 32 minutes on television and the same again on radio. The report does not print the per-list minutes, and that number assumes every list took an equal share. Equal is not the same as effective. The blocks aired back to back and mostly outside prime time, and the observers judged their reach to voters limited.

Paid television advertising was already off the table in several European democracies long before the platforms closed. Peer-reviewed research from 2019 found that television political advertising had been banned outright during elections in many European democracies for decades, while online rules differed from country to country (Dobber, Fathaigh and Zuiderveen Borgesius, 2019). A Council of Europe comparative study (Furnémont and Kevin, 2020), reference material rather than a tested finding, states: "There is no common or harmonised approach in Europe to the regulation of political advertising." Its 2020 country list has the broadcast ban outright in Ireland, France, Belgium, Portugal, Switzerland and the UK, and allows paid broadcast only in a pre-election window in Germany and, for local broadcasters, in Italy. Ireland puts the ban in primary law, and the one route left open is a party political broadcast that must be allocated fairly rather than sold (Ireland, 2009).

The Netherlands voted on 29 October 2025, weeks after Meta's ban took effect, and the observers there found access uneven. "Traditional media remained an important means of campaigning, but access was uneven." What sorted parties was not effort. "Parties with parliamentary representation or strong public opinion polling generally received greater media coverage and more frequent invitations to media events than smaller or non-parliamentary parties" (OSCE Office for Democratic Institutions and Human Rights, 2026c). Coverage and invitations cannot be bought at all. They are given, and the test for who receives them is seats or poll standing, which the platform exit covers. Hungary arranged things differently for its 12 April 2026 election. There the public broadcaster had to carry political ads free and on equal terms, commercial broadcasters could choose whether to offer airtime, and only one said it would (OSCE Office for Democratic Institutions and Human Rights, 2026b).

Nothing measures what happened to the money once the platforms closed. No study takes a post-October-2025 campaign budget and splits it across television, print, outdoor, direct mail, online display, paid creators and field, and our article the platform exit records the same gap. Every claim in this article about the mix is therefore either a national rule or an observer's impression. The one hard number is the Czech CZK 23.2 million, and that is Who Targets Me's estimate, carried in a footnote, covering the main competitors only. Treat it as an order of magnitude, not an audited total. No causal estimate exists of the change's effect on any election result. Prior research on how little campaigns move votes points the same way, as our article on minimal effects sets out: any such effect would be small and hard to detect, and no study has tried.

The money was the smaller half of the story. The Czech numbers settle that: the whole platform market there was small next to what one contestant alone could legally spend. The stronger claim, that parties lost access rather than a budget, is built from the gates. Some gates are outright prohibitions, such as Ireland's ban on political advertising on broadcast. Others are thresholds that decide who gets a debate invitation or a mention in the news. Nobody has priced what parties put in place of the closed channel, so that claim remains an argument. It is a good argument, and it is not a measurement.

So the channels that carry a European campaign now are the ones nobody buys. Free airtime on public broadcasters, which the law allocates equally and the schedule pushes out of prime time. News coverage, which goes to parties voters already know. Unpaid posting on social platforms, which costs staff time instead of media budget. None of these was designed as a replacement for paid social, and none is new. What changed is that they are no longer the backdrop to a bought campaign. They are the campaign, and no study measures what they do to a vote.

So what

How far this travels is worth being honest about. The picture comes from Czechia, Ireland, the Netherlands, Hungary and Cyprus: five member states that happened to hold an election with an observer report in the window, not a sample of 27, and all in one cycle. Read it as national rules and observer impressions, not a trend. The questions still open are specific. Who sets the polling threshold for a debate invitation, and on whose polls? Does a free, equal, off-peak airtime block change anything at all? Did field contact rise, the question our article what a contact is worth deals with? Where did the money go? And once it is populated, will the EU's own ad repository restore a cross-country baseline, which our article the transparency gap covers?

For political parties

Parties that tell observers they will campaign mainly through social networks should read that promise closely. In Cyprus, which votes on 24 May 2026, contestants told the observers' needs assessment mission they "will campaign primarily through social networks", alongside traditional methods, and said the traditional side was held back by insufficient money, a complaint that smaller and newly-established parties made loudest (OSCE Office for Democratic Institutions and Human Rights, 2026d). With paid social gone, that plan can only mean unpaid posting. So the budget line that decides how well it works is not placement. It is production: the staff, video and design time that make something worth sharing, and the check on what actually reached anyone. A channel nobody measures cannot be corrected.

For government

For a regulator, the cheapest useful act is also the dullest: publish the map. The only cross-country list of broadcast political advertising bans is from 2020, in a Council of Europe study (Furnémont and Kevin, 2020), and only its Irish entry has been checked against the statute since. No comparable table covering all 27 member states has been published after 2022. That leaves parties, journalists and voters working from a picture taken before the online market closed. A current table would let a party planning a campaign know what it may buy where, and let a citizen see who is allowed to pay for attention and who has to earn it. Rules nobody can compare are rules nobody can audit.

References

Dobber, T., Ó Fathaigh, R. and Zuiderveen Borgesius, F.J. (2019) 'The regulation of online political micro-targeting in Europe', Internet Policy Review, 8(4). Available at: https://doi.org/10.14763/2019.4.1440 (Accessed: 11 September 2026).

European Union (2024) Regulation (EU) 2024/900 of the European Parliament and of the Council of 13 March 2024 on the transparency and targeting of political advertising, OJ L, 2024/900, 20.3.2024. Available at: https://eur-lex.europa.eu/eli/reg/2024/900/oj/eng (Accessed: 11 September 2026).

Furnémont, J.-F. and Kevin, D. (2020) Regulation of political advertising: a comparative study with reflections on the situation in South-East Europe. Strasbourg: Council of Europe. Available at: https://rm.coe.int/study-on-political-advertising-eng-final/1680a0c6e0 (Accessed: 11 September 2026).

Ireland (2009) Broadcasting Act 2009, section 41. Available at: https://www.irishstatutebook.ie/eli/2009/act/18/section/41/enacted/en/html (Accessed: 11 September 2026).

OSCE Office for Democratic Institutions and Human Rights (2026a) Czechia, parliamentary elections, 3 and 4 October 2025: final report. Warsaw: ODIHR, 1 July. Available at: https://odihr.osce.org/odihr/666202 (Accessed: 11 September 2026).

OSCE Office for Democratic Institutions and Human Rights (2026b) Hungary, parliamentary elections, 12 April 2026: interim report (26 February to 25 March 2026). Warsaw: ODIHR, 27 March. Available at: https://odihr.osce.org/odihr/662971 (Accessed: 11 September 2026).

OSCE Office for Democratic Institutions and Human Rights (2026c) The Netherlands, early parliamentary elections, 29 October 2025: ODIHR election expert team final report. Warsaw: ODIHR, 11 March. Available at: https://odihr.osce.org/sites/default/files/documents/official_documents/2026/03/NL_Early%20Parliamentary_2025_EET_final%20report_11.03.2026_0.pdf (Accessed: 11 September 2026).

OSCE Office for Democratic Institutions and Human Rights (2026d) Cyprus, parliamentary elections, 24 May 2026: needs assessment mission report. Warsaw: ODIHR, 18 March. Available at: https://odihr.osce.org/odihr/662848 (Accessed: 11 September 2026).