This post isn’t about the content of the advertising or the outcome of the Holborn and St Pancras by-election, but rather the sheer volume of digital material campaigns paid to put in front of voters during the campaign. The by-election wasn’t especially unusual (though Reform and Restore spent much more than parties with no chance of winning normally would), but instead it’s a good example of the growing gap between how we regulate elections and how campaigns actually operate.

As best we can (the data we can glean from the platform ad libraries is blurry), we think around £215,000 of digital advertising was bought by the various campaigns across a range of different advertising accounts (Meta, YouTube, candidate, party, local party and party leader pages).

If you take a fairly standard cost of £8 per 1,000 ad impressions for social media ads, that means 27M ad impressions were bought during the campaign. 

That’s the equivalent of 382 ad impressions for each of the 70,302 eligible voters in Holborn and St Pancras, or 10-15 per voter per day of the campaign.

Of course, not all these impressions will have reached registered voters. Some will have gone to commuters, visitors and people living outside the constituency. And some people will have seen far more advertising than others. But the scale is really extraordinary.

The cost of reaching voters

Let’s try to compare that to direct mail and leafleting.

A British parliamentary candidate is entitled to one piece of direct mail delivered, for free, by the Royal Mail, to either a household, or a named individual voter. They still have to cover the printing costs, but the delivery is free. If you’ve got everything perfectly lined up with the Royal Mail, your message could hit the doormats in about a week. 

If you want to send a second piece of direct mail, you’ll have the same production cost, plus the cost of postage to get it to the letterbox. That’s potentially tens of thousands of pounds to reach all the households a second time. With a campaign expenditure limit of £180,050 for your entire by-election campaign (much less for a normal parliamentary election), you’re going to need to really, really believe in the power of direct mail to send that second piece.

Which brings us to the next option – leaflets.

Leaflets and door knocking are the lifeblood of British political campaigns. They give your members something to do, while helping you gather data about how voters feel “on the doorstep”. Most importantly, canvassing conversations tell you who to “knock up” (unfortunate phrase) on polling day, in order to make sure they’ve voted.

Leaflets turn volunteers into an informal postal service. Give each of them a few streets to aim at and a couple of bundles of leaflets and you’re “getting the message out”, every evening and weekend for the duration of the campaign. Leaflets are pretty cheap to produce (often just a few pence each in large print runs, though can be more for more sophisticated designs and layouts) and can be “targeted”, in that you can focus your efforts on areas or even houses, where you know your voters, or those most likely to be persuaded, live. They’re also persistent, in that they still arrive whether or not someone is paying attention and can be delivered when no one is home. 

That said, they do take a long time to distribute, and they’re less likely to reach people with loud, barking dogs, letterboxes they never check or who are quick to recycle. 

Despite the limitations, the economics of traditional constituency campaigning are built around the physical work of reaching households. The typical way a general election candidate’s spending limit is set is by multiplying the number of voters by a figure (which varies depending on whether it’s an urban or rural constituency) and adding a chunk for fixed costs.

Volunteers vs Platforms

Delivering 10 leaflets to 10 households using volunteers takes around 10 minutes and costs perhaps 20–50p in printing, depending on the specification and print run.

Therefore, for a typical urban constituency like Holborn and St Pancras, of roughly 50,000 households, delivering 500,000 leaflets (10 per household) might cost £10,000–£25,000 to print, and 8-10,000 hours of volunteer time, assuming an average one leaflet delivery per minute. If we were to pay our volunteers a London Living Wage (we won’t and don’t), we’d have added up to £150,000 of implied labour cost to the campaign.

So even a modest programme of ten leaflet deliveries per household represents a significant logistical undertaking. It requires printing, sorting, organising volunteers and thousands of hours of walking around streets.

Now compare that with digital advertising. At £8 CPM, £200,000 buys 25 million impressions. That’s 50 times as many ad impressions as our 500,000 leaflet deliveries, without needing an army of volunteers to distribute them or a week to get the Royal Mail moving.

These aren’t equivalent contacts. A leaflet on a doormat isn’t the same as an ad appearing in someone’s feed. Neither guarantees a voter will read the message. But the difference in the scale and speed of distribution is enormous.

And yet our approach to election spending limits still behaves as if reaching voters means physically delivering things to them.

Inflation: The wrong measure

Before the last election, the Conservative government increased national party spending limits by 80% (to £34.1m for a party running in all seats across England, Wales and Scotland). By-election candidate limits also rose by 80%, while ordinary general election candidate limits increased by around 32%.

They were able to do this, without any broad public consultation, because the limits hadn’t been increased with inflation for many years. If there had been a debate, we’d have argued that inflation was the wrong measure to base it on. 

When you look at the cost of campaigning over time, the core inputs either have not risen at all (volunteers are still free, as is Royal Mail delivery for the first item) while others have risen much more slowly than inflation. The ONS says the cost of printing services has gone up by around 37% since 2006, compared with overall inflation of around 73%.

Added to this, digital advertising (let alone organic social media campaigning) has introduced an entirely new way of distributing campaign messages, at a scale that would have been difficult to imagine when the national spending limits were originally established by the Political Parties, Elections and Referendums Act of 2000. 

What are spending limits actually for?

The point of spending limits isn’t to preserve the purchasing power of a political party. It’s to constrain the amount of money parties can deploy to influence an election. These are completely different objectives.

If campaigning becomes cheaper, maintaining or increasing spending limits can allow parties to purchase substantially more political communication than they could previously. That’s something an inflation calculation alone cannot tell us.

In the light of bigger and bigger money coming into UK politics, we risk a situation where only the wealthiest parties can take full advantage of these higher limits. In 2024, Labour spent £30.1m, the Conservatives around £23m, Reform £5.4m and the Liberal Democrats £5.6m.

The problem isn’t simply that some parties have more money than others. That’s always been the case. It’s that higher limits create more room for the wealthiest parties to turn a financial advantage into an even greater one when communicating with voters.

If the UK is able to reshape spending limits based on the true cost of campaigning, rather than simply adjusting historic figures for inflation, using laws written before the internet age began, could make for a more competitive democracy.

Political parties absolutely do need to reach voters, explain their policies and challenge their opponents. But they should never have a virtually unlimited entitlement to buy voters’ attention.

The government has already asked the Electoral Commission to review spending limits by summer 2027. As the Commission undertakes that review, we suggest the first thing it considers is what it actually costs to reach voters in 2026, and how much more communication the same campaign budget can buy than it could 25 years ago.

That means looking beyond inflation and current spending levels and considering the actual costs of printing, postage, distribution, digital advertising and campaign technology. It also means asking how different spending limits affect the ability of smaller parties to compete.

We think such a line of enquiry would result in the spending limits shrinking substantially. But whatever the appropriate figures, they should reflect the economics of campaigning today, rather than the inflation-adjusted costs of campaigning in the past.

Because if the purpose of spending limits is to stop money overwhelming democratic competition, we want a better democracy than money can buy.

—–

On spending in Holborn and St Pancras

Estimated digital advertising expenditure during the campaign, based on platform ad-library data. These figures are approximate and represent advertising across different accounts, not verified statutory campaign expenditure.

The Green Party

• £99,279 on Meta, with all but 0.4% of it targeting postcodes in and around Holborn and St Pancras.
• Approximately £10,000 more from the London Green Party

Labour

• £25,000 on YouTube ads
• £25,000 on Meta ads
• Around another £10,000 on Meta from the main party account

Reform

• Roughly £20,000 on Meta

Restore

• Roughly £25,000 on Meta

Total = Approximately £215,000

All data is sourced via the platform ad libraries and our own tool, Who Targets Me Trends.

Note: This is another in an occasional series of speculative ideas about how we might make digital politics work better.

When parties and candidates buy digital ads on Facebook or YouTube, the comments are frequently full of people “debating” their content. The quality of arguments varies.

This engagement with political content provides platforms with signals about users’ interests and preferences, potentially informing which content and advertising they see in future.

Many of us in civil society, journalism and government (and, under some duress, the platforms themselves) also think political ads are worth preserving in transparency archives. The theory is that transparency constrains bad behaviour, through democratic norms and election laws, and allows opposing parties to see what their rivals are doing so they can respond.

The underlying idea is that ad libraries facilitate “counter-speech”. By making political arguments visible, they allow opponents to challenge them and offer alternatives. In theory, this is an unalloyed democratic good, with politics decided through a battle of ideas.

But the way this is delivered seems woefully insufficient, particularly in an era where so much of the content we see has no intermediary journalist or editor to help us contextualise information and decide whether it’s worth believing.

Worse, because of the way ads are presented to users on social media platforms, if you want to respond directly, it’s nearly impossible to find them in the wild. There are too many of them, and they aren’t targeted at you. Essentially, if you want to push back against the content of an opponent’s ad, using the comment feature, you can’t do it unless the message appears directly in your own feed.

So, what can be done?

Rather than attempting the Sisyphean task of independently fact-checking every political ad, we could give parties and candidates the privileged opportunity to respond to ads about them or the elections they’re contesting.

Simply put, they should have an automatic and highly visible right of reply.

For example, if a Republican candidate posts an ad about the policies or character of their Democratic opponent, that opponent gets to reply, and their reply is pinned at the top of the comments or even displayed immediately below. Voters can then debate whatever the issues are in the normal way below, but with important added context.

The point of this facility isn’t to give politicians yet another privilege, but to give voters a better opportunity to hear competing arguments before making up their minds. The principle isn’t new with broadcast regulators having long grappled with fairness and rights of reply in political speech. But so far we’ve done nothing to bring these principles into online advertising.

What about the practicalities?

The immediate question is “who” gets to reply. This isn’t as easy to manage as it sounds. Most formal political debates are extremely structured and rule-bound, and something similar should be true here.

Parties might get to nominate one account to use for their replies. They might be limited in who they can reply to (designated party accounts, rather than individual candidates). They might have to be of a certain size to qualify for the right-to-reply (after all, not every party makes it to the TV debate). They could get the same right-of-reply if they’re mentioned by non-party campaigners (e.g. PACs and third parties) too.

What might the consequences be?

The most foreseeable consequence is an explosion in comments by opposing parties. This could be a good thing.

It could make political advertising more substantive. If your ad makes an argument that’s trivial to rebut, you will probably think twice about posting it. And if you do choose to run it, you’d have to work through how your opponent might fact-check your statements, lay out positive arguments of their own or post that killer joke.

But the right of reply could also be a bad thing, as parties look for ways to lure their rivals into denying things that aren’t true (e.g. LBJ’s “pig fucker” politics).

The targeted ad then, hitherto seen as an unalloyed good from a campaigner’s perspective, suddenly becomes a double-edged sword. Buying an audience for your own argument also buys an audience for your opponent’s response.

There might also be a positive effect on political competition. A party spending £15m or $150m on advertising would, in effect, be helping distribute its opponents’ arguments too. That could modestly narrow the gap between well-funded campaigns and those with fewer resources, while reinforcing the idea that elections should be contests between competing arguments, ultimately settled by voters. Equally, a well-resourced party posting thousands of different ads might be able to overwhelm their poorer counterpart (in the past, we’ve argued that fewer ads would also be a good thing).

When it comes to political speech on the internet, there are always unintended consequences. Systems create incentives. Incentives change behaviour. And the consequences aren’t always predictable. Any proposal like this would need careful design and testing before being rolled out. Campaigners should be invited to wargame the system, looking for ways to exploit it, so the worst unintended consequences can be predicted and designed out.

The right of reply as a new intermediary

We’re all familiar with the rat-a-tat of partisan politics: two politicians across a desk, a journalist between them, each firing back and forth. It’s not perfect, and it can be undignified, but voters hearing both sides and making a judgement based on the conversation is generally a good thing.

The modern information environment isn’t like this. The political communication we see via ads looks much more like a speech or press statement, where the speaker comes out, makes their point, turns on their heel and heads back inside without taking questions.

Who Targets Me has spent nearly a decade trying to make political advertising more transparent. For millions of ads, we can now see who said what, to whom and how much they spent doing it.

The next challenge is trying to make sure that those who are spoken about can answer.

In October 2025, the European Union’s “Transparency and Targeting of Political Advertising” regulation (TTPA) came into force. At the same time, Meta and Google left the EU political advertising market, arguing the new law made things too operationally complex. 

We’ve argued before that the regulation over-stretched, adding “targeting” to what should have been a primarily transparency and advertiser verification-based remit. A more limited scope would likely have seen the platforms continue to offer political advertising services to all-comers in Europe.

This matters because mainstream political parties used these platforms much more than the far right did. Although German parties were never among Europe’s biggest buyers of digital political advertising, parties and other political actors still spent around €14 million on ads in the three months before the 2025 federal election. The AfD accounted for just 2.75% of political ad spending on the major platforms. The Greens spent around ten times as much, while the CDU and SPD each spent roughly five times more. When online political advertising effectively stopped, the biggest losers weren’t the extreme right, but parties in the political centre.

In Sachsen-Anhalt all parties had an effective digital ad spend and reach of zero. (Meta’s imperfect implementation of its own ban meant a few hundred ads mentioning the AfD, CDU and SPD still slipped through over the past months, but they generated little reach and very few originated from official campaigns.)

But what’s left when millions of paid political ads, one of the few transparent, attributable and verified campaigning tools, disappear?

Everything else.

Algorithmic feeds still reward emotionally charged content. Political influencers continue to reach millions. Highly engaged networks, domestic and foreign, remain able to reach voters without buying a single advertisement. 

These are also the parts of the online ecosystem in which parties willing to push emotionally charged messages, particularly around immigration, crime and national decline, have proved most effective. The immediate mutual gratitude between Ulrich Siegmund, the AfD’s main man in Saxony, and Elon Musk speaks to how highly these supportive channels are valued by the party (and its European counterparts). 

The TTPA clearly wasn’t designed to help the far right, but it should be judged by its effects. By removing a campaigning tool that mainstream parties relied on disproportionately, while leaving alone the mechanisms that reward polarising content, the regulation shifted the balance of online campaigning in favour of the AfD and other European far-right parties.

There still seems to be little appetite in the EU to revisit the regulation to make it fairer. Even if that happened, it’s not obvious that Meta and Google (let alone others) would re-enter the political advertising market. 

Our argument isn’t that unrestricted political advertising should return. Setting a baseline for political advertising transparency is an extremely valuable thing to do. But in trying to solve one problem too many (the way ads were targeted, over and above what is allowed by the GDPR), the overall situation has been made worse.

The ironic result is that Europe now has less transparency in digital campaigning than much of the rest of the world, while making life easier for the parties that depend least on paid political advertising.

Digital political advertising carries well-documented risks — opaque funding, misuse of personal data, and the amplification of mis- and disinformation. It also offers clear benefits: campaigns can reach voters directly, efficiently, and at scale, which is especially valuable in an algorithmically mediated information environment that often favours extreme content. Over more than a decade of study, our research has found that the benefits of digital political advertising generally outweigh the costs, but this gap could be (positively) widened with specific policy interventions.

Government should use legislation to create an online political advertising market with the following features:

1/ Transparency: All political advertisers should register and verify their UK identity and disclose their funding sources. Near-real-time reporting should provide details on ad content, targeting parameters, spend, and impressions. Archiving ad content ensures accountability over time. Transparent, accessible data allows regulators, media, and civil society to scrutinise campaigns effectively, prevent “shadow” spending, and enforce rules consistently. Transparency rules should be “always on”.

2/ Trustworthiness: Verification of advertisers ensures that most digital advertising comes from UK-based actors with a verifiable track record in politics, standing or likely to stand candidates, and with known financial backers. Polling-day advertising restrictions align with broadcast rules, preventing last-minute misuse and supporting electoral integrity. Trustworthiness measures, reinforced by transparency, build voter confidence that campaigns are legitimate, auditable, and compliant with spending limits.

3/ Competition: Common standards for transparency and trustworthiness make it easier for new services to offer political advertising to parties and candidates. Regulation should actively support opportunities for political advertisers beyond the two (current) largest digital platforms (Meta and Google).

4/ Plurality: Voters should encounter multiple candidates and viewpoints. Digital advertising spending caps ensure campaigns are treated fairly and no single campaign locally or nationally is able to dominate the information space.

This last point is crucial. At the 2024 election, party and candidate spending limits were significantly raised. These new ceilings allowed digital advertising spending to reach record levels and enabled the largest parties to vastly outspend smaller opponents on digital. Without intervention, there is a clear risk of an escalating arms race in digital political advertising.

We propose a framework of pragmatic, proportionate, and future-proof spending caps:

For party campaigners:

“No more than one third of regulated campaign spending, or £1 million, whichever is higher, may be spent on paid digital political advertising, subject to an absolute maximum of 50%.”

For candidates:

“A candidate’s campaign may spend up to £5,000 on digital advertising, or one third of their spending limit, whichever is higher.”

For by-elections:

“Digital advertising spending by candidates is capped at £20,000.”

For national parties targeting individual seats:

“A party campaigner can spend no more than the candidate limit on ads targeting a specific seat (or equivalent multiples of that limit if the ads target more than one seat).”

For non-party (third party) campaigners: 

“A registered non-party campaigner may spend no more than £10,000 or 50% (whichever is higher) of their total regulated campaign spending on digital advertising.”

And:

“A registered non-party campaigner may spend no more than £2,000 on digital ads targeting a specific seat (or equivalent multiples of that limit if the ads target more than one seat).”

And:

“Non-party campaigners below the £20,000 registration threshold (£10,000 in Scotland, Wales or Northern Ireland), can spend no more than £10,000 on digital advertising and no more than £1,000 on digital ads targeting a specific seat (or equivalent multiples of that limit if the ads target more than one seat).”

Taken together, these measures:

• Prevent unlimited and escalating digital advertising spend
• Protect smaller and digital-first campaigns by giving them a proportionally higher allowance
• Preserve the balance between national and constituency-level campaigning, which is central to the UK’s electoral system
• Ensure the primacy of candidates and parties in UK campaigns with comparable and fair limits on non-party campaigners
• Increase confidence in a system the public distrusts

Conclusion

By adopting proportionate spending caps alongside transparency and trustworthiness requirements, the UK can maintain the benefits of digital political advertising while ensuring a level playing field, promoting fair competition, and safeguarding voter trust.

The framework we propose is simple to enforce, scalable, and adaptable to future technological and campaign innovations, providing a strong, practical foundation for a transparent, trustworthy, and pluralistic political advertising ecosystem.

There are four challenge areas to overcome:

1/ Opaque funding: Voters cannot easily see who is paying to influence them

2/ Loss of territorial control: UK elections are increasingly exposed to foreign actors and cross-border influence

3/ The collapse of the money > reach link: Digital communication has radically reduced the cost of reach, meaning spending limits no longer effectively constrain influence.

4/ Changing media: Regulation still rests on broadcast/non-broadcast distinctions that no longer reflect how voters encounter political messages.

Meeting these challenges requires a single, coordinated reform package across campaign finance, digital transparency, and media regulation – a “big bang” to restore public trust. 

We hope the forthcoming Elections Bill can rise to meet the challenge.

Campaign finance

Campaign finance is the most important pillar. The Government should:

1/ Reduce campaign spending limits which, following increases in the last parliament, are now too high. Current limits have two negative effects. They (a) make parties excessively dependent on large individual donations; and (b) entrench a divide between “haves” and “have-nots”, where only the best-funded parties can campaign fully. Lowering the limits would reduce large donor dependency and improve perceived fairness. As part of this, the Government should consider spending caps for social media advertising, which has got out of control.

2/ Cap individual and corporate donations at around £1 million per election cycle. £1m is a significant contribution, but limits outsized influence. All donors (or the owners of the companies in question) should be registered to vote and straightforwardly tax-resident in the UK. 

3/ Shift towards a greater proportion of public funding for political parties, allocated on the basis of past election results. This would offset the impact of lower donation caps while tying public funding to democratic support.

4/ Introduce “always-on” political spending reporting, making it as close to real time as possible. New digital tools, housed within the Electoral Commission, should automate and standardise reporting, reducing compliance burdens while dramatically increasing transparency.

5/ Rebalance national and constituency spending limits. High national limits, combined with modern data-driven targeting, allow parties to concentrate their resources on a small number of seats by drawing on very large national budgets. This concentrates influence while weakening accountability. 

Digital campaigning and media regulation

Digital campaign regulation is no longer fit for purpose. The Government should:

6/ Require digital platforms, including AI services, to provide real-time, privacy-preserving random samples of political content (posts, ads, and prompt-response pairs), enabling regulators, researchers and the press to assess the prevalence of narratives and potentially systemic risks as they emerge.

7/ Mandate platforms provide real-time public leaderboards of highest-reach and highest-engagement political content in the UK, allowing journalists, researchers, and voters to understand which messages are actually shaping public discourse.

8/ Radically update the system of Party Political Broadcasts to reflect modern media consumption. This should extend the principle of universal access to voters through “must-carry” obligations for streaming platforms and/or free or low-cost political advertising quotas on major social media platforms.

9/ Set clear, statutory transparency and verification standards for political advertising (on- and offline), while ensuring these requirements are technically workable. This should include financial, content and targeting transparency, and be made available via Ad Libraries and APIs. Regulation should be (within reason) co-designed with providers to avoid the unintended consequence of de facto bans on lawful political advertising (as we have seen in the EU).

10/ Introduce explicit transparency rules for influencer marketing and paid “verified/blue tick” programmes, with meaningful penalties for accounts, parties, or platforms that fail to comply. Platforms should bear responsibility (within reason) for enforcement.

11/ Align the campaign quiet period across all media, extending existing broadcast principles so that election day messaging is limited to “get out the vote” activity, with no new persuasive campaign messages published online.

12/ Ban all foreign political advertising, including advertising delivered via paid verification schemes, which function as a form of unwanted political interference.

13/ Address coordinated “organic” political campaigning. Platforms should have a statutory duty during election periods to detect and expose coordinated inauthentic or misleading political behaviour, working with researchers and regulators to counter it.

14/ Extend regulation to third-party and issue-based digital campaigning. Update the definition of regulated political content to include digital activity that is reasonably likely to influence electoral behaviour, even where it does not explicitly promote a named party or candidate. This will help avoid actors trying to circumvent the rules.

15/ Introduce a new cross-platform election integrity obligation. Require major digital platforms to cooperate during election periods through standardised metadata, secure information-sharing, and coordinated reporting to the Electoral Commission. Campaigns are cross-platform, and regulation should account for this.

16/ Set clear rules for synthetic and AI-generated political content. Platforms should be required to support provenance and labelling standards for synthetic political media during election periods, with enhanced penalties for undisclosed impersonation of real individuals.

Improve implementation, enforcement and accountability

17/ Strengthen enforcement and sanctions for digital campaign violations. Ensure new transparency and integrity rules are backed by credible, timely enforcement. This should include escalating fines proportionate to campaign spend or platform turnover, expedited enforcement powers during campaign periods, and personal accountability for responsible officers in cases of systematic or deliberate non-compliance.

18/ Ensure the Electoral Commission has the powers and capacity to deliver. The proposed reforms above will require explicit statutory powers, adequate funding, and in-house technical capability for the Electoral Commission, including data science and platform oversight expertise.

19/ Support independent journalism and civil society. Recognise the critical role of investigative journalism, fact-checkers, and civil society organisations in holding campaigns, parties, and platforms accountable. The Government should provide sustained support for independent monitoring, including access to public datasets, platform transparency tools, and funding for capacity-building initiatives. Strengthening this ecosystem ensures that oversight extends beyond regulators and enhances public trust in the electoral process.

20/ Actively build societal support for democracy. Polls show support for democracy is in decline. This is partly a result of successive political failures, partly due to persistent efforts to undermine it but also due to an unwillingness to foster and invest in renewing it. This must go far beyond the usual policy tool of “changing the school curriculum”, into a much wider, better funded and more consistent effort, ideally with cross-party support.

Conclusion: Restoring trust

Taken together, these reforms would modernise important aspects of election law for a digital, global media environment – restoring public confidence without suppressing political debate. Too much is changing, too quickly, and the UK has (along with many other governments around the world) fallen behind. Incremental change isn’t enough. The Government should act decisively, through the Elections Bill, to safeguard UK democracy.