Today, we are going to check on some key lessons for us working in political campaigns, drawn from a small island nation that just delivered one of the more instructive referendum campaigns Europe has seen in years.



Lesson 1: Make the Abstract Threat Concrete



The Yes campaign, led by Prime Minister Kristrún Frostadóttir's government, had geopolitics on its side and used it. Trump's talk of annexing Greenland put Arctic security on the front page for weeks. The government argued that EU and NATO solidarity mattered more than ever, that a small nation needed friends, and that the króna's volatility made Iceland more exposed, not less, to a dangerous world. Iceland's inflation was running well above the euro area's. Interest rates sat close to 8% while the European Central Bank's stood near 2%. On paper, the economic argument for the euro should have worked in the government's favour too.



It did not land the way they hoped, because a rival economic argument was louder, simpler, and closer to home. Fisheries account for roughly 8% of Iceland's GDP and around 40% of its exports. A Gallup poll found 90% of fishing sector companies opposed EU membership, fearing the Common Fisheries Policy would open Icelandic waters to foreign trawlers. That is not an abstract sovereignty argument. It is a mortgage, a job, a town's entire economic base. As one Euronews analysis put it after the result, geopolitical framing was the tempting story for outside observers, but the actual driver of the vote was fish.



The lesson is not that geopolitics never matters. It is that campaigns have to translate an abstract threat into something voters can price into their own lives. The No campaign could point to boats, jobs, quotas and communities whose economic future appeared to be at stake. “European solidarity” had to compete with something voters could see from the harbour.



There is also a useful warning here for campaign teams reading international coverage of their own races. Foreign media often emphasise the dimension most interesting to their audience, not necessarily the issue actually moving voters. Never mistake the international narrative around an election for the voter motivation inside it.



Lesson 2: Distribution Is Built Before the Campaign Starts



I always check with enthusiasm and expectation to have the next “eureka” moment when writing these articles. My curiosity is always expecting the most clever strategy. But, ask what digital tools decided this referendum and the honest answer is less about clever targeting and more about who already controlled the channels. 



Morgunblaðið, Iceland's only remaining daily print newspaper and a major digital publisher, reaches close to 90% of Icelanders every week. It has been majority owned by fishing companies, including Samherji, since 2009. Its own former chief executive admitted the investors wanted to influence debate on the fishing quota system and EU membership. When the referendum was called, those same companies injected a further €4 million into the paper to keep it running through the campaign.



That is a digital and media strategy that took seventeen years to build and was simply switched on when the moment came, in exactly the way the ECI campaign against conversion practices spent a decade building supporter networks before its six day sprint to a million signatures. Iceland's referendum campaigns also face no donation limits and no requirement to disclose funding sources, so this concentration of media ownership operated with almost no transparency at all. 



The Yes side, by contrast, was described by its own supporters as cautious to the point of near silence. One government figure told reporters he had stopped reading the news because the debate had become so polarised.



The campaign lesson is to audit distribution before you audit content. Who already owns the channels your persuadable voters read, watch and trust? Which newsletters, media brands, creators, community groups and first-party lists can repeatedly reach them without buying every impression? And what has the opposition spent years building that you are trying to recreate six weeks before polling day?



A brilliant six-week content strategy cannot compensate for seventeen years of distribution infrastructure.



Lesson 3: Borrow Trust Locally, Not Prestige Internationally



We should have learned this by now: a high-profile foreign endorsement can easily matter more to journalists than to the voters it is supposed to persuade.



Before the vote, European figures did weigh in. Polish Prime Minister Donald Tusk voiced support for Icelandic membership. The EU's Enlargement Commissioner, Marta Kos, welcomed the referendum announcement. Neither appears to have moved a single Icelandic vote. The most quoted MEP reaction of the entire episode came after the result, not before it, when Finnish EPP member Mika Aaltola offered the most clear eyed line of the whole saga: "Icelanders did not vote against Europe, they voted against redundancy. When a country already has the single market, Schengen and open borders, the marginal value of a seat must outweigh shared control of its core resource."



That is a sharp piece of analysis. It is also useless as a campaign tool, because it arrived once the votes were already counted. The pattern across this referendum, and frankly across most European referendums, is that endorsements from Brussels or from friendly European capitals reassure people who have already decided and do almost nothing to persuade people who have not. Icelandic voters were choosing about their own fish, their own currency, their own coastline. What an MEP in Strasbourg thought about it was, to most of them, close to irrelevant.



For campaign teams, the question is not “Who is the most prestigious person we can get to endorse this position?” It is “Whose opinion reduces uncertainty for the voter we still need to move?” A Commissioner may generate press coverage. A fishing cooperative leader, local employer or trusted community figure may change a vote. Those are different objectives and should not be confused.



Lesson 4: Your Opponent’s Campaign May Have Started Years Ago



The most striking international dimension of this campaign was not European institutions leaning on Iceland. It was Icelandic money and ideas flowing outward first.



Fishing companies tied to Morgunblaðið’s ownership had, over the previous decade, donated to Brexit-adjacent European groups including the Alliance of Conservatives and Reformists in Europe and the now defunct Alliance for Direct Democracy in Europe. One prominent No campaign writer produced fisheries research for Vote Leave operative Lee Rotherham’s think tank. A Swiss former court president who has written that EU membership is a one-way street addressed a No rally organised by Áfram Ísland, the campaign group that ultimately won.



By the time Iceland’s referendum arrived, the arguments, framing and, in some cases, personnel had already been road-tested in the UK’s Brexit campaign a decade earlier. Cross-border ideological networks do not usually show up as an MEP publicly backing your campaign. They show up years earlier as funding relationships, shared researchers and imported framing that only becomes visible in hindsight.



The practical lesson is to map networks, not just campaign organisations. Track who funds whom, which researchers and strategists move between campaigns, which arguments are being tested elsewhere, and which organisations provide the infrastructure for ideas to travel across borders. By the time a successful frame appears in your own race, it may already have been tested, funded and refined somewhere else.



If you start studying the opposition when the election is called, you may already be years behind.



Lesson 5: Build the Counter-Disinformation Stack Before You Need It



Icelandic authorities went into this referendum expecting the same wave of AI generated misinformation that hit Malta and Cyprus earlier in the year. A viral claim that EU membership meant conscription into an "EU army" spread fastest among the least politically engaged voters, despite the EU having no standing conscript force. Cybersecurity officials detected what they described as coordinated automated activity during the campaign and could not trace its origin.



What was new was the institutional response. Rather than leaving fact checking to overstretched newsrooms, the University of Iceland built an AI assisted research tool called Evrópuvefur, drawing only on peer reviewed Icelandic academic sources. A separate independent project called ESB Vaktin used Anthropic's Claude to help generate fact checks, with every output reviewed by a statistician before publication. The Media Commission's Netvís centre opened a public portal in July specifically for reporting suspected disinformation and coordinated inauthentic activity ahead of the vote.



Notice the architecture: trusted source material, AI-assisted monitoring and drafting, mandatory human verification, and a public distribution channel. That is much more useful than simply telling a communications team to “monitor misinformation.” It is a repeatable workflow.



SOURCE → DETECT → DRAFT → VERIFY → DISTRIBUTE



None of this stopped the No side from winning, and it was never designed to settle the outcome. What it represents is a genuinely new category of campaign infrastructure, institutional and academic actors using the same AI tools that produce disinformation to build faster, more credible rebuttal capacity in real time. 



For a European political landscape that has mostly treated AI in campaigns as a production tool, as Labour did in Malta, this is the other half of the story: AI as public interest infrastructure, built before the crisis, not scrambled together during it.



Lesson 6: Audit the Language After Every Election



Listen to how Áfram Ísland's leadership talked about their win and the strategy becomes obvious. 



Halldór Benjamín Þorbergsson, the movement's founder, closed the campaign with a line that could have come from any self reliance movement anywhere, "God helps those who help themselves." Independence Party leader Guðrún Hafsteinsdóttir framed the result not as a rejection of Europe but as a domestic correction, telling reporters "the best way to heal the wounds is for the government to listen to the conclusion and put this path aside." Pressed on whether Brussels could help solve Iceland's economic problems, the same side was blunt, "these are not problems that Brussels will solve for us, these are problems that Icelandic politicians must solve."



Compare that with Prime Minister Frostadóttir's own concession, "we've been waiting for this for years to be able to have a vote," a sentence about process, not about people's lives. The winning campaign talked about mortgages, quotas, and who gets to decide, in the first person plural. 



The losing campaign talked about timelines, opportunities, and international positioning. Voters chose the side that sounded like it was solving their problem, not managing a foreign policy file.



Lesson 7: Win the Motivation Before You Win the Channel



Strip away the fjords and fishing quotas and Iceland leaves campaign professionals with one overriding lesson: distribution amplifies a motivational frame; it does not create one.



The No campaign had a frame built around control: control of fishing grounds, control of economic decisions and control of Iceland’s relationship with Europe. That frame could travel through established media, local economic interests and campaign networks because it translated an institutional question into a personal motivation.



That is the order campaign teams should work in:



MOTIVATION → FRAME → MESSENGER → DISTRIBUTION → CONTENT



Too many digital strategies start at the opposite end. Teams choose the platform, format and creative before answering the harder question of what actually motivates the voters they need to move.



Iceland’s referendum was not won because one side discovered a new platform or targeting trick. It was won by the side that connected an abstract institutional choice to a concrete motivation, had powerful distribution infrastructure already available, and communicated through language voters could attach to their own economic lives.



The tools matter. The channels matter. AI-assisted fact-checking matters. But they sit downstream from the decision that matters most: what motivation is actually capable of moving the voters you need?



In Iceland, the answer was control. The campaign that recognised it first won.