Playing on difficult mode

An exclusive new mega-poll sets out the challenge for Labour

In the run up to the 1997 election, Tony Blair’s Labour had a clear political strategy: winning over aspirational middle class voters in marginal constituencies. They could focus on that because their existing seats were genuinely safe. Neither the Tories or Lib Dems were going to threaten in city centres, northern industrial heartlands, Scotland or Wales.

In 2026 things look very different. There are no safe seats. Those city centres are shifting Green and the northern heartlands Reform. Nationalists are winning in Scotland and Wales.

Which is one reason many Labour MPs were somewhat frustrated by Blair’s latest intervention. His much discussed 6,000 word essay didn’t acknowledge this change in context at all. As party leader he was able to pursue a centrist platform (albeit one to the left of his new manifesto) because it fit with a viable political strategy. The challenge now is far harder. Asserting that the party just needs to identify “the right answer” and then persuade people of it is disingenuous because it assumes agreement on underlying values between voters with different worldviews. Governing reactively, without any guiding principles, is a terrible idea. But so is ignoring the interests of the people you need to vote for you.

The challenge for whoever leads Labour next will be to pursue a coherent governing project that also acknowledges this very different political context. To help them do that I teamed up with Convergent Opinion to poll 10,000 people about how they voted in the local elections and why. This allowed them to build an MRP looking at the shift in voting behaviour between 2024 and 2026 in every council that held elections, down to ward level. Unlike most MRPs this isn’t trying to predict the future but looks back at something that’s happened, so it can be calibrated against actual results.

As well as looking at how people voted compared to 2024, we asked why those who switched party did so, which other parties they considered, and what would make former Labour voters switch back. It’s the most comprehensive dataset available on voter shifts between parties and the reasons for them.

I’m going to start by looking at three very different councils Labour lost. Doing so makes it clear that simply trying to appeal to one type of lost voter isn’t going to work. But it also shows that there are ways to frame an agenda that would deal with real problems facing the country, while also appealing to different groups of defectors. In the second half of the post I set out what that might look like.

At the end of the post is a link to a dashboard which lets you look at detailed voter movement analysis for every council and ward that held elections.