Most political innovation dies because people either over-build or never test. A lot of my work has been about creating practical ways to experiment inside campaigns and institutions with limited time, money and appetite for risk building evidence rather than anecdotes, and adapting experimentation methods to the constraints of UK politics (short campaigns, small budgets, electoral law, nervous stakeholders).

Through Campaign Lab I've helped run what we believe is the largest political RCT in UK history, 101,600 households across 8 constituencies, analysing data from over 600,000 electors, and the first RCT ever run during a UK General Election. But some of the most useful lessons come from smaller, scrappier tests, and especially from the surprises:

• Null results are findings too. Another mailer RCT found neither locally nor nationally framed issue leaflets moved turnout at all — knowledge that saves campaigns real money on print runs that don't work.
• Cheap observational tests can be surprisingly revealing. A 900-leaflet QR code pilot got a 7.3% scan rate (nearly double the industry average) and showed that 25 of 27 school-petition scans came from within 400m of the school — hyperlocal relevance driving engagement, mapped street by street.
• Experiments often tell your something unexpected. A petition RCT with 28,000 supporters found explicit personalisation lifted signing by 2.5 points — but asking supporters to recruit others did nothing, because just 4% of people generated all the peer recruitment. Relational campaigning strategies built on "everyone shares" quietly fail.

I can help fellows think about how to run a political experiments: choosing hypotheses, defining success, running pilots, evaluating outcomes and deciding whether something should scale. This includes campaign messaging tests, organising interventions, behavioural experiments, tool adoption and operational workflow changes. I’m particularly interested in helping people build evidence rather than anecdotes and in adapting experimentation methods to the constraints of UK politics.