Case Study: Advance Together, a new wave of politics

Advance Together was set up in the wake of the Grenfell tragedy, with the aim of creating a political party for people, not politicians. It wanted to address the issues facing the Royal Borough of Kensington & Chelsea by having local people represented by local people, with a new form of politics. This is how we worked with them in the run-up to the May local elections. 

THE BACKGROUND

Kensington and Chelsea has some of the biggest disparities between rich and poor in the country;

1. It has the highest gap between the median and mean salaries in the country, with a £90,000 difference between the two

2. In Notting Dale, 34% of children live in poverty, compared to <10% in most other wards

This, combined with Grenfell, proved that something had to be done. People were fed up with the status-quo;

1. Recent events at the 2016 General Elections saw Jeremy Corbyn’s Labour on the march, with a 10% swing in votes from Tory to Labour and the first Labour MP sitting in the seat since its creation

2. Royal Borough of Kensington & Chelsea became a battleground at Local Elections; Labour and opposition parties looked to exploit Tory weaknesses, whilst the Tories desperately sought to maintain influence

OUR OBJECTIVES

Starting from scratch, Advance had three clear objectives to measure success by:

1. Change the discussion around local politics: Instead of being driven purely by national issues, it wanted to bring the conversation back to local issues

2. Get local people engaged in the discussion and issues in the borough: Grenfell had highlighted just how out of touch the Council had become from many of the ordinary voters. Advance wanted to increase engagement with the political system especially in the poorest boroughs where turnout was under 30% in 2014

3. Get candidates elected into office: the initial aim was to win between 1-3 seats across the borough

WHAT WE DID IN SOCIAL

The diverse nature of Kensington and Chelsea’s different wards, combined with a limited social budget, meant that a blanket paid-for social approach wasn’t going to be possible. To make sure the right creatives were being delivered to the right wards, we:

1. Tested the response rate across every ward in the borough, narrowing down the wards we then tested variables with from fifteen, down to three

2. Tested party messages vs. topic-led posts with the three wards, to understand what resonated best, discovering that topic-led content performed 40% better on a CPM basis

3. Having discovered that topic-led content performed best, we set out to find which of the campaign topics, out of Schooling, Housing, Governance, Accountability and Air Quality, would resonate best. We found that there was little variance between topics

4. Tested which formats (video vs. image) performed best. As we wanted to get the party message ‘out there’, we prioritised CPM over CPE for these tests, discovering that video-based content was more effective than image-based posts

RESULTS & LEARNINGS

Campaign Results

1. Across the campaign, we reached 184,900 individuals, within the targeted area. Given that the 2011 census put the populations the borough at 160,000, we are confident that nearly everyone in the borough would have seen an Advance Together post in the run-up to the elections

2. People watched our content for a cumulative time of 117 hours

3. 1,000 click-throughs to the Advance Together website

4. Voter turnout increased by an average of 9% across the ward

5. On election day, Advance took 9% of the vote across the wards that they ran in. An impressive result but unfortunately not enough to break the two-party dominance of the current council

Provable effect on votes cast

In the aftermath of the election, we wanted to know if we could see any relationships between vote share and seeing one of our adverts online. Because the vote tallies for each ward were published we had a great data set to look into the effectiveness of our campaign on driving vote share

1. We ran a linear regression model to see if views, engagements, or reach helped drive votes

2. We found that there was a significant relationship between vote share and reach (significant at the 0.01 level) – essentially a statistically provable relationship between social media outcome and electoral outcome

Shaping the debate: what did we learn about who took part in the conversation?

We explored whether Advance was a significant part of election debate by tracking the conversation throughout April and May and found that:

1. The official Advance account was second-most influential within this conversation

2. Three of the top 12 accounts were Advance candidates

3. The appearance of several personal accounts within the top 12 reiterated the need for candidates to push the party message, demonstrating that they are the most likely to get cut through within this conversation

4. We were also mentioned by some prominent celebrity Labour supporters, the most notable being Lily Allen, who had a significant network effect. This Tweet was one of the most discussed and most influential and shows how well Advance did in gaining cut through

Below is a graph of the most influential accounts within this conversation, measured by a variety of factors such as: how many times an account mentions a key term, follower interaction with tweets, and the number of followers of that account.

What are the learnings which other progressive political organisations could deploy?

We found that there were a variety of learnings that could be applied to other progressive political campaigns wishing to use social to help mobilise support:

1. Campaign messages need to resonate: Our most successful messages were edgy, with this type of content outperforming on a CPM and CPE basis by some margin (+70%)

2. Online buzz does not necessarily result in votes: 477 people talking about the election had Labour in their bio, compared to six for Tory or Conservative. This means people were 53x more likely to identify as a Labour supporter than a Tory. This was not reflected in final results, where the Tories made gains

3. People respond to people: content needs to be authentic, with the candidates and issues concerned being key, not production values

CONCLUSION

What was so fascinating about this project was that it demonstrated statistically, there was a clear relationship between being seen online and the amount of votes cast for particular candidates. However, in order to be seen, brands must understand not only who their audience is, but also how and why they consume, share and act on online content, in order to maximise effectiveness. We have been able to prove a clear model for alternative political movements to drive political change.