Thread by @RobertCPalmer13

Want to know the digital strategy adopted by Vote Leave during the referendum?

This thread summaries it based on expert evidence given to the courts.

This is how they set out to steal our democratic process:

1.

Platforms like Twitter and FB give campaigners outlets to communicate w/out worrying about how journalists/editors may change, interpret, or fact-check campaign messages. Removing editors/journalists from the flow of info removes checks on opinions & facts in circulation.

2.

By using social media, the claims of the Leave campaign went unchecked & they circumvented investigations by journalists into whether inpolitical claims were accurate. Instead, the opposition was left to counter erroneous & inaccurate claims. They could say what they wanted.

3.

Ordinarily, campaigns win support by using as much personal data on ppl as possible to tailor & customize political messaging. They also use this data to 'redline' (exclude the groups that would be a waste of resources).

The Leave campaign found new ways to manipulate data.

4.

It's commonly known that most voters in elections and referenda make up their mind in the last few days of a campaign. In all elections between 20% & 30% of voters decide within a week of the vote (half of them on the last day) - VL et al exploited that phenomenon to the max.

5.

Numbers are higher in a ref as ppl tend to think their vote
is more significant. A key finding in an LSE report showed that 54% of ppl saw the ref as the most important vote in a generation (81% put it in the top 3).

Exploiting such factors was pivotal to Leave's victory.

6.

Broadly, Vote Leave’s (VL) digital strategy was conducted largely by the Canadian firm, Aggregate IQ (‘AIQ’). AIQ began by building a ‘core audience’ for VL’s adverts, by 1st identifying the social media profiles of those who
had already ‘liked’ Eurosceptic pages on FB.

7.

VL advertised to this 'core audience' to try & attract them onto its website, where they would be invited to add their details to its database. AIQ also used an advertising tool within FB called the ‘Lookalike Audience Builder’.

8.

The Lookalike Audience Builder was applied to the demographic features identified by FB in the ‘core audience’ group to the UK population.

This 2nd group, known as the ‘persuadables’, comprised of ppl FB had identified as having the same demo features as the core audience.

9.

The ‘persuadables’ hadn't previously expressed interest in Eurosceptic content on FB by ‘liking’ Eurosceptic pages. This group of ‘persuadables’ contained ‘a number of better-educated and better-off people’.

10.

Dom Cummings stated in a blog post that the persuadables were ‘a group of about 9m ppl defined as: "between 35-55, outside London and Scotland, excluding UKIP supporters and associated characteristics, and some other criteria"’.

11.

VL then began a process known as ‘onboarding’. This turned sympathisers into committed supporters of, donors to, and volunteers to the campaign. To achieve this, VL & AIQ used advertising targeting the ‘persuadables’ & employed a 3 step ‘onboarding’ process.

12.

Step 1 was to invite the reader to click on an online advert, which would be displayed on FB or another digital channel, to be taken to VL’s website. Once there, the 2nd step was to invite the reader to provide their personal
details, which would also populate VL’s database.

13.

The 3rd step was to invite the reader to either make a donation, share VL’s messaging around (with individuals using their own s/m accounts & generating organic growth of the message w/out any cost to VL), or volunteer their time to the campaign.

14.

At each step, the ads & messages were tested on an interactive basis,
so that ads or messages that failed to convince enough readers to move to the next step were re-worked or changed entirely until a success threshold was reached ( this is aka the ‘conversion rate’).

15.

Different ads were developed for the ‘core’ & ‘persuadable’ audiences. The
former were served more ‘demonstrative’ ads, the latter reacted better to ads that appealed to their curiosity in weighing the arguments e.g. ‘is this a good idea?’ or ‘do you want to know more?’

16.

4 weeks before polling day, VL launched a competition to gather voters for its database. The competition boasted a £50m prize to anyone who could predict the winner of all 51 games at Euro 2016 football cup to attract ppl who don't
usually have any interest in politics.

17.

The images used were released by FB. Participants were asked to provide their name, mobile number, address and email details during the entry process. This data was fed in to VL's database. Over 120K entered, all were sent a reminder
on 23 June 2016 to vote in the ref.

18.

VL launched an app for smartphones. This app was ‘gamified’: users were awarded virtual points for performing actions such as completing quizzes on VL’s talking points & from watching VL’s video material etc. Users sent 70K messages on polling day to remind friends to vote.

19.

VL identified that swing voters were very confused, & liable to change their decision on which way to vote based on the last message seen from either side of the ref campaign. Cummings implemented the ‘Waterloo Strategy’ to ensure that a VL ads were seen late by swing voters

20.

Cummings stated: ‘We ran loads & loads of experiments for months, but on relatively trivial amounts of money. And then we basically splurged all the money in the last 4 weeks & particularly the last 10 days.’

Remain reached their spending cap by the last few days.

21.

In his blog, Cummings said that ‘one of the few reliable things we know about advertising amid the all-pervasive charlatanry is that, unsurprisingly, adverts are more effective the closer to the decision moment they hit the brain’.

22.

VL pressed their NHS, immigration, & impending membership of Turkey lies throughout the campaign & especially in the last few days of the campaign. There were dozens of ads with such messages.

People were conned into believing made up claims up until & including polling day.

23