The report by Ben Collier, James Stewart, Shane Horgan, Lydia Wilson and Daniel R. Thomas on Influence Policing: Strategic communications, digital nudges, and behaviour change marketing in Scottish and UK preventative policing states
Influence policing is an emerging phenomenon: the use of digital targeted ‘nudge’ communications campaigns by police forces and law enforcement agencies to directly achieve strategic policing outcomes. While scholarship, civil society, and journalism have focused on political influence and targeting (often by malicious actors), there has been next to no research on the use of these influence techniques and technologies by governments for preventative law enforcement. With grant funding from SIPR and support from The Scottish Centre for Crime and Justice Research (SCCJR), we have studied how this novel mode of police practice is developing through an in-depth study of Police Scotland’s strategic communications unit and a wider systematic overview of these campaigns across the UK.
Key findings: Police Scotland
1. Since 2018, Police Scotland has had a dedicated team devoted to strategic communications marketing campaigns - developing methods for crime prevention through communications. These adapt classic forms of ‘strategic communications’ and ‘social marketing’ to incorporate novel techniques and tools, tailoring them to crime prevention - particularly the use of behaviour change theory and digital targeting and segmentation infrastructures
2. These influence (including ‘nudge’) communications go beyond ‘information’ campaigns or those which simply tell or ask the public to do something, and instead incorporate psychological design elements which attempt to alter the decision environment in which members of the public make choices about their behaviour - often linking up with other interventions such as the redesign of public services.
3. In a wider policing context, these innovations can be understood as a development of problem-oriented and intelligence-led policing models in a digital context
4. The campaigns - focusing on areas with a perceived ‘online’ component, such as violence against women and girls, online grooming, and hate crime - are conceived as part of a public health prevention approach, often using perpetrator-focused adverts to deliver messages to those profiled as ‘at risk’ of offending. This is part of a move away from campaigns which simply rely on telling the public what to do, or which focus on putting the responsibility for crime prevention on victims.
5. There are two main elements to campaigns - the first are attempts to directly change people’s behaviour through ‘nudge’ communications, and the second are wider attempts to shape the cultural narratives that are perceived to contribute to crime.
6. In Scotland, digital targeting is mostly used at the broad demographic level (i.e. age and gender), although some use of fine-detail location and interest- based targeting is evident.
7. Online targeted paid advertising is used in conjunction with conventional media buying, and organic and ‘earned’ communication with stakeholders and civil society partners.
8. Campaigns are largely developed in house, but the media buying and some campaign development is done with commercial advertising and marketing partners. Civil Society stakeholders play a key role in consultation and development of campaigns, and in the ‘organic’ promotion.
9. The digital platforms themselves play a major role in shaping what is possible, sometimes redirecting the intervention through algorithmic processes or promoting organic sharing.
10. Evaluation of the campaigns is able to use some outcome measures but also still relies heavily on ‘vanity’ metrics (such as apparent views and click- throughs) provided by the platforms - and effects are difficult to measure.
11. The use of influencers (usually well-known public figures) in some campaigns to amplify messages is a clear innovation - though raises some concerns. These influencers have legitimacy with and knowledge of targeted communities, and generally retain their audiences across multiple platforms (even when these platforms change or fail).
12. We suggest the term Influence officers to describe the professional police communications specialists who design and develop these campaigns, who represent a growing new role in ‘frontline’ policing.
13. The centralised unit and single national force structure in Scotland has had some positive effects when compared to English forces, providing mechanisms for accountability (and saying ‘no’ to unsuitable or harmful campaigns) where more formal structures don’t yet exist. However, this is reliant on the tacit knowledge and expertise of a small group of practitioners - and some aspects of this approach would benefit from being on a more formal institutional footing.
14. Despite its proliferation across the UK, this is a distinctively Scottish mode of ‘influence policing’ and the ‘influence officer’ as a possible emerging role within policing.
Key findings: law enforcement and security in the wider United Kingdom
1. The use of targeted digital influence, or behaviour change, campaigns is widespread across UK law enforcement, but there is enormous variation in models, practices, theories, and institutional relationships between different forces and agencies; this often looks very different to how things are done in Scotland.
2. Within an extremely diverse range of practices, we identify six broad models underpinning these campaigns: based respectively on choice, risk, co- ordination, community, opportunity, and territory.
3. Campaigns are run by many different kinds of law enforcement bodies and their partners - including local divisions and forces, centralised agencies, government departments, security services, charity partners, local PCCs, and others.
4. There is a major focus on violence prevention (often through local VRUs), counter-radicalisation, and security themes. More recent campaigns include action around online harms and there has been a large increase in communications prevention activity around violence against women and girls in the last two years; though topics vary extremely widely in line with local police strategic priorities.
5. Targeting ranges from the very broad (all men in Manchester, or all people in England) to the hyper-specific, incorporating interests, behaviours, and very detailed location data (combining MOSAIC marketing, Census data and police intelligence with social media platform profiling techniques). Police are able to develop ‘patchwork profiles’ built up of multiple categories provided by ad platforms and detailed location-based categories using the platform targeting categories to reach extremely specific groups.
6. Paradoxically, as platforms have refined or in some cases removed the ability to target people by ethnicity and religion directly, this appears to have caused many law enforcement actors to adapt by adopting extremely invasive targeting via proxies for these protected characteristics, such as low-level postcode location data or microtargeting via interests and behaviours.
7. We have found evidence of some unethical and harmful practices in particular areas - most notably, the invasive targeting of extremely vulnerable refugees by the Home Office. More generally, we see the repeated use of language, interest and fine-detail location as proxies to access protected characteristics, such as ethnicity and sexual orientation. There is some particularly troubling and invasive targeting of Muslim and Black communities. The ethics of targeting are extremely complex, and law enforcement and public sector campaigns add a new dimension to these issues. Paradoxically, as platforms have refined or in some cases removed the ability to target people by ethnicity and religion directly, this appears to have caused many law enforcement actors to adapt by adopting extremely invasive targeting via proxies for these protected characteristics, such as low-level postcode location data or microtargeting via interests and behaviours.
8. The different forms this takes are clearly shaped by local strategic priorities, wider political environments, and the operational history of particular forces and agencies - particularly where there is a history of significant counter-terror activity and counter-radicalisation work.
9. Commercial marketing and advertising partners are promoting their expertise to forces, and work as contractors in development and media buying.
10. Some forces are being ‘left behind’ and doing essentially no strategic marketing - in these areas, much is being done by PCCs, charities, NGOs, delivery partners, charities, and other non-police actors, including national agencies
11. There is generally more emphasis on ‘hotspots’ and predictive policing approaches (often under risk or opportunity models) in England, with nudge communications offering a ‘smart policing’ complement to operational deployments.
12. There appears to be little central oversight at the bureaucratic or political level, and the legal and political position of these approaches is unclear - campaigns are often visible but the targeting is hidden. Some centralised agencies (e.g. the NCA) are much more developed in terms of accountability, legal case, proportionality, and ethics
13. Emerging platform Ad Libraries of political and social campaigns, particularly that of Meta, are making the use and reach of public sector advertising more transparent. We note that this has resulted in the platforms potentially revealing information intended by government agencies to be kept secret.
14. The strategic priorities of the Government Communications Service (GCS), including innovative use of new technology, sharing of audience data, professionalisation, and partnership working can draw directly on developments within policing in the UK, but also needs to support and reinforce the unique requirements of policing and the particularly sensitive nature of the social issues addressed in police communications
Outline and recommendations
The spread of this approach shows the clear (if still embryonic) emergence of a novel mode of operational policing (influence policing) - and a new kind of ‘frontline’ police sub-profession (the influence officer). Police communications officers are attempting to shape the culture and the behaviour of the public directly through behavioural marketing campaigns in order to achieve core strategic priorities around crime prevention. The digital tools provided by social media platforms allow them to find specific groups that are harder to reach - to project power and stake ownership of crime problems and online harms in digital spaces in which the police presence has struggled to establish itself.
We begin with an overview of the current landscape of digital communications, setting out the history and context of Police Scotland’s use of strategic communications approaches. We then discuss our study design and methods. This empirical section of this study is split into three sub-reports. The first of these is a study of Police Scotland’s strategic communications team, involving interviews and document-based research. The second is a review of four case studies of campaigns run by law enforcement in Scotland, based on analysis of key campaign documents. The third report extends the scope to all law enforcement digital campaigns run in the UK on Meta platforms, and is accompanied by a dataset of 12,000 campaign segments, how they were targeted, and audience data. We then discuss our key findings across the research project. In our final chapter, we give a set of key conclusions and recommendations. Based on our research, we make a number of recommendations within the following five key areas. Our detailed recommendations can be found in Chapter 7 of the report.
Recommendation 1: Transparency
There is a clear need for improved transparency around the use of digital targeting for strategic communications campaigns by law enforcement; we recommend the creation of a public register of digital campaigns and the targeting approaches used.
Recommendation 2: Formalising regulation and oversight
While some areas of good practice are emerging, there is little consistency between different forces and agencies, and no formal regulation within law enforcement other than that used for traditional communications. Much of the good and ethical practice is reliant on ‘load-bearing’ experts and their professional experience, and there are numerous examples of unethical practice in the wider UK. We recommend the formalisation of law-enforcement-specific best practices and accountability frameworks that recognise the unique environment of law enforcement, and the additional capacities and risks of these approaches compared to traditional communications.
Recommendation 3: Improving monitoring and evaluation
Although some campaigns appear to have promising pro-social results, evaluation is extremely difficult and based largely on ‘vanity’ metrics provided by the platforms. Significant additional investment is required for monitoring and evaluation, especially to assess the potential for unintended consequences and ‘blowback’.
Recommendation 4: Co-production and consent
The emerging Police Scotland model contains many distinctively ‘Scottish’ aspects, particularly an emphasis on partner engagement and some examples of fully co-produced campaigns. We recommend the further development of democratic and deliberative involvement of the public - especially targeted communities - in the development of campaigns.
Recommendation 5: Public awareness and engagement with political actors/civil society
There is still little public knowledge (including among politicians, policymakers, and law enforcement themselves) of the use of digital targeted influence campaigns for preventative law enforcement. This needs to be part of a wider public discussion of these approaches. Further, these approaches are becoming a core part of law enforcement reactions to ‘online’ or ‘emerging’ harms - and need to be systematically considered and explicitly set out in strategy and law relating to online harms, including the potential harms and risks which they themselves may pose.
Further Research
This report covers an initial overview of the use of strategic communications in the UK, with a particular focus on Scotland. However, several key questions remain unanswered or only partly-understood; substantial additional research is needed into this important emerging phenomenon. Further research is needed into how law enforcement practitioners across the UK (and across a range of policy areas) are adapting these communications techniques for preventative law enforcement. The role of private sector consultancies across this landscape is a crucial aspect of this developing set of practices; we also have little understanding of how the platforms and social media sites - Facebook, Twitter, Google, and others - are adapting to these new use cases. Complementing this research, there is a need for a better appraisal of how these campaigns are being experienced by the public (especially communities that are heavily targeted).
