Sex Workers As Collateral Damage, Once Again: A Critique Of The New ‘Sex Trafficking Identification Matrix’ Tool

This article is a guest post by Victoria Holt, Emily Kenway and Addy Berry, with thanks to Lydia Caradonna

Context and background

Evidence-based research is necessary to help improve the lives and working conditions  of sex workers and this is especially the case where research lies at the intersection of anti-trafficking and the sex sector. Some discourses and discussions end up positioning people who are trafficked as mere victims of crime, forgetting the whole social, economic and political context that puts people in that situation, often in a voluntary and conscious way. 

Responses to both sex work and trafficking, such as the Swedish/Nordic model are ideology based and have proven harmful to everyone in the sex industry, consenting or otherwise. Anti-trafficking measures are routinely used against sex workers under the guise of protection, resulting in intrusive or destabilising ‘rescue missions’ such as Operation Lanhydrock (2016), or Operation Pentameter (2009). The article critiques a ‘bespoke tool of analysis’, created to flag ‘risk indicators’ on sex worker advertising profiles for use in police investigation. The tool of analysis is the Sex Trafficking Identification Matrix (STIM) by Xavier L’Hoiry, Alessandro Moretti & Georgios Antonopoulos

The aims of this article are threefold:

  1. To highlight the risks of adding to a body of anti-trafficking research which is ideologically driven and methodologically flawed, whether or not the researchers are aware of these issues.
  2. To highlight the importance of sex worker involvement in sex work related research, and to articulate the shortcomings of research which knowingly sidesteps sex worker involvement at any stage. We do this in part by examining distorted inferences of data taken from Adult Service Websites (ASWs).
  3. To explore the use of the STIM and to lay out the danger posed by creating a tool to strengthen police intervention.

We take each of these points in turn. 

  1. Ideology and its flawed understandings of the sex industry

In brief, the Sex Trafficking Identification Matrix (STIM) is a tool created with the aim of aiding law enforcement to distinguish between independent sex worker adverts, and those created by traffickers. STIM refers to a list of ‘risk indicators’ allowing police to sift through profiles and categorise them as ‘low priority’, ‘considered for investigation’ or ‘high risk’.  STIM was created to tackle an extremely specific view of sex trafficking, backed by ‘experts’ in the field. We show how their vision is not supported by evidence. 

STIM is made to target the ‘industrial scale’ commercial sexual exploitation it has been claimed is happening via Adult Service Websites (ASW). The authors’ claims of this ‘industrial scale’ have been taken from the All-Party Parliamentary Group on Commercial Sexual Exploitation (APPG CSE) which has recommended bringing in FOSTA/SESTA (1) style laws to shut down ASW as well as promoting the criminalisation of the purchase of sex. The APPG CSE’s research and reports are widely recognised to be based on flawed evidence and are ideologically driven, evidenced by the report considering all forms of transactional sex to be commercial sexual exploitation, a far reaching accusation and against all evidence such as that provided in the recent report commissioned by the Home Office. The APPG CSE’s views are also not matched by those of prominent anti-trafficking experts (for example here and here).

The APPG report cited in L’Hoiry et al contains numerous fundamental flaws. First, it fails to differentiate between migrant sex workers and victims of trafficking: the existence of foreign sex workers is adopted as proof of sex trafficking, whereas it is well known that economic migration is a source of foreign-born sex workers. Further, it presents the reported increase in the number of brothels (324 according to the APPG) as indicative of ‘industrial scale’ sex trafficking, but offers no evidence of this. Over the same period, the number of construction sites in the UK has also increased, but few would accept this as evidence of human trafficking.

Further, the STIM risk indicators are informed by the research of Diba et al, but Diba acknowledges that most women know that they will be working in the sex industry and willingly wish to travel; most of the time there is no actual coercion during recruitment and so police intervention will disrupt workers not victims..

Indeed, trafficking is a field notorious for hyper-inflated figures and dubious research methods, justified by claims that the true figures can never be known. One notorious example of this is research by Kelly and Regan (2000) found 71 cases of trafficked women,but this was extrapolated without justification up to 4000 cases, a figure reported and repeated; a figure then used to justify invasive/violent measures and over-policing.

Further indication of the problematic basis for the L’Hoiry research is to be found in the article referenced as ‘gathered knowledge’ The article, Latonero (2011), examines sporting events and sexual exploitation, and tracks an increase of ASW listings during the 2011 Superbowl 2011, taking this as evidence of forced sexual exploitation at the time of the event. However, Latonero does not distinguish between forced sexual exploitation and consensual commercial sex meeting an increase in client demand. 

In fact, there is ample research disproving the link between sexual trafficking increasing around major sports events, including Finkel and Finkel, Dolinsek in Al Jazeera, and the Anti-Trafficking Review.  It is therefore surprising that, while the link between sexual trafficking and sports events has been proven to be an ideological myth, it is  considered “gathered knowledge” by L’Hoiry et al. 


2. Shortcomings of no sex worker involvement

No sex workers were consulted in L’Hoiry et al’s research. The authors claim that this is a result of lack of available time, although they did have time to interview 26 experts. Research which aims to identify people at risk within an industry, but which does not gather information from anyone within that industry, may at first sight be exposed to shortcomings. And concerns may further be raised in the light of other research showing that policing often harms sex workers, so a study using police input without the input of sex workers may be subject to charges of bias and ethic. 

All of the research they reviewed to inform the creation of STIM was research that excluded knowledge from sex workers. The absence of sex workers in the research becomes most apparent when they discuss the ‘established’ indicators of trafficking. When sex workers are not involved in the research at any stage, there is no opportunity to intervene in their analysis of ASW and say “Hey, that doesn’t mean what you think it means. Let me explain”. 

The vast majority of the risk indicators of the STIM, as well as the underlying assumptions around victimhood or sexual exploitation used to create them, would also capture independent sex workers if applied. This implies that the tool will not achieve its anti-trafficking aims. The lack of sex worker involvement, and dangerous assumptions, has resulted in a set of potentially useless ‘risk indicators’. To demonstrate, we go through the list of risk indicators below, and interrogate the misinformation of each.

Repetitive and distinct patterns of behaviour: Antonopoulos and colleagues claim that traffickers display ‘repetitive and distinct’ patterns of behaviour and the STIM’s risk factor is based on this. However, we know that sex workers display ‘repetitive’ behaviours that could easily be picked up by STIM, such as use of specific words that they know work well to attract clients; similar spelling mistakes because one worker can speak English well and does the text for other all the profiles; sex workers copying and pasting from other workers for ease, etc. 

Poor English language on profiles was listed as a risk indicator, but this is often just a feature associated with migrant workers, not trafficked people. Also for some reason STIM assumes that most traffickers are also migrants, when data does not currently exist to demonstrate that.

‘Inconsistency of story across different profiles’ is based on research by Ibanez and Suthers and considered to be indicative of trafficking. Different ages, ethnicities or names for the same worker across different profiles are somehow suspect. In reality, this is extremely common because it’s advertising. Sex workers often have separate personas to try and capture more markets. 

Lying about ethnicity or nationality is also common, and while it may seem dubious, it is usually in response to customers shaping the demand, and the racist policies of ASW, such as not allowing people from some South Asian or Eastern European countries to register.  One of our authors has worked in outreach with sex workers in their local community and assessed a ‘rapid respond needs assessment’ of sex workers by the local council.  Even though the assessment was fully anonymised, often the Eastern European sex workers lied and said they were Italian or Greek because – they said – if they wrote their Eastern European or South Asian nationality, they feared that the police would assume they were part of criminal networks or trafficked and thus arrest them or issue closure orders of their working flats. Lying about nationality may not be an indication of being trafficked, but rather a way of avoiding intrusion and possible further harm. 

‘Frequent movement/advertising in more places’. Every sex worker who is willing to travel to their clients will advertise in more than one place – in fact, some people don’t work locally because they don’t want to feel too visible sex working in their own community. Even people who are incall only (meaning that they only work from their own premise, or a hotel which they book) will advertise in more than one location to attract more profile views, and to gauge interest in a city before organising a working tour there. It is not indicative of anything except trying to attract more clients.

Somehow, both travelling and not travelling are risk indicators. ‘Restricted movement’ is also listed as an indication of trafficking, but in the context of sex woker advertisements this simply means sex workers not offering outcalls to clients (outcall is when a sex worker goes to the client’s place of residence or a hotel they have booked). This example in the STIM makes absolutely clear the abject failure of the research to speak with sex workers. Outcalls are more dangerous for sex workers as they will be alone in the home of a stranger, and so many workers do not provide them and so will not travel. How can both travelling and not travelling be risk factors? Who, then, are the only ‘real’ sex workers?

‘Profile posted by third party’/language suggesting it isn’t posted by workers themselves. Some workers have fake assistants or pretend that they’re not directly responding to things themselves to weed out time wasters. This is also true of the point we made earlier about sex workers asking others to write their profiles, or copying from other profiles. 

Phrases indicative of ‘youth’ e.g new in town, fresh. Firstly, service providers often advertise as new or create new profiles to access the new client rush – many clients enjoy seeing ‘new’ workers. But also, advertising ‘youth’ is a response to cultural fetishisation of youth. Iin some gay communities talking of “boys” means adult twinks (18-21) and doesn’t mean literal boys. Youth is not necessarily indicative of trafficking. 

Weights under 115lbs. The authors seem to think that low body weight is an indication of either starvation or of a literal child, but it is one example of many that we found in L’Hoiry et al’s paper of taking ad copy at face value without sex worker input. Sex workers lie about their weight because of social fatphobia: often men want to sleep with slim, young women. Also, skinny women exist, and we refer to our point above about fetishisation of youth.

Links to external websites, which sex workers do all the time to bolster their traffic and remove third party interference. 

References to ‘spa massage therapy’ is listed as a risk indicator with no justification. Many workers will either actually offer it or say they do it because of criminalisation. Spas and massage parlours are convenient and socially accepted fronts for the sex industry and migrant workers often work together there for safety.  

Subjects from countries associated with high levels of trafficking –  we contend that ultimately, the STIM is based on racist and racially coded algorithms, just like in the Foxglove case.  The English Collective of Prostitutes has previously highlighted that their Romanian members are erroneously and harmfully targeted by this kind of assumption – an assumption upheld by the APPG report cited earlier – and this further proves our point above about why sex workers lie about their nationality or country of origin. 

Low cost – Poverty can push people into sex work, and low cost of services is an indicator of desperation. Desperation is not linked to trafficking but more likely to be due to being caught up in the Universal Credit waiting time issue or client scarcity because of the COVID pandemic.  

PAYG phones suggest ‘involvement with criminal networks’ sex work is criminalised, and sex workers use burner phones to avoid stalking and harassment.

Analysis of photos  – overly explicit photos are listed in the STIM as being indicative of trafficking, along with uneasy facial expressions of women. Without wanting to sound crude, these are photos of actual, literal prostitutes, so the contention of photos being ‘overly-explicit’ is oddly paternalistic.  Further, uneasiness is incredibly subjective and not conclusively indicative of trafficking. It is possible to feel uneasy in front of a camera without being forced in front of it. Low quality/resolution was listed as well, but having a cheap work phone is not the same as being trafficked, it’s just taking photos on a cheap phone to save money. Multiple workers having similar backgrounds is not indicative of trafficking either, because workers will visit the same hotels, in-calls or shoot locations. They may all be working from the same location. 

Subject appears to be between 18- 24 years old – Sex workers, along with many other people in appearance focused work, lie about their age all the time. It’s so common that it’s simply ludicrous to suggest that listing one’s age as being between 18-24 is suggestive of trafficking. Even if the workers are not lying, it is not de facto suspicious to be working in the sex industry between 18-24. Many sex workers enter the industry in this age bracket. 

Wide range of services advertised such as anal or bareback – again an indicator of poverty, not trafficking. Also people will just tick all of the boxes/services so that they show up in searches, and then negotiate these with clients outside of the website. 

Non-itemised’ or ‘all inclusive pricing’ is seen as a risk factor, when actually people negotiate their boundaries and costs with each client even if only one price is listed on the profile. When offering lots of services as a strategy to earn more is considered by police to be a sign of trafficking, it means that the workers most likely to be caught up in the policing outcomes of STIM are the ones who are at their most vulnerable.

3. Potential Harms of a trafficking tool

L’Hoiry et al argue that a ‘key challenge’ for law enforcement is being able to differentiate between independent workers, and trafficked persons. No definition of trafficking is given in the article, and while the authors state that they do not want the STIM to replace the ‘craft’ of policing, it is clear that the STIM is intended to act as a convenient tool for the police to increase the power of their ‘craft’ – a craft, which we will show to have been effective neither in preventing trafficking nor in responding to it. 

Failure of police intervention 

Policing is one of the main drivers of violence and displacement of sex workers. STIM is a tool for further police intervention and involvement, which, we believe, is harmful to sex workers, especially those most vulnerable to police violence. In this last section we assess the failures of police intervention in suspected trafficking cases, and show how the STIM as a tool will fail in its aims of assisting them. STIM should not be employed as a tool for law enforcement, nor should the paper be cited in further research, for risk of propagating false notions of trafficking, victimhood and rescue. 

‘Welfare’ checks  and brothel raids by the police posit sex workers as being physically trapped in their circumstances, or imprisoned in brothels, in need of being freed through closure orders and arrests. Assistant Chief Constable Dan Vajzovic – who also acts as the National Police Chiefs’ Council lead on sex work and prostitution – has acknowledged that “these welfare visits [can] appear to some as a raid”.  We argue that police intervention through checks or raids are are frightening and violent intrusions. We have heard reports of sex workers having their earnings and passports confiscated or being forcibly repatriated. Is this outcome really in the best interests of victims?

Research conducted by Laura Connelly and the English Collective of Prostitutes found that growing numbers of migrant workers changed their working practices due to a perceived risk of deportation and arrest. The same report showed that nearly half (44%) of respondents said their existing relationship with the police was ‘poor’ or ‘very poor’. 

Pop up brothels, identified by the APPG, is hardly an indication of trafficking, but more likely to be in response to destabilizing and non-consensual police interventions. Sex workers wish to hide from the police so they move around before they can be located. Worryingly, police often conflate workers with managers and this results in increased arrest of sex workers for managing a brothel or pimping when they are really only working with a friend or colleague, and sharing a premise for safety. 

Increased policing has been shown time and again to result in sex workers being less safe, not more safe. We strongly advocate against any tools being utilised which increase police power to intrude on the lives of sex workers in the name of safety. 

Failures of the STIM

We have shown that the STIM is the result of research based on incorrect assumptions about the industry, and uninformed by contact with workers in the industry. If used by law enforcement, police will be attempting to locate ‘victims’ in the wrong place. 

We already know that the vast, vast majority of ‘suspected’ trafficking cases lead to no victim being identified. For example, out of 501 Modern Slavery Helpline calls referred to the Gangmasters and Labour Abuse Authority in 2018, there were 51 investigations and only 1 victim referred.  The authors acknowledge the probability of false positives, and stress the need for the STIM to be used alongside existing knowledge and technology. They do not, however, acknowledge whether any legitimate trafficking victims have been discovered with the help of STIM. The tool has no validity, even if it were based on sound theoretical arguments, unless there is a way of testing it in the field and gathering evidence of its effectiveness, accuracy, and benefits. The authors present no methodology for any of these outcomes. Does the tool actually work, or is it just convenient? L’Hoiry et al also acknowledge that many risk markers are just risk markers of independent migrant workers which vitiates the tool, so why did they not attempt to improve the tool? Are they accepting them as collateral damage?

It must be acknowledged that the STIM does nothing to address the state’s own role in the construction and maintenance of the sexual labour market, such as low wages, precarious employment, or borders nor does it assess risks or vulnerabilities to trafficking or being trafficked. It is unfortunate that, while L’Hoiry et al acknowledge that there is a lack of empirical research into the accuracy of tools already used by law enforcement they risk exacerbating this problem by making yet another inaccurate digital investigation tool. 


Nowhere, at any point, did the authors of the STIM consult those who would be affected most by the tool, and who would arguably contribute the most knowledge:  sex workers themselves. Perhaps this was simply an oversight or, perhaps, because sex workers are often reluctant to become involved in academic research, for reasons that have been explored elsewhere. The present authors, who have collective lived experience of sex work, as well as academic interests in the sex industry and trafficking, did approach L’Hoiry et al following their pre-publication announcement. We question, therefore, whether neglecting to include sex workers in their research is  is at best, a failure to respect the knowledge and agency of workers in the sex industry and, at worst, a deliberate attempt to reproduce knowledge which delegitimises sex work as a form of labour and further entrenches harm and violence against sex workers.

As discussed above, enhanced police powers makes vulnerable people –in this case migrant sex workers—increasingly fear the police. The tragic irony is that this plays directly into the hands of exploiters. If sex workers fear that the police are mining ASW for data, they will simply stop working on those sites, and turn to the streets or brothels where they will be less detectable, and also at greater risk. Working independently online allows sex workers to exercise agency over who they see and how. Instead of locating the problem, and thus any kind of solution, on ASW, we need to fully decriminalise the sex industry to empower women with rights. Decriminalisation creates the opportunity to strengthen anti-trafficking activities, as explained in this briefing

Like L’Hoiry et al, we wish that no person would be subjected to trafficking and associated harms.  However, for the reasons given above, it is clear that the STIM as a tool will fail in its stated aims and will quite likely lead to increased harm if it is deployed. We must therefore conclude that the STIM should not be employed as a tool for law enforcement. Further, researchers and practitioners must listen to ‘experts by experience’, that is, those with lived experience of the industry itself. 


(1) This would mean websites would be responsible for the content that their users post, holding them accountable for potentially aiding human trafficking. To cover their backs many websites such as Backpage, removed all adult service advertising resulting in “dire financial situations” for sex workers.

About the authors

Adeline Berry is an MSCA fellow and PhD student researching the experiences of older intersex people as part of INIA. A former member of SWAI, they are a member of Red Umbrella Front and the author of How Sex Workers Understand Their Experiences of Working in the Republic of Ireland.

Emily Kenway is a PhD scholar in political science at Edinburgh University focusing on exploitation. Prior to this, she was a policy adviser on human trafficking. She is on the boards of the thinktank Commonwealth and the Public Interest Research Centre. Her first book, The Truth about Modern Slavery, was published in January 2021 and described as a ‘powerful treatise’ by the Guardian. She is currently working on her second book.

Victoria Holt is a PhD student at the University of Roehampton, researching sex workers’ experiences of domestic violence and abuse. Her work has been published in The Independent, The Quietus and Novara Media. She is a member of SWARM and Decrim Now and campaigns for the full decriminalisation of sex work.

Leave a Reply

Your email address will not be published. Required fields are marked *