"This app is so laggy, add me on the other one."
"I can't really talk here, my mum checks this."
"It's more private over there, just us."
None of these messages would trip a keyword filter. They are polite, ordinary and, most of the time, completely harmless. Children move conversations between apps constantly, usually with friends they already know. Yet when the person asking is someone the child has only just met online, this small, friendly request is one of the most consistent features of how online grooming unfolds. It is also one of the last points at which the platform where the contact began can still see what is happening.
That is why I think the request to switch apps deserves far more attention than it gets.
Key takeaways
- Off-platforming, moving a child from a public or moderated space to a private one, appears consistently across research, charity and regulator evidence on grooming.
- 19% of UK children aged 11 to 17 have been asked to move their chat by someone they barely know (Ofcom, 2026).
- The request is framed as convenience and privacy, not as anything sinister.
- It is often the last moment the original platform can intervene, and the first moment the pattern becomes legible.
- Detection must read the sequence, not the sentence.
A step that keeps appearing in the evidence
Researchers, charities and regulators describe online grooming in different ways, but the move to a more private space shows up again and again.
How WeProtect describes off-platforming
WeProtect Global Alliance calls it "off-platforming": perpetrators who find children on social media, in chat rooms or in games divert conversations to a private messaging app or an end-to-end encrypted environment "due to the lower risk of detection" (WeProtect Global Alliance, 2023). The same report describes how, in gaming environments, once a child has answered sexualised questions, the perpetrator "quickly looks to move them to a private messaging platform that allows the sharing of images, and voice and video calls", with "a strong preference for apps that are encrypted or where chats are perceived as unmoderated".
The same pattern in UK police data
The NSPCC describes the same pattern in UK police data. Perpetrators typically use mainstream platforms, games, chatrooms and dating sites for first contact, and then "encourage children to continue communication on private and encrypted messaging platforms where abuse can proceed undetected" (NSPCC, 2024). Its more recent analysis of police-recorded sexual communication offences identified "cycles of behaviours" in which perpetrators create multiple profiles and manipulate children into engaging with them across different platforms (STV News, 2025).
Academic and industry research
A scoping review of 93 peer-reviewed studies of grooming strategies found that research on online grooming identified strategies not seen in pre-internet studies, including "media progression" alongside isolation, secrecy and deception (Ringenberg et al., 2022). Industry guidance from the Technology Coalition describes perpetrators switching platforms as the relationship develops, starting in a more open setting and moving to direct messaging, potentially encrypted, as they "double down" on their connection with the child (Technology Coalition).
How common is the request to switch apps?
More common than most adults assume.
In Thorn's 2022 survey of young people in the United States, roughly two in three minors said an online-only contact had asked them to move from a public chat into a private conversation on a different platform (Thorn).
The most detailed recent UK evidence comes from Ofcom's 2026 survey of 4,097 children aged 11 to 17, carried out by Savanta (Ofcom, 2026). It found that:
- 19% of children had been asked to move their chat to a different app or platform by someone they did not know well or did not know at all, rising to 25% of 16 to 17 year olds.
- Around three in ten of these requests came either at the very first contact (15%) or within the first day (a further 16%).
- 21% of children who received such a request continued chatting on the new platform.
- The most common reasons given were that the other person said it would be "easier / better" on the other platform (43%), or that the other platform was "more private" (36%).
The two figures are not directly comparable, because the surveys asked different questions of different populations in different countries. Ofcom's question also asked specifically about people the child did not know well, whereas Thorn's asked about online-only contacts. Together, though, they tell us this is not a rare or exotic experience. For many children, it is a routine part of being online.
The last finding matters most for anyone designing safety systems. The request is not framed as anything sinister. It is framed as convenience and privacy, which are exactly the things children value in their apps.
Why moving apps removes protective safeguards
Criminology has a simple way of explaining why. Routine activity theory holds that harm becomes more likely when a motivated offender and a suitable target come together in the absence of a capable guardian (Cohen and Felson, 1979). Online, a "guardian" is not only a parent looking over a shoulder. It is the platform's moderation, its reporting tools, its safety defaults, the friends who can see a public thread, and the parental controls that may be configured on one app and not another.
A request to switch apps is, in effect, a request to leave those guardians behind. The person asking may not have the child's best interests in mind, but they understand the environment well. One platform is useful for finding children. Another is useful for being alone with them. Each service can be working exactly as designed while, together, they form a pathway neither intended.
Three risks once the conversation moves
This creates three problems.
1. Visibility collapses after the move
Once a conversation is on a different service, and particularly an end-to-end encrypted one, the first platform has no further view of it. The NSPCC has warned that gaps in the Online Safety Act mean private messaging services have fewer protections for children than public spaces, raising fears that harm will migrate from public to private settings (NSPCC Learning, 2025).
2. The escalation that follows happens out of sight
WeProtect's account of off-platforming is explicit that the destination is chosen for its image sharing, voice and video features and its lower risk of moderation (WeProtect Global Alliance, 2023). The requests for images, the pressure and, in some cases, the sextortion that follows, frequently take place somewhere the original platform cannot reach.
3. The request is often the clearest signal the first platform will ever get
Before the move, the conversation may consist of compliments, shared interests and questions that would look innocent in isolation. The request to move is often the moment where the intent of the contact becomes clearer, especially when it arrives alongside questions about age, whether parents check the phone, or an emphasis on keeping things "just between us".
In other words, the switch is a pinch point. It is the last place where a platform can still intervene, and often the first place where the pattern starts to become legible.
Not every request is grooming
This needs saying clearly, because getting it wrong carries real costs. Children move chats between apps for entirely legitimate reasons all the time. In Ofcom's survey, WhatsApp was the most widely used app among 11 to 17 year olds (74%), and two thirds of children used one-to-one direct messages (Ofcom, 2026). A classmate saying "add me on Snap" is not a safeguarding concern.
A system that treats every mention of another app as a red flag will drown moderators in false alarms, frustrate young users and teach them to work around the tool. The request only becomes meaningful in context. The questions that matter are about the relationship and the sequence, not the words:
- Who is asking? Is this a long-standing connection or someone who appeared in the child's messages hours ago?
- When is it happening? A request at first contact or within the first day looks very different from one made months into a friendship between peers.
- What came before it? Has the contact been probing age, supervision, loneliness or secrecy? Has there been intense flattery, gifts or attempts to create a special bond?
- What is being offered? Is the new space being sold as "more private", somewhere parents won't see, or somewhere the child can "be themselves"?
This is consistent with the wider evidence that grooming does not follow a single fixed script. Lorenzo-Dus and colleagues, analysing chat logs shared by UK law enforcement, argue that recognising online grooming "is not simply a matter of spotting tell-tale words or phrases" but of understanding how language functions across a dynamic exchange (Evans and Lorenzo-Dus, 2025). The request to move is one move in that exchange, and it carries most weight when read alongside the others.
What platforms can do about off-platforming
The regulatory direction in the UK is already moving towards reducing the opportunities that make off-platforming possible. Ofcom's Illegal Content Codes recommend that services at high risk of grooming set defaults so that children are not suggested to strangers, their connection lists are hidden, and they can only receive direct messages from accounts they are connected to (Ofcom, Illegal Content Codes of Practice). In 2026, Snap, Meta and Roblox committed to further anti-grooming measures in response to Ofcom's demands, including Snap preventing adult strangers from contacting children by default in the UK and Meta planning AI tools to detect likely sexualised conversations between adults and teens in Instagram direct messages (Ofcom, 2026).
These measures matter because they reduce how often an unknown adult gets the chance to make the request at all. But defaults will not catch everything, and they do not help once a contact has been accepted. For that, platforms need to be able to recognise the request when it happens and respond proportionately. In practice, that means:
- Treating the move request as a risk signal, weighted by context. A request from a new, unknown contact, early in the relationship, alongside secrecy or age probing, should raise risk. The same words from an established peer should not.
- Intervening before the child leaves. A gentle, well-designed prompt at the moment of the request ("You don't know this person well. Moving to another app means we can't help if something goes wrong") gives the child information while they are still in a space where support exists. Ofcom's own codes already recommend reminders and safety options when a child exchanges a first direct message with a new account (Ofcom, Illegal Content Codes of Practice).
- Routing high-risk patterns to human review. Where the request is part of a wider pattern, trained moderators should be able to see the sequence, not just a single flagged message.
- Sharing signals where it is lawful and proportionate. Because perpetrators move between services to evade detection, research commissioned by the NSPCC recommends cross-platform signal sharing and on-device protections that can work in private and encrypted spaces (NSPCC Learning, 2025).
Reading the sequence, not the sentence
At Tuteliq, we build detection around behavioural sequences rather than word lists, because the request to switch apps is exactly the kind of signal a per-message filter misses. "Add me on the other app" is unremarkable on its own. "Add me on the other app", from an account that appeared yesterday, after questions about age and whether parents check the phone, is a different situation altogether. The words are the same. The pattern is not.
The opportunity here is a narrow one. Once the conversation has moved, the platform where the child was found has, in most cases, lost the ability to help. The moment of the request is when protection is still possible, and that is the moment safety systems should be designed to see.
If you run a platform that children use and want to understand how conversation-level detection could help you recognise off-platforming attempts in context, we would be glad to talk.
Dr Nicola Harding is Chief Scientific Officer at Tuteliq.
Frequently asked questions
What is off-platforming in online grooming?
Off-platforming is when someone who has made contact with a child on one service, such as a game, social app or chat room, persuades them to continue the conversation on a different, usually more private or encrypted, app. WeProtect Global Alliance describes perpetrators doing this because of the lower risk of detection.
How common are requests to switch apps?
Ofcom's 2026 survey of 4,097 UK children aged 11 to 17 found that 19% had been asked to move their chat to a different app by someone they did not know well or at all, rising to 25% of 16 to 17 year olds.
Is every request to move apps a sign of grooming?
No. Children move chats between apps with friends all the time. The request becomes meaningful in context: who is asking, how early it happens, what came before it and how the new space is being sold.
What can platforms do about off-platforming?
Treat the move request as a risk signal weighted by context, intervene with a well-designed prompt before the child leaves, route high-risk patterns to human review, and share signals across services where it is lawful and proportionate.
---
References
- Cohen, L.E. and Felson, M. (1979) 'Social change and crime rate trends: a routine activity approach', American Sociological Review, 44(4), pp. 588-608. Read the journal article
- Evans, C. and Lorenzo-Dus, N. (2025) 'A corpus-assisted discourse analysis of children's and groomers' talk in online grooming interactions', Applied Corpus Linguistics, 5(3), 100147. Read the research
- NSPCC (2024) Online grooming crimes against children increase by 89% in six years. Read the NSPCC analysis
- NSPCC and PUBLIC (2025) Tools to combat online harms: protecting children in private messaging spaces. London: NSPCC. Read the private messaging report
- Ofcom (2026) Understanding Online Communications Among Children (research report by Savanta). Read the report
- Ofcom (2026) Illegal content Codes of Practice for user-to-user services. Read the codes
- Ofcom (2026) Tech firms commit to stronger anti-grooming measures in response to Ofcom demands. Read the Ofcom update
- Ringenberg, T.R., Seigfried-Spellar, K.C., Rayz, J.M. and Rogers, M.K. (2022) 'A scoping review of child grooming strategies: pre- and post-internet', Child Abuse & Neglect, 123, 105392. Read the scoping review
- STV News (2025) Thousands of online grooming crimes recorded in Scotland over past five years (reporting NSPCC analysis of police data). Read the STV report
- Technology Coalition (n.d.) Considerations for Detection, Response and Prevention of Online Grooming. Read the paper
- Thorn (2022) Online Grooming: Examining risky encounters amid everyday digital socialization. Read the Thorn research
- WeProtect Global Alliance (2023) Global Threat Assessment 2023. Read the assessment
Note on evidence
The Ofcom and Thorn figures are self-reported by children and young people and come from different countries, age ranges and question wordings, so they should not be compared directly. Ofcom's timing and response figures are based on a smaller group of 209 children who reported receiving a move request in the past year; the survey excludes gaming platforms. Ofcom also notes that its multiple-choice reasons question allowed children to select options suggesting either they or the other person initiated the move. The opening example messages are illustrative and not drawn from real cases.