A made-up name, an unsigned comment, an “anonymous” survey: they all look like places where who you are does not count. For anyone who wants to know who is behind them, though, each of these places leaves clues. Nobody needs to break into an account or steal a password: often it is enough to put together what is already visible.
Deanonymisation is the process by which someone links an activity, an account or a piece of data that looked anonymous back to a specific person. Sometimes it is the work of a researcher studying “anonymised” data. More often, in everyday life, it is a curious, angry or interested person cross-referencing scattered details: a photo, a time of day, a username, a way of writing.
This post explains the most common techniques and why they work better when the level of anonymity chosen does not match the situation. It is the threat side of the recommendation on how to decide how much anonymity you need. How to recognise the signs that it is happening is the subject of the post on the signs your anonymity is slipping.
One useful clarification: this post explains how these techniques work from the point of view of the person on the receiving end, so that you can recognise and prevent them. It contains no operational instructions.
Starting scenario
For three years you have been writing, under a made-up name, on a forum about your hobby. You have never written your name. But in three years you have mentioned plenty of small things: that you work in an office near the station, that you have a dog of a particular breed, that you go swimming on Tuesday evenings, that you went on holiday to a mountain village in August. You have posted a few photos: the dog in the park, the view from your office window. Your username is the same one you use on a gaming app.
None of these details, on its own, says who you are. Together, though, they describe a person whom, in a medium-sized town, only a handful of people might match. And anyone who knows you, even by sight, might need nothing more.
Usually nobody does this on purpose. Anyone who wants to know who you are, on the other hand, is counting on exactly this.
Why mismatched anonymity increases the risk
All the techniques described here have one thing in common: they are not looking for a secret, they are looking for a link. They work best when the protection chosen is weaker than you think, for example a pseudonym believed to be anonymous, or when it is so complicated that sooner or later you slip up.
One concept helps to understand the mechanism: the anonymity set, meaning the group of people who could be you. At the start it is enormous. Every detail shrinks it. A study of 1990 US census data estimated that postcode, sex and date of birth, taken together, were enough to single out 87% of the population. Another, from 2013, using fifteen months of mobility data from one and a half million people, showed that four points of place and time were enough to single out 95% of individuals. The principle is the same for the details you leave on a forum.
1. Cross-referencing a few details
In plain terms. Someone lines up details you shared at different times until the group of people who match shrinks to one.
How it works. Each detail, such as your town, your job, a recurring time, a pet, a trip, acts as a filter. In a page published in February 2024, the CNIL, France’s data protection authority, describes exactly this cross-referencing of public information: scattered data, none of which identifies anyone on its own, can reveal who is behind a pseudonym, where they work and where they live. The same mechanism underlies the re-identification of data published as “anonymised” from which only the name had been removed.
Why mismatched anonymity helps it. People who believe their pseudonym is enough share details less carefully. The longer they use it, the more filters they pile up.
Possible impacts. Your name linked to opinions, activities or groups you wanted to keep apart.
What to watch for. Someone asks you about details you wrote in different posts; someone you know quotes something you wrote under a pseudonym.
How to protect yourself. For each pseudonym, write a short list of details you will not use: workplace, area, names of relatives, personal dates. The resource on why a few details can identify you explains how to estimate how quickly the group shrinks.
2. Linking accounts
In plain terms. Something two accounts have in common joins them up: an email, a number, a username, a photo, an address book.
How it works. An identical username on two services is the simplest bridge. But there are others: the same recovery phone number, an uploaded address book that makes the account appear among your contacts’ suggestions, a sign-in through the “Sign in with…” button that ties a service to your main profile, a profile photo used elsewhere that a reverse image search finds again. That is why the CNIL advises using different pseudonyms and email addresses for the accounts you want to keep apart, and not posting the same photo on several platforms.
Why mismatched anonymity helps it. For convenience, a pseudonym is often opened with the usual email and number. The name is new; nothing else is.
Possible impacts. The account is linked to you by anyone who sees the bridge: the service, your contacts, anyone searching for the username.
What to watch for. The pseudonym appears among the suggestions of people you know; searching for your username also turns up your other profiles.
How to protect yourself. Separate recovery details for the pseudonyms that matter, no “Sign in with…”, usernames and photos never used elsewhere. Why it is worth limiting sign-in through your social media account is explained in the recommendation on limiting social login; how identities kept apart get linked, in the post on identity correlation attacks.
3. Metadata in files and communications
In plain terms. Files and messages carry information you cannot see: where, when, with what, with whom.
How it works. A photo can contain the coordinates of the place where it was taken and the model of the phone. A document can contain the author’s name and editing dates. Even when the content of a message is encrypted, information about who is talking to whom, and when, remains: the Electronic Frontier Foundation’s (EFF) digital self-defence guide points out that end-to-end encryption does not protect this metadata.
Why mismatched anonymity helps it. People who focus on the name forget the file. A photo posted under a pseudonym, with the location still inside it, says more than any post.
Possible impacts. The location of your home or office; the name your computer is registered under; how regularly you are in contact with someone.
What to watch for. You post photos straight from your phone’s gallery; you share documents without checking them.
How to protect yourself. Remove the location from photos, as the recommendation on turning off photo geotagging explains, and check documents before sharing them, as described in the post on document metadata.
4. Writing style
In plain terms. The way you write, the words you prefer, your punctuation, your recurring mistakes, can resemble that of a text you signed.
How it works. The study of style to attribute a text is called stylometry. It gives probabilities, not certainties. Research presented in 2012 at a computer security conference tried to identify the author of a text among 100,000 blogs: the first name suggested was the right one in just over 20% of cases. It is not proof, but for someone who already suspects a particular person, it may be enough to firm up an idea.
Why mismatched anonymity helps it. People who write a lot, for a long time and about the same topics under two different names leave more material to compare.
Possible impacts. A suspicion that becomes a conviction, right or wrong; a mistaken attribution to someone else.
What to watch for. Someone notices that you “write like” another person; you use the same tone and expressions in both settings.
How to protect yourself. For what really needs to stay separate, write less and more neutrally, and avoid expressions you also use under your own name. You do not need to do this for everything: only for the rows on your sheet that call for it.
5. Social engineering
In plain terms. Someone asks you, kindly or on some pretext, for exactly the details that are missing.
How it works. Social engineering is the set of techniques used to persuade someone to hand over information. Where anonymity is concerned, it might be a friendly user who asks in private which area you live in “to recommend a place”, a fake survey, or a message pretending to come from the service and asking you to “confirm” a phone number.
Why mismatched anonymity helps it. Inside a community you know, people lower their guard: the pseudonym feels like a safe place, and the questions seem normal.
Possible impacts. The missing detail that completes the picture; sometimes, access to the account.
What to watch for. Persistent questions about where you live, work or study; requests to move to another app; “official” messages asking for details.
How to protect yourself. Answer the question, not the details: you can talk about a topic without saying where you live. Be wary of messages asking you to confirm details: the real service does not usually do that in a chat.
6. Doxxing
In plain terms. Someone publishes your personal information, such as your name, address, job or photos, to intimidate or embarrass you.
How it works. Doxxing is often the last step after the techniques above: having linked the pseudonym to you, someone makes the link public. Police Scotland, for example, describes it as sharing someone’s information online without their consent, usually to scare or punish them, and advises reporting the content to the platform, keeping the evidence, blocking the sender and contacting the police.
Why mismatched anonymity helps it. The more links are left open, the easier it is to complete the picture and make it public.
Possible impacts. Unwanted messages, tension at work or in the family, fear; in the most serious cases, contact in real life.
What to watch for. Your details appear in a group or on a forum; someone writes to you mentioning where you live or work.
How to protect yourself. Do not reply to whoever is provoking you, keep screenshots and links, report it to the service and, if there are threats, contact the police (101, or 999 in an emergency). If the person exposing you is someone close, GOV.UK lists specialist helplines, such as Refuge’s National Domestic Abuse Helpline (0808 2000 247, free, 24 hours a day). How to reduce the information others can find is covered in the resource on how to prevent doxxing.
Summary table
| Technique | Main risk | What to watch for | Effective defences |
|---|---|---|---|
| Cross-referencing a few details | The group you blend into shrinks to you | Questions about scattered details, quoted sentences | A list of details you will not use |
| Linking accounts | Email, number, username or photo in common | Suggestions to your contacts, profiles found together | Separate recovery details, no “Sign in with…” |
| Metadata | Location, author and contacts inside files and messages | Photos posted from the gallery, unchecked documents | Removing location and metadata before sharing |
| Writing style | A suspicion strengthened by resemblance | “You write like…” | Writing less and more neutrally where it matters |
| Social engineering | The missing detail, asked for on a pretext | Persistent questions, fake official messages | Talk about the topic, not about yourself |
| Doxxing | Personal details made public | Details in groups or forums, messages mentioning where you live | Evidence, reporting, specialist help |
What they have in common
Six different techniques, three defences that work against almost all of them:
- A clear sheet — knowing what you want to protect and from whom tells you where care is needed, and where it is not.
- No bridges — separate emails, numbers, usernames, photos and sign-ins for the pseudonyms that matter.
- Few details — every detail shrinks the group you blend into; the ones you never write cannot be cross-referenced.
These are not advanced measures: they are habits you can pick up in a few evenings.
Protection checklist
- ☐ I have written the sheet for the pseudonyms that matter
- ☐ Every important pseudonym has email and recovery details that do not lead to me
- ☐ I do not reuse usernames and profile photos across settings I keep apart
- ☐ I have a list of details I do not write under a pseudonym
- ☐ I remove location and metadata from photos and documents before sharing them
- ☐ I do not answer persistent questions about where I live or work
- ☐ I know whom to contact if someone publishes my details
How this connects to the Cyber Welfare Framework
| Pillar | What this content contributes |
|---|---|
| Awareness | Understanding that deanonymisation looks for links, not secrets, and that a few details can be enough |
| Skills | Recognising the techniques, from cross-referencing details to metadata, from writing style to social engineering |
| Secure Behaviour | Keeping a clear sheet, no bridges between separate settings and few details where it matters |
Reference level: FL3 — Autonomous, with elements of FL2 — Beginner in the parts on the sheet and on the difference between a pseudonym and anonymity.
Conclusion
The techniques described here do not need to get into your accounts. They rely on what is already visible: scattered details, bridges between accounts, metadata, habits, trust. They work above all when a pseudonym is believed to be stronger than it is.
That is why the most effective defence is not to disappear, but to choose: a clear sheet, no bridges, few details where it matters. To find out whether a link is already opening, the next step is the signs your anonymity is slipping; for a wider view of your digital habits, you can start with the digital resilience self-assessment.
Something to think about. If someone read every post of your oldest pseudonym, how many details would they need to reach you?
Related resources
Short reads from the Resources section, for anyone who wants to focus on a single aspect:
- Re-identification: Why a Few Details Can Identify You
- How to Prevent Doxxing: Spot the Signs, Reduce Your Exposure
Related content
- Deciding how much anonymity you need — the recommendation this post comes from
- Signs your anonymity is slipping — the indicators to check
- How to protect your anonymity online — the steps to prevent and put things right
- Anonymity technologies — what each tool protects
- What happens when anonymity breaks — what these techniques cause when they succeed
- Identity correlation attacks — the related topic of how identities kept apart get linked
Start with the first step: the Cyber Welfare Programme guides you free of charge, one recommendation at a time.




