A separate identity seems safe behind its different name: a blog signed with a pseudonym, an account for a hobby, a nickname on a forum. For anyone who wants to know who is behind it, though, the name is the least interesting part. What counts is what that identity has in common with the others: an address, a number, a photo, a time of day, a turn of phrase.
Identity correlation attacks are the ways someone compares details from different identities to work out whether they belong to the same person. They do not require getting into your accounts: they work on what your identities let others see, and on what other services have already collected.
This post explains the most common techniques and why they become more effective when identities share details. It is the threat-side companion to the recommendation to keep separate online identities apart. The techniques used to break stronger anonymity, the kind needed by people who do not want to be linked to themselves in any way, are a different topic, covered by the post on how deanonymisation works.
A useful clarification: this post explains how these abuses work from the point of view of the people on the receiving end, so they can be recognised and prevented. They are not operating instructions.
Starting scenario
You have a profile with your name, used by friends, family and colleagues. You also have a separate identity: a pseudonym you have been writing, drawing or posting under for a few years. You set it up in a hurry, one evening, with whatever email address and phone you had to hand. Now and then you post the same photo on both. The evenings you write under the pseudonym are the same evenings you are online under your own name.
None of these details is a problem on its own. Together, though, they form a set of traces that can be laid side by side. And anyone with a reason to do so, out of curiosity, spite or self-interest, can start from just one of them.
Hardly anyone thinks about this when opening a new identity. Anyone who wants to link it to you, on the other hand, is counting on exactly that.
Why points in common increase the risk
All the techniques described here have one thing in common: they do not need to discover your name; they work it out from what your identities share.
Each shared detail works like a filter: it narrows down the group of people who could be behind the pseudonym. One detail on its own leaves many possibilities. Two or three details together can leave just one. It is the same principle by which a few ordinary details are enough to single out one person among many.
Without points in common, anyone wanting to link two identities has to rely on weak, uncertain clues. With a shared recovery email, a photo and a number, they have a good chance of getting there with little effort.
1. Linking through recovery details
In simple terms. Someone uses the email address or number a separate identity is registered with to reach the identity that carries your name.
How it works. Every account has recovery details: the address and number to which the service sends codes when you forget your password. Some services, on their recovery page, show part of those details, such as the first letters of the address or the last digits of the number, so that you can recognise them. If the separate identity and the personal one use the same address or the same number, those fragments match. And if the address itself contains your name, the link is immediate.
Why points in common help it. Shared recovery is a direct bridge between identities, which exists even when no detail is visible on the profiles.
Possible impacts. The pseudonym linked to your name; and, if someone manages to get into the recovery inbox, the loss of both identities.
Warning signs. Password reset requests you did not make; the separate identity’s notifications in your personal inbox; a sign-up address that contains your name.
How to protect yourself. Give each identity its own address, without your name, and its own recovery. Protect both with a unique password and two-step verification.
2. Uploaded contacts and your phone number
In simple terms. Your number, saved in the contacts of someone who knows you, leads that person to your separate identity.
How it works. Many services offer to upload your contacts to “find friends”, and let people search for others by phone number. If the separate identity is registered with your number, anyone who has that number in their contacts may see it suggested, or find it by searching. Services can also use other clues, such as mutual contacts; exactly how this works is not public and varies from service to service. The Electronic Frontier Foundation (EFF) guide to social networks names phone numbers and photos as the two types of information to keep an eye on when you want to keep identities separate.
Why points in common help it. The number is one of the details services use most easily to connect people, and almost everyone who knows you has it in their contacts.
Possible impacts. The separate identity suggested to colleagues, relatives or acquaintances; the link known to people who were not looking for it.
Warning signs. Someone tells you the social network suggested your pseudonym to them; the separate identity suggests people who know you by name.
How to protect yourself. Do not use your personal number for the separate identity if it is not needed; do not upload your contacts from that identity; check the settings on searching by number, where the service offers them. The recommendation on sharing your phone number only where it is needed helps you decide.
3. Photos and their metadata
In simple terms. A photo posted by the separate identity is found elsewhere, or reveals a place that leads to you.
How it works. Image search lets people find identical or very similar photos on other sites: a profile picture also used on your personal account turns up in seconds. Photos can also contain metadata, meaning information hidden in the file such as the date and sometimes the location. And even without metadata, what can be seen matters: a window, a shop sign, a recognisable object. The EFF warns that photos, in particular, can sneakily link accounts you intend to keep separate.
Why points in common help it. A shared photo is a link visible to anyone. A recognisable place is a link for anyone who knows that place.
Possible impacts. The pseudonym linked to your name; your home or workplace identified from a photo.
Warning signs. One of your photos appears somewhere you did not post it; someone comments on a place they recognise.
How to protect yourself. Use images for the separate identity that appear nowhere else, and remove location and recognisable places. The ways photos give places away are described in the post on tracking people through photos.
4. Timing and writing style
In simple terms. Someone compares when and how two identities write, and notices that they are alike.
How it works. Two identities that always post at the same times, and go quiet on the same days, can look like the same person. The same goes for the way you write: favourite words, punctuation, how you open and close messages. Style analysis, known as stylometry, compares these habits across different texts. It gives probabilities, not proof, and it works best when whoever uses it already has a short list of candidates and plenty of text from each: for example, among the people in the same team or the same club.
Why points in common help it. On their own, timing and style are weak clues. They become strong when they add up with the other links on this list.
Possible impacts. A suspicion that turns into certainty in the eyes of people who know you; the separate identity attributed to you, rightly or wrongly.
Warning signs. Someone tells you the pseudonym “writes like you”; persistent questions from a person who already knows one of the two identities.
How to protect yourself. You do not need to change your voice. Take out of your posts the details of your life that lead to you, which weigh far more than style, and avoid posting the same content on both identities within minutes of each other.
5. “Sign in with…” and connected apps
In simple terms. A service used by the separate identity knows it is yours, because you signed in with your personal profile.
How it works. Social login lets you open an account on a new site using a profile you already have. According to the guidance of one of the main account providers, the first time you use it you agree to share your name, email address and profile picture with the site. If the separate identity uses “Sign in with…” your personal profile, that site, and anyone who manages to read its data, holds the link. The ways connected apps are exploited are described in the post on attacks through connected apps.
Why points in common help it. A single profile that opens many doors joins identities you wanted to keep apart.
Possible impacts. The name and photo of your personal profile associated with the separate identity; data from both exposed if the service suffers a breach.
Warning signs. Apps or sites used by the separate identity appear in your personal profile’s settings; the site shows your real name next to the pseudonym.
How to protect yourself. For the separate identity use the dedicated address, not “Sign in with…”. The recommendation on limiting social login explains how to choose.
6. Cross-referenced data breaches
In simple terms. Data from several breached services is put together, and the separate identity appears next to the one with your name.
How it works. When a service suffers a data breach, its members’ email addresses, phone numbers and usernames can end up in collections that circulate. By comparing collections from different services, whoever holds them can spot the same number or email address under different names. The GDPR, and its UK version, the UK GDPR, require in Article 34 that the controller inform the people concerned when a breach is likely to result in a high risk to their rights and freedoms.
Why points in common help it. If identities share an email address, a number or a username, every exposed collection shows them together.
Possible impacts. The link known to whoever holds those collections; more convincing scam messages, because the sender knows both identities.
Warning signs. A breach notice from a service used by the separate identity; messages calling you by your real name in the pseudonym’s inbox or chats.
How to protect yourself. Different sign-up details for each identity and unique passwords: that way an exposed collection shows only one identity at a time. After a breach notice, change the password and check the recovery details of the account involved.
Summary table
| Technique | Main risk | Warning signs | Effective defences |
|---|---|---|---|
| Recovery details | A direct bridge between identities | Unrequested resets, notifications in the wrong inbox | Dedicated address and recovery |
| Contacts and number | Identity suggested to people who know you | Suggestions to colleagues and relatives | No personal number, no uploaded contacts |
| Photos and metadata | Visible link, places identified | Photos found elsewhere, places recognised | Unique images, no location |
| Timing and style | Suspicion that becomes certainty | “Writes like you”, persistent questions | No life details, no twin posts |
| “Sign in with…” | The service knows it is you | The identity’s apps in your personal profile | Dedicated address, no social login |
| Cross-referenced breaches | Identities shown together in collections | Breach notices, your name in the pseudonym’s chats | Different sign-up details, unique passwords |
What they have in common
Six different techniques, three defences that work against almost all of them:
- No shared sign-up details — its own address, recovery, number and password for each identity remove the most direct links.
- No shared content — photos, posts and places that appear on one identity only cannot be compared.
- No life details — your neighbourhood, your job, the people close to you lead to you more than any style analysis.
These are not advanced measures: they are habits you adopt when you open an identity, and check again now and then. If, however, you have reason to fear someone who is looking for you persistently, for example someone close to you who is monitoring you, the first step is not technical: in England, GOV.UK lists Refuge’s National Domestic Abuse Helpline, free and open 24 hours a day.
Protection checklist
- ☐ Each separate identity has an address without my name and its own recovery
- ☐ Each identity has a unique password and two-step verification
- ☐ I do not use my personal number for the separate identity if it is not needed
- ☐ I have not uploaded my contacts from the separate identity
- ☐ The separate identity’s photos appear nowhere else and carry no location
- ☐ I do not use “Sign in with…” my personal profile for the separate identity
- ☐ I reread my posts and take out details of my life
How this connects to the Cyber Welfare Framework
| Pillar | How this content contributes |
|---|---|
| Awareness | Understanding that two identities are linked back together by the details they share, and that links add up |
| Skills | Recognising the techniques, from recovery details to uploaded contacts, photos and cross-referenced breaches |
| Secure Behaviour | Keeping sign-up details, content and life details distinct for each identity |
Reference level: FL3 — Autonomous, with elements of FL2 — Beginner in the parts on the dedicated address and recovery.
Conclusion
The techniques described here do not need to get into your accounts. They rely on ordinary details, an address, a number, a photo, that identities share for convenience.
That is why the most effective defence is not to give up pseudonyms, but to separate: sign-up details of their own, content of their own, no life details. To find out whether your identities are already crossing over, the next step is the signs your identities are crossing over; for a wider view of your digital habits, you can start with the digital resilience self-assessment.
Something to think about. If someone knew only your separate identity’s profile picture, how long would it take them to reach your name?
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
- Metadata: The Invisible Traces That Tell Who You Are
Related content
- Keep separate online identities apart — the recommendation this post builds on
- Signs your identities are crossing over — the indicators to check
- How to manage multiple online identities — the steps to put things right and prevent problems
- Digital compartmentalisation — the tools that keep identities apart
- What an exposed pseudonym can lead to — what these techniques cause when they succeed
- Attacks through connected apps — the related topic of abusing permissions granted with “Sign in with…”
Start with the first step: the Cyber Welfare Programme guides you free of charge, one recommendation at a time.




