The wrong starting point
Most people discover they want better digital security after an incident — a hacked account, a data breach notification, or one too many targeted ads that feel suspiciously accurate. The instinct is to immediately start installing things: a VPN, a password manager, maybe Signal. That’s not a bad instinct, but it’s backwards.
The right starting point isn’t “what tools should I use?” It’s “what am I actually protecting, and from whom?”
That question is called a threat model, and working through it honestly will tell you more about your security posture than any app.
What a threat model is (and isn’t)
A threat model is a structured way of thinking about risk. It answers four questions:
- What do you want to protect? Your email archive, financial records, communication history, browsing habits, physical location.
- Who might want access to it? An ex-partner, a landlord, an employer, a targeted phishing campaign, a government agency, an opportunistic criminal.
- What happens if they get it? Embarrassment, financial loss, professional consequences, physical danger.
- How much effort are you willing to spend? Because every security measure has a usability cost, and that cost compounds.
Notice that “everything from everyone forever” isn’t on that list. It can’t be, because it’s not achievable — and chasing it is a reliable way to exhaust yourself while protecting nothing well.
Why “perfect privacy” is the wrong goal
Perfect privacy would require never using any networked device, never communicating with anyone, and never appearing in any database. It would also make modern life impossible. This is not the bar you should be aiming for.
The useful question is: what are the realistic threats given who you are and how you live? A journalist covering corruption has a different threat model than a teacher managing a class roster. A person leaving an abusive relationship has different priorities than someone who’s annoyed that LinkedIn keeps guessing their interests.
For most private individuals, the realistic threat landscape looks like this:
- Mass surveillance and data brokerage — the broad, commercial kind, where your behavior is sold as a product. Browser compartmentalization, DNS filtering, and ad-blocking address the majority of this.
- Credential compromise — phishing, database breaches, reused passwords. A password manager and two-factor authentication address roughly 90% of this category.
- Device theft — full-disk encryption and a strong PIN handle this.
- Social engineering — someone calling, emailing, or messaging to trick you or someone who has access to your accounts. This is a human problem that requires training and procedures, not tools.
Scoping it honestly
Once you’ve identified your realistic threats, scope your defenses to match. You don’t need to protect against nation-state adversaries if you’re trying to prevent a credential-stuffing bot from accessing your email.
This isn’t defeatism — it’s resource allocation. Security attention is finite. Spending it on the wrong threats leaves real ones unaddressed.
A few principles that hold across most individual threat models:
- Separate identities where exposure is asymmetric. Use a different email address for financial accounts than for newsletters. Separate work and personal browsing. Isolation limits blast radius.
- Protect what can’t be changed. Your government ID number, biometrics, and medical history are permanent; a password is not. Prioritize accordingly.
- The threat model evolves. A change in job, relationship, or public profile changes your exposure. Revisit annually, or after major life changes.
A place to start
If you’ve never worked through this deliberately, here’s a minimal starting exercise: write down three things you’d genuinely not want an adversary to access, and one type of adversary that actually concerns you. That’s enough to start making better decisions about where to focus.
From there, the tools follow naturally — not the other way around.