[Beginner] — how people actually lose money, and the two habits that stop nearly all of it.
The kid version first
Almost nothing that reaches you is a technical attack. It’s a message designed to make you act before you think.
The whole craft is manufactured urgency: your account will be closed, your parcel is held, there’s suspicious activity. Panic is the exploit. Everything else is decoration.
Two habits stop most of it:
- Never act from a link in a message. Open the app or type the address yourself.
- Verify urgency through a channel you chose — a number you looked up, not the one in the message.
The tells
Urgency and threat. “Within 24 hours or your account will be suspended.” Real organisations don’t work this way. Manufactured time pressure exists to prevent you thinking, and it’s the single most reliable indicator.
An unexpected link or attachment, even from someone you know — their account may be the one that’s compromised.
A request that breaks procedure. “I’m in a meeting, can you just pay this invoice quickly?” — the boss who never texts, texting.
Slightly wrong details. paypa1.com, micros0ft-support.net, a display name that doesn’t match the address behind it. Hover over links before clicking; on mobile, long-press to preview.
Wrong tone or context. A bank that suddenly uses your first name oddly, or a service you don’t use.
Two old tells that no longer work: bad spelling and poor design. AI has removed both. Modern phishing is well written and pixel-perfect, so “it looked professional” is no longer evidence of anything.
The varieties
Bulk email phishing — millions of copies. Fake login pages for Microsoft 365, banks, delivery firms.
Spear phishing — targeted, researched, using real details about you or your workplace. Far more convincing, far fewer copies.
Smishing (text) and vishing (voice). “Your parcel couldn’t be delivered — pay £1.99.” Voice is now the dangerous one: AI voice cloning means a distressed relative’s voice can be faked from a few seconds of social media audio.
Agree a family code word. A word only your family knows, for verifying an urgent call. It costs one conversation and defeats voice cloning entirely.
Business email compromise (BEC) — the most expensive category in the world by value. A supplier emails new bank details for an outstanding invoice. The email is real, from a compromised account, and the details are the attacker’s. Verify any change of payment details by phone, on a number you already had.
Tech support scams — a pop-up or call about a virus, leading to remote-access software and “refund” theatre. Microsoft and Apple will never call you.
Romance and investment fraud (“pig butchering”) — weeks or months of genuine-feeling relationship, then an investment opportunity on a fake platform showing fake gains. Small withdrawals succeed; large ones require “fees”. Often operated from trafficked-labour compounds — organised crime, not opportunists. The largest category by money lost, and the victims are not foolish; they were groomed for months.
Recovery scams — someone who “can get your lost money back”, targeting people already defrauded. Being scammed once puts you on a list.
Marketplace and rental scams — the deposit for a flat you can’t view, the item that never ships.
The rules
1. Never use the link. Ever.
Not because you can’t spot fakes, but because you don’t have to. Open your banking app. Type the address. Use your bookmark. Nothing legitimate is lost by arriving under your own steam.
2. Verify out of band.
Urgent request about money or credentials → contact them on a number or address you already had. Never the contact details in the message. If it’s genuine, a five-minute delay costs nothing.
3. Nobody legitimate asks for these:
- Your password
- A code that was sent to you → two-factor
- Remote access to your computer
- Payment in gift cards, crypto, or bank transfer to “keep your money safe”
- A crypto seed phrase → the crypto fraud surface
4. Slow down when you feel urgency. The feeling of needing to act right now is the attack working. Treat it as the alarm itself.
5. Use a password manager as a phishing detector. It fills credentials based on the domain — so if it doesn’t offer to fill, that’s not the real site. A genuinely useful side effect.
If you clicked
Don’t panic, and don’t stay quiet — shame is what turns a small incident into a large one.
- Entered a password? Change it immediately, everywhere you reused it, and sign out all sessions → when it goes wrong
- Entered card details? Call your bank, freeze the card
- Installed something or gave remote access? Disconnect from the network, and treat the device as compromised — a wipe and reinstall is the only reliable fix
- Sent money? Contact your bank immediately — same-day recall sometimes works. Report it (Action Fraud in the UK, IC3 in the US, your local equivalent)
- At work? Tell IT now. They have seen it before and would far rather hear it early. Concealment is the thing that causes real damage
If you’re helping someone else
Most people reading this will at some point be the person a parent or friend calls.
- Don’t lead with “how did you fall for that”. They’ll stop telling you things, and that’s worse than any single scam
- Set the six basics up for them → the six things. Doing beats explaining
- Give them one rule they’ll remember: “If it’s urgent and about money, hang up and call me.”
- Agree the family code word
- Consider being a trusted contact on their accounts
Key insight
Phishing targets the person, not the computer — so no amount of software fixes it, and no amount of technical skill makes you immune. What works is procedural: never act from a link, and verify anything urgent through a channel you chose yourself. Those two habits require no expertise and defeat the overwhelming majority of what will ever be aimed at you.
Related
- passkeys — the only defence that’s phishing-proof by design
- when it goes wrong
- social engineering — the professional treatment
- using AI — why the spelling-mistake tell stopped working
Software engineering
Infrastructure & security
Computer science
Hardware & engineering