How to Choose a Password Manager (and Why You Actually Need One)

If you reuse the same password across sites — or keep them in your head, a notes app, or your browser — you already have a security problem, whether or not anything bad has happened yet. The average person now juggles well over a hundred logins, and the human strategy for that (reuse a few passwords, tweak them slightly) is exactly what attackers count on. A password manager fixes the root problem: it generates and remembers a different strong password for every account, so a breach at one site can't unlock the rest. This guide covers why that matters, whether these tools are actually safe after high-profile breaches, and how to choose one — with a clear-eyed look at what really keeps your vault secure.

Our full terms are on our disclaimer page.

Who this is for, and how we chose what to cover

This is for an ordinary person with too many logins and no security background, deciding whether to start using a password manager and which one.

How we chose what to put in this guide:

• We lead with the three things the reader controls, because they determine the outcome far more than the product does — and every "best password manager" article inverts that.
• We do not rank or score the managers. We have tested none. More importantly, no public, current, comparable security assessment exists across these products: audits differ in scope, are commissioned by the vendor, and cannot be turned into a league table without inventing a methodology.
• We name them anyway with links, because "pick a reputable manager" is not something you can act on without knowing which ones exist.
• We give the methods behind the criteria — how to build a passphrase, how to check for breaches, how to migrate — because a criterion without a method is only half an instruction.

Who this is not for: if you already use a manager with a strong unique master passphrase and 2FA, you are done. The rest of this article will not improve your security materially.

Why a password manager, and why now

The core threat is credential stuffing: attackers take username/password pairs leaked in one breach and try them on dozens of other sites, because so many people reuse passwords. One reused password, one breach, and every account sharing it is exposed — which is also how a great deal of "identity theft" actually begins, rather than through anything exotic. This is why NIST's digital identity guidance (SP 800-63B) moved away from forced complexity rules and periodic mandatory changes toward what actually helps — long, unique passwords for every account, checked against known-breached lists, plus multi-factor authentication. That is precisely what a password manager makes practical: it creates a unique 16+ character password per site and fills it for you, so uniqueness stops being a memory problem. CISA states the same guidance in consumer terms.

How large is the underlying problem? Verizon's annual Data Breach Investigations Report has for years found the "human element" — stolen credentials, phishing, error — behind the majority of breaches. The exact percentage moves each edition and the methodology changed part-way through the series, so treat any single figure you see quoted as a snapshot of one year rather than a constant. The direction has been consistent, and that is the part worth acting on.

Bottom line: the real win isn't convenience, it's uniqueness at scale — a different strong password everywhere, which is the single most effective thing an ordinary person can do against account takeovers.

"Isn't my browser's saved passwords good enough?"

Usually not, and the reason is concrete. Browsers do store passwords in an encrypted database — but the encryption key is often left unprotected on your device, which security researchers compare to hiding the house key under the doormat. Two real consequences: infostealer malware (families like RedLine and Raccoon) specifically targets browser password stores and can exfiltrate the whole set; and on a typical Windows PC, anyone who sits at your unlocked computer can view your Chrome-saved passwords in plaintext from the settings page — there's no separate vault password. A dedicated manager keeps a separate, encrypted vault behind its own master password, so a compromised browser or device doesn't hand over everything.

In short: the browser feature is convenient but weakly protected; a dedicated manager's separate master-password vault is a meaningful security upgrade, not just a nicer interface.

A fair word for the OS-level managers, since they are not the same thing as the browser store: Apple's Passwords and Android's Google Password Manager are genuinely better protected than a browser's saved-password list, are encrypted, and support passkeys. Their real limitation is portability — they work beautifully inside one ecosystem and awkwardly outside it. If every device you own and will own is on one platform, and you never need to share a credential with someone on another, an OS manager is a legitimate choice and much better than reuse. For everyone else, cross-platform is worth the small effort.

How they work — and what "zero-knowledge" really means

The mechanism is honest and worth understanding. Your master password never leaves your device; the manager runs it through a slow key-derivation function (PBKDF2, scrypt, or Argon2) to produce an encryption key, and that key encrypts your vault on your device with strong ciphers (AES-256 or XChaCha20). The company's servers store only the encrypted blob — this is what "zero-knowledge" means: the provider can't read your passwords because it never has your key.

But be precise about the limit, because the marketing isn't: zero-knowledge protects your data if the provider's servers are breached — an attacker gets ciphertext, not plaintext. It does not make a weak master password safe. If your master password is short or reused, a stolen encrypted vault can be attacked offline until it cracks. Recent security research has also shown some managers' "zero-knowledge" implementations are less airtight than promised, which is exactly why independent audits matter.

Bottom line: "zero-knowledge" is a real and valuable property, but it's a floor, not a force field — its protection is only as strong as your master password and the quality of the implementation.

The flagship: the three levers that actually determine your safety

Choosing between reputable password managers matters far less than three things you control. Get these right and almost any well-regarded manager is fine; get them wrong and the best product on the market won't save you.

Worth separating this from network privacy, which is a different problem with different tools — a VPN does nothing about a reused password, and a password manager does nothing about your ISP.

The point: the first two are entirely yours. The third is the only one that's genuinely a product-choice question — and every mainstream manager gets it broadly right, which is why the choice between them matters less than the marketing suggests.

Lever 1, done properly: how to actually build a master passphrase

"Several unrelated words" is a criterion, not a method, and the difference matters enormously — because words you choose yourself are not random, and an attacker's guessing model knows that. People pick words that relate to each other, to their interests, and to things visible from where they are sitting.

The method is to let something else pick the words. Either use your password manager's own passphrase generator (most have one, set it to five or six words), or use physical dice with a numbered word list — the technique usually called diceware. Both produce words drawn genuinely at random from a large list, which is the whole point.

Why five or six words rather than three. With a standard 7,776-word list, each randomly chosen word contributes about 12.9 bits of entropy. So:

Three words is not enough against an offline attack on a stolen vault, which is the LastPass scenario exactly. Five is a reasonable floor and six is comfortable, and the difference in what you have to memorise is one word.

Three practical rules: write it down on paper and keep that paper somewhere physically safe until you have it memorised — the risk you are managing on day one is forgetting it, not a burglar reading it; type it fifteen or twenty times deliberately on the first day, which is what actually commits it to memory; and never reuse it anywhere, because a passphrase that also unlocks your laptop is one breach away from being ordinary.

Bottom line: generate it, don't invent it. Five or six words. Practise it the day you set it.

Lever 2, done properly: which second factor

In preference order: a hardware security key (strongest, phishing-resistant, costs money), an authenticator app or your manager's built-in TOTP (very good, free), then SMS (weak against SIM-swap attacks, and still much better than nothing).

The bootstrapping trap to avoid: do not store the only second factor for your password manager inside the password manager itself. If your vault is locked behind a code that lives in the vault, you have made a circle. Keep the manager's own 2FA on a separate device, an authenticator app, or a hardware key — and keep its recovery codes with your master passphrase, on paper.

What the LastPass breach actually teaches

In 2022, attackers reached LastPass's systems and eventually made off with backup copies of customer vaults — the passwords inside were encrypted, but some account metadata was not. The important, non-panic lesson: the encryption held, so users with strong, unique master passwords were protected; the users genuinely at risk were those whose master password was weak or reused, because a stolen encrypted vault can be attacked offline indefinitely.

Bottom line: the breach is a case for the three levers, not against password managers — it's real-world proof that your master password strength and the vault's encryption are what stand between an attacker and your accounts.

What to look for (the checklist)

✓ Zero-knowledge, end-to-end encryption · ✓ Recent independent security audit · ✓ 2FA support · ✓ Breach/weak-password monitoring · ✓ Cross-device sync · ✓ Emergency/recovery access · ✓ Passkey support

A note on two common questions:

• Open-source vs. proprietary: open-source code can be independently inspected, which many security-minded users prefer, but a well-audited proprietary manager is also legitimate. Audits matter more than the license label.
• Passkeys: the newer passwordless standard (a login tied to your device biometrics) is rolling out, and good managers now store and sync passkeys. You still want a manager, because you'll have both passkeys and ordinary passwords for years. The FIDO Alliance, which publishes the standard, explains what they are.

What an audit does and does not tell you. A published third-party audit is a real positive signal and it is not a certificate of safety. It covers a defined scope at a point in time, it is commissioned and paid for by the vendor, and its findings may be summarised rather than published in full. What to look for is the pattern, not the badge: does the provider publish audits regularly, name the auditor, and describe what was fixed? A vendor that publishes a report containing findings it then remediated is telling you more than one that publishes a clean summary.

What to check: buy on verified security (audits, zero-knowledge, 2FA) and the recovery/sync features you'll actually use — not on a feature-count arms race.

The managers themselves, named and unranked

Five a general reader will encounter, alphabetically, not ranked, not tested by us, with no price asserted here. Each link goes to that company's own page:

• 1Password — proprietary, paid, cross-platform, family plans
• Bitwarden — open-source, capable free tier, paid upgrade, self-hosting possible
• Dashlane — proprietary, paid, cross-platform
• KeePassXC — open-source, free, and a local file with no cloud account at all (see the alternatives below)
• Proton Pass — open-source, free tier, bundled with a wider privacy suite

What to compare across them: whether the free tier syncs across the device types you actually use (this is the most common reason a free tier turns out not to be sufficient); the recovery mechanisms offered; whether family sharing is included; whether passkeys are supported on your platforms; and the renewal price after any introductory year.

Free vs. paid

A reputable free tier is genuinely sufficient for most individuals — the core security (strong encryption, unique passwords, 2FA) doesn't cost extra. Paid tiers mostly add convenience and sharing: family vaults, more secure sharing, richer breach monitoring, priority support. And note the honest point: even well-regarded managers get scrutinized — the security research mentioned above examined several of them — which is exactly why you want one that publishes audits and fixes what they find, not one that claims to be flawless.

Bottom line: start with a reputable free or low-cost manager today — the security that matters is in the free tier; upgrade only for features you'll use.

What you need before you start

• A device you control and trust. If you suspect the device itself is compromised, fix that first — a manager on an infected machine protects nothing.
• Access to your email account, since it is used for setup and recovery. If your email password is currently reused, this is the one to change first.
• A piece of paper and a pen, for the master passphrase and recovery code. This is not a joke — paper is offline, and offline is the point.
• About an hour for setup and the first pass at migration. Not a weekend: see the priority order below.
• A second device if you want cross-device sync working from day one.

A worked example: moving 120 logins without losing a weekend

Take the case of someone with about 120 accounts accumulated over fifteen years, most of them in Chrome, several sharing four variations of the same password. The instinct is to sit down and change all 120. Almost nobody finishes that, and the ones who abandon it halfway end up worse off than when they started, with a half-populated vault and no idea which passwords are where.

Here is the version that finishes.

Hour one — set up and import.

1. Install the manager, generate the master passphrase (five or six words, generated not invented), and write it on paper.
2. Turn on 2FA for the vault itself, and store its recovery codes with the passphrase.
3. Import the browser's saved passwords. Every manager has an importer; this takes minutes and brings all 120 in at once, weak passwords and all. You have not improved anything yet — you have made the problem visible, which is the prerequisite for fixing it.
4. Run the manager's built-in weak-and-reused-password report. This is the to-do list, sorted by risk.

The rest of hour one — change the six that matter. Not 120. Six, in this order:

1. Email — it is the reset route to everything else, so it is not merely one account among many.
2. Your password manager (if the passphrase you just made replaced an old reused one).
3. Banking and anything holding money.
4. Your phone-carrier account — because control of your phone number defeats SMS 2FA everywhere else.
5. Primary cloud storage.
6. Anything that is currently reused and holds a payment card.

Week one — clear the browser store. Delete the saved passwords from the browser and turn off its offer to save new ones. Skipping this is the most common half-migration: two copies of everything, one of them weakly protected, and no idea which is current.

Month one — change the rest opportunistically. Do not schedule this. Each time you log into a site, if the manager flags the password as weak or reused, change it then. It costs ninety seconds in context, and after a month the accounts you actually use are all unique. The accounts you never log into are, by definition, the ones with the least at stake.

What this assumes, and what would change it. It assumes none of the accounts is already compromised — if you know one is, that one goes first regardless of this order. It assumes your email is not currently sharing a password with a breached site; if it is, that changes from step one to right now. And it assumes a personal, not a work, context: work accounts follow your employer's policy, and importing them into a personal vault may breach it.

Bottom line: the migration that finishes is the one that changes six passwords in the first hour and the rest by accident over a month.

How to know it's actually working

Four checks, and the last is the one nobody runs until they need it.

1. Open the manager's security dashboard. Reused and weak counts should be falling week over week. If they are not, the manager is a filing cabinet rather than a security tool.
2. Check your email address at Have I Been Pwned. It tells you which breaches your address appeared in, which tells you which accounts to prioritise. What it cannot tell you is whether a specific current password is safe — absence from a list is not evidence of safety, and many breaches are never published.
3. Confirm the browser is no longer saving passwords. Open its settings and look. Two stores is one too many.
4. Test your recovery — deliberately, while everything is fine. Log in on a device you have not used before, using only what you would have if your phone were lost: the passphrase from memory and the recovery code from paper. If you cannot get in, you have discovered it on a Tuesday rather than in an emergency. This is the single most valuable ten minutes in this article and almost nobody spends it.

The test six months on: could you get into your vault today if your phone fell in a river? If the answer is not an immediate yes, recovery is the thing to fix, not your password strength.

What happens to your vault if you die

The most-asked question in this category and the least-answered, because the honest answer is uncomfortable: with a genuine zero-knowledge manager, if nobody has your master passphrase, your accounts are unreachable — permanently. That is the design working as intended, and it is a real problem for the people who survive you.

Three approaches, and most households want one of them:

• The manager's own emergency access. Several offer a named contact who can request access, with a waiting period during which you can decline. This is the cleanest option where it exists — set it up when you enrol.
• A sealed physical copy. The master passphrase and recovery code, written down, in a safe or with a solicitor, with instructions. Low-tech, entirely effective, and no worse than where your other important documents live.
• A shared vault with a partner for the credentials that genuinely need to be joint — utilities, shared finances — kept separate from your personal vault.

What not to do is email it to yourself, put it in a note on your phone, or assume a family member can talk the provider into it. They cannot; that is the whole property you selected the product for.

If you don't want a cloud provider at all

A legitimate position, and there is a real option for it: a local-file manager such as KeePassXC keeps the encrypted vault as a file on your own machine, with no vendor account and nothing synced anywhere unless you arrange it.

The honest trade-off. You gain: no third party holds your ciphertext, no vendor breach can touch you, and no subscription. You take on: sync and backup become entirely yours. A vault file on one laptop is one hardware failure from total loss, so you need a backup discipline you will actually maintain. For most people the cloud managers' zero-knowledge design is the better trade; for a technically confident user with a backup habit, local-only is genuinely more private and no less secure.

Common mistakes

• A weak or reused master password. The one unforgivable mistake — it undoes everything (the LastPass lesson).
• Skipping 2FA on the vault. Your most important account to protect; protect it hardest.
• Trusting the browser store for everything. Convenient, weakly protected, and a prime malware target.
• Losing your recovery options. If you forget the master password of a true zero-knowledge manager, no one can reset it for you — set up the manager's recovery/emergency access when you sign up, not after.
• Choosing on brand hype over audits. A published independent audit beats a marketing superlative.
• Inventing the passphrase instead of generating it. Human-chosen words are far more guessable than they feel.
• Importing everything and clearing nothing, leaving two copies of every password, one of them weakly protected.
• Storing the vault's own 2FA seed inside the vault, which makes the second factor circular.
• Trying to change all 120 passwords at once, and abandoning the migration halfway.
• Never testing recovery, and discovering it does not work at the worst possible moment.

Putting it together

Reuse is the real vulnerability, and a password manager solves it by making a unique strong password per account effortless. Whether these tools are "safe" comes down to the three levers you control: a long unique master passphrase, 2FA on the vault, and an audited zero-knowledge manager. Get those right — as the LastPass breach showed — and even a server-side compromise leaves an attacker with unreadable ciphertext. Pick any reputable, audited manager that fits your devices, set it up carefully (especially recovery), and you've closed the single biggest gap in most people's personal security.

Your next three moves, in order: (1) pick any manager from the table and generate — do not invent — a five or six word master passphrase, on paper; (2) turn on 2FA for the vault and store its recovery codes with that paper; (3) change six passwords tonight, starting with email, and let the rest happen over the next month.

Where to go from here

• Credentials that are already out are a different problem with a different fix — the free protections against identity theft covers what to do about it, starting with a credit freeze.
• Account security and network privacy are separate problems that get conflated constantly; what a VPN actually does draws the line.
• If a camera or smart-home account is among the ones you are securing, how to choose a home security system covers why that particular account matters more than most.
• To check your own exposure right now, Have I Been Pwned is free and takes thirty seconds.

Our full terms are on our disclaimer page.

FAQ

(Only questions the body doesn't fully answer.)

• What if I forget my master password? With a genuine zero-knowledge manager, the company cannot reset it — that's the point. Your only path back is the recovery method you set up in advance (a recovery key, emergency contact, or biometric unlock). Configure it on day one.
• Can the company read my passwords? No, if it's truly zero-knowledge — your vault is encrypted with a key derived from your master password on your device, and the server only ever holds ciphertext. This is the property to verify (via their audit), not just take on faith.
• Is a free password manager safe? A reputable one, yes — the encryption and unique-password generation that provide the security are in the free tier. "Free" is a risk mainly with unknown, unaudited apps, the same as with free VPNs.
• Do I still need one in the passkey era? For now, yes — passkeys are spreading but far from universal, so you'll manage both passkeys and passwords for years, and a manager handles both.
• Should I keep my 2FA codes in the same vault as my passwords? It is a genuine trade-off rather than a settled question. Keeping them together means a single compromised vault yields both factors; keeping them apart means real friction, and friction is why people abandon 2FA. For most people, TOTP codes in the vault plus a separate second factor on the vault itself is a reasonable balance. For your highest-value accounts, keep the second factor somewhere else.
• How do I share a password with my partner? Use the manager's sharing feature, not a message. Messaged passwords persist in two message histories and any backup of them. Where the account is genuinely joint, a shared vault is cleaner than sharing individual items.
• My password manager got breached — what do I do? In order: change the master passphrase; confirm 2FA is on and rotate its recovery codes; read the provider's own disclosure for what was actually taken (ciphertext or metadata is very different from plaintext); then rotate the passwords for the six categories in the migration section above rather than all of them at once. Do not immediately export everything to a different manager in a panic — a rushed migration creates more exposure than it removes.
• Do I need to change all my passwords now? No. Change the six that matter tonight and let the rest happen as you log in. A plan you complete beats a plan you abandon.
• Is the manager built into my phone good enough? If your whole life is on one platform and you never need to share credentials across ecosystems, yes — it is far better than reuse. Cross-platform managers exist for the other case, which is most people eventually.