oversharingmuch?

How we score

Every score on this site is calculated, never typed in by hand. We collect documented facts about a company, such as a regulator's fine, an encryption feature, or a court order, and each fact moves one category of the score up or down. The numbers below come straight from the code that calculates the scores, so this page can't drift out of date.

What counts as evidence

  • One claim, one fact. Each piece of evidence is a single factual sentence with a date.
  • Always sourced. Every claim links to at least one source. Primary sources (regulator and court decisions, the company's own policies, transparency reports and blog posts) count fully. Reputable journalism counts slightly less.
  • Archived. Where possible we link a Wayback Machine snapshot, because privacy policies change quietly.
  • Fixed issues still count, for less. A breach that was fixed still happened, but it matters less than one that is still going on.

What we don't do

  • We don't accept money, ads or affiliate links from the companies we rate.
  • We don't usually score products separately. Trust is a property of the company: the same data pool, the same privacy policy and the same legal exposure. Apps are listed under their company, with a note when one is clearly better or worse. The exception is a product that competes in a category where the rest of the company's evidence doesn't apply, such as Proton VPN among VPNs. It gets its own score there, built from company-wide evidence plus evidence about that product, so encrypted email doesn't lift a VPN's rating.
  • We don't guess. If we have no evidence for a category, it is left out rather than assumed to be good or bad.

The formula

  1. Every category starts at 50, a neutral score.
  2. Each piece of evidence adds or subtracts points based on its severity, multiplied by the weight of its best source.
  3. Issues the company has fixed still count, at 50% of their points.
  4. If a company can't read your data, there is less to leak or hand over. Negative evidence about government requests, jurisdiction and track record is reduced by up to 50%, in proportion to its "Can they read your data?" score.
  5. The points are mapped onto 0–100 with a curve: every fact counts, but each extra one moves the score a little less, so a category never hits a hard floor or ceiling.
  6. The overall score is the weighted average of the categories that have evidence.
  7. A company needs evidence in at least 6 of the 9 categories before we publish a score.

Categories and weights

CategoryWeight
Can they read your data?
Whether your content is end-to-end or zero-access encrypted, so the company itself can't read it.
20%
Business model
How the company makes money. Advertising and data-sharing models reward collecting more about you.
15%
Track record
Breaches, scandals and regulator fines, and how the company responded.
13%
Government & legal requests
What the company hands over to police, courts and governments, and whether it pushes back or reports on it.
12%
Data collection
How much the company collects, including tracking outside its own products.
10%
AI & your conversations
Whether your chats, posts and files are used to train AI, reviewed by humans, or kept after you delete them.
10%
Your control
Can you see, export and delete your data, and opt out without dark patterns?
8%
Transparency
Open-source code, independent audits and transparency reports.
6%
Jurisdiction
Which country's laws apply, and how much those laws protect you from surveillance.
6%

Severity

SeverityPoints
Minor±8
Moderate±15
Major±30
Severe±50

Source tier

Source tierMultiplier
Primary×1
Secondary×0.8
Other×0.5

Severity guide

How we decide how much a fact counts. The same rule applies whether the fact is good or bad news, so the same kind of fact always carries the same weight.

Can they read your data?
Minor ±8
Covers metadata or a narrow feature.
Moderate ±15
Covers one product or feature, or protection that is opt-in or partial.
Major ±30
Covers a major product or most of your content, such as messages, backups or files.
Severe ±50
Covers the core service and nearly all your content, by default: either end-to-end/zero-access encrypted, or readable by the company.
Business model
Minor ±8
Small or secondary ad products, ads that don't use your personal data, or a limited commitment.
Moderate ±15
Some ads or sponsored content; or structural safeguards such as non-profit ownership or a public no-ads pledge.
Major ±30
A large advertising business alongside other income; or, positively, funded by users/hardware with no advertising.
Severe ±50
Most revenue comes from targeted advertising or monetising user data.
Track record
Minor ±8
A small incident that was quickly fixed, or one at a supplier with limited data.
Moderate ±15
A smaller fine or settlement, a limited exposure, or a product change that exposed people's data.
Major ±30
A fine or settlement of €/$100 million or more, data of 1 million+ people exposed, staff abusing access to user data, or a regulator banning the service.
Severe ±50
A fine or settlement of €/$1 billion or more, data of 100 million+ people exposed, or deliberate misuse of personal data against people.
Government & legal requests
Minor ±8
A single handover of a small piece of metadata.
Moderate ±15
A documented handover of limited data, dropping protection under government pressure, or publicly resisting a demand in court.
Major ±30
It can (or can't) hand over the contents of your data: messages, files, backups, encryption keys, or bulk chat logs.
Severe ±50
The company can hand over essentially nothing, even when forced; or it gives authorities broad or real-time access by design.
Data collection
Minor ±8
A small extra piece of data, such as a phone number.
Moderate ±15
Collects your content plus the usual account and device data, or privacy-preserving design for a specific feature.
Major ±30
Broad or sensitive collection (location, voice, biometrics, keystrokes, browsing), or collection proven by independent audit to be minimal.
Severe ±50
Tracks you across other websites, apps or the physical world at massive scale.
AI & your conversations
Minor ±8
A small AI data practice.
Moderate ±15
Training on public posts, an AI feature that exposed data, or an opt-out, temporary chat or no-training plan.
Major ±30
Your conversations are used for training by default, reviewed by humans or kept for years; or AI designed so the company can't keep your data.
Severe ±50
Your conversations are used for ads or training with no way to opt out.
Your control
Minor ±8
Basic download or delete options.
Moderate ±15
Good tools to export and delete your data, or to use the service without giving up personal details.
Major ±30
A regulator found you were forced to consent, or a system-wide protection that stops tracking by default.
Severe ±50
You can't see, export or delete your data at all.
Transparency
Minor ±8
Publishes a periodic transparency report.
Moderate ±15
Publishes every government request and its response.
Major ±30
Apps are open source or independently audited.
Severe ±50
Everything is open source and independently audited, including servers.
Jurisdiction
Minor ±8
A proposed legal change that could weaken or strengthen protection.
Moderate ±15
US law (such as the CLOUD Act) can compel access, or strong privacy law like the EU's GDPR or Swiss law protects you.
Major ±30
Data is regularly accessible to staff in such a state.
Severe ±50
Data is held in, or the company answers to, a state where companies must assist intelligence services and courts aren't independent.

Tiers

  • 80–100Tight-lippedCan't read your data even if it wanted to.
  • 60–79DiscreetCollects some data, but mostly keeps it to itself.
  • 40–59ChattyShares more about you than you'd probably like.
  • 25–39Loose-lippedYour data is part of how it makes money.
  • 0–24OversharerTreats your data as widely available.