The data you give a wellness app — your moods, your journal, your habits, what you're struggling with at 2 a.m. — is some of the most sensitive information that exists about you. And here's the uncomfortable part: in most places, it's barely protected. The health-privacy laws people assume apply (like HIPAA in the US) generally cover clinical providers, not consumer apps. The app holding your anxieties is often regulated more like a game than like a doctor.
The track record matches. When Mozilla's Privacy Not Included project reviewed mental-health apps in 2022, it called the category among the worst it had ever audited — most of the apps earned its warning label for things like vague policies, weak security answers, and broad data sharing. And in 2023, the U.S. Federal Trade Commission settled with BetterHelp, one of the biggest online-therapy platforms, over sharing users' health information — including answers to intake questions — with advertising platforms after promising it wouldn't. The order included $7.8 million for refunds and a ban on sharing health data for ads.
Full disclosure before we go further: we build a product in this category. Which is exactly why this post is deliberately product-agnostic — these are the seven questions we think you should ask of any app that wants your inner life. Ours included. Honest products can answer them crisply. And evasive answers are themselves an answer.
The seven questions
1. "Is my sensitive content encrypted at rest — and what counts as sensitive?"
Not "do you use encryption" — every app encrypts traffic in transit; that's table stakes. The question is whether your journal entries, chats, and answers sit encrypted in the database, so a leaked backup or a curious insider reads nothing. And ask what's covered: an app that encrypts your password but stores your trauma journal in plaintext has answered the question.
2. "Is my data used to train AI models — yours or anyone's?"
The defining question of this era. Three parts: does your content train the company's own models; does it go to third-party AI providers, and on what terms; and is "improving our services" in the policy defined narrowly, or is it a blank check? You're looking for a flat, unqualified no on training. Watch for qualified answers — "not without consent," where consent was a pre-checked box in onboarding.
3. "What leaves your servers when AI features run — and what gets stripped first?"
If the app has an AI coach or AI summaries, some of your content is traveling to a model somewhere. The question that separates products: are identifying details redacted before the prompt leaves, or does the raw entry ship wholesale? Few products can answer this concretely. The ones that can — in specifics, not vibes — are telling you something real about how they're built. (Our earlier AI post covers why this layer matters even when the policies look good.)
4. "Who do you share data with — and does 'anonymized' survive scrutiny?"
The BetterHelp case is the cautionary tale here: the sharing that triggered the FTC was with mainstream ad platforms, of exactly the data users assumed was private. Ask specifically about advertising and analytics SDKs embedded in the app — data flows through code, not just policies. And treat "anonymized" and "aggregated" as claims that need detail. Re-identifying thin datasets is, unfortunately, a solved problem.
5. "Can I export everything — and does deletion actually delete?"
Two halves of one right. Your reflections are yours, so there should be a real export (structured data, not a PDF of screenshots). And deletion should remove your content from production systems on a stated timeline — not an "account deactivation" that strands your journal on their servers forever. Here's a useful tell: an app that makes leaving easy is an app that plans to keep you by being good.
6. "What happens to my data if you're acquired or shut down?"
The policy that matters most isn't this company's. It's the next owner's. Wellness startups fail and get acquired constantly, and databases are assets in both events. Look for an explicit commitment about data on change of control — and notice whether the question has been thought about at all. As our journaling post put it: the real test is whether you'd be comfortable with what they have if the company changed hands tomorrow.
7. "Where's the honest sentence about your limits?"
The meta-question, and our favorite. Every real privacy posture has limits — redaction can miss things, employees exist, subpoenas exist — and a trustworthy product says so somewhere. Marketing that promises perfect privacy is either naïve or lying, and either one disqualifies. The strongest trust signal in this category isn't a bold claim. It's a precise one.
How to actually use these
Nobody interrogates a signup flow with a seven-point checklist. So here's the practical version: before you type anything into a wellness app that you'd mind a stranger reading, spend five minutes on questions 2, 4, and 5 — training, sharing, and exit. They're the highest-stakes three, they're usually answerable from the privacy policy plus one support email, and an app that fails any of them has told you how it thinks about the rest.
And calibrate to what you're sharing. A habit tracker holding your gym schedule deserves less scrutiny than a journal holding your marriage. The rule from our safety-tier post applies here too: unbounded uncertainty is a tax, and vague data practices are exactly that — a little background worry you carry every time you're honest with the app. Part of what these questions buy you is the ability to stop wondering.
Where NexTier fits, briefly
Our answers, for the record, live on our trust page: journals and coaching content encrypted at rest, no training on your data, pattern-based redaction plus a second sanitizing layer before anything reaches an AI provider, self-serve export, and deletion that removes your account and everything attached to it rather than deactivating it. Two honest limits while we're here: no redaction is perfect, and our export doesn't yet cover every feature area — we name the gaps on the trust page rather than calling it complete. We publish the answers because we think answering is this category's real test.
Whatever tool you choose — including a paper notebook, which passes all seven questions effortlessly — ask the questions.
—
This post is part of a series on personal development organized around what we call Maslow's extended hierarchy — our synthesis of his later work. Sources: Mozilla Foundation, "Privacy Not Included" review of mental-health apps (2022, updated 2023) — mozillafoundation.org; U.S. Federal Trade Commission, In the Matter of BetterHelp, Inc. (2023 settlement: $7.8M for consumer refunds; ban on sharing health data for advertising) — ftc.gov. This post is educational content, not legal advice.