First: who is behind it, and can they keep it alive?
Every app is a bet on the team that maintains it, because apps are living products, not purchases. The cheapest check is the developer listing on the store page: is it a known company, a named studio with other apps, or a string of letters you have never seen? Neither is automatically bad — solo developers make wonderful apps — but unknown developers deserve more of the other checks, because there is less reputation at stake if they behave badly.
Then look for signs of life. Recent update date, ideally within the last six months. A changelog that says something real ("fixed sync conflicts, faster search") rather than nothing ("bug fixes and improvements" forever can mean a skeleton crew, or it can mean a stable app — read several versions back for the pattern). An active response from the developer to store reviews, especially critical ones, is one of the strongest signals available: it shows someone is home.
The failure mode you are checking for is abandonment. Apps do not die loudly; they die quietly — updates slow, bugs pile, and one day a change in your phone's system breaks something and nobody is coming to fix it. The cost of abandonment is not the download; it is whatever you stored inside the app.
Second: what it costs later
The store page tells you the app is free. What matters is what the app does after the honeymoon, and app pricing has a small set of standard patterns worth recognizing. The free tier that is a demo: core features locked, trial of everything for a week, then a wall. The subscription creep: free at launch to grow, features migrating behind a paywall over the following year — a pattern you can spot by reading reviews from a year ago and comparing what people complain about now. The bait pricing: a one-time purchase that turns out to be one feature, with the others as separate purchases. And the data price: "free" apps whose real currency is attention, advertising tracking, or your data.
None of these patterns is automatically disqualifying — developers need to eat, and subscriptions fund genuinely good software. The vetting question is whether the pricing model is honest and visible now, so you can decide with your eyes open. Read the in-app purchase list on the store page before downloading, and read recent reviews specifically for pricing complaints. If people say "it changed," believe them about the direction it is moving.
Third: what the permissions say
Permissions are the fastest honest signal in the whole vetting process, because they are stated by the developer and hard to fake. A flashlight app that wants your contacts, a photo editor that wants your microphone, a notes app that wants your location in the background — each of these is a mismatch, and mismatches are how you filter. The rule is not "never grant permissions"; it is "each permission must have an obvious job." Contacts for a messaging app: obvious job. Background location for a game: no.
Two permissions deserve extra care: accessibility (it can read your screen — appropriate for a screen-reader or password manager, alarming for almost anything else) and notifications (annoyance is the floor; re-engagement pings are the business model of many free apps). Deny generously at install. A well-built app asks again, politely and with a reason, when it truly needs something — and an app that cannot function without every permission on day one is telling you its priorities.
For anything that handles sensitive material — notes, photos, health, finances — the check is the privacy policy plus the data-safety section on the store listing. Read what they claim to collect and whether they say they sell it. Vague language ("we may share with trusted partners") is itself information.
Fourth: read the reviews like a skeptic
App store reviews follow the same laws as product reviews, with two twists. Twist one: ratings are polluted by feature complaints ("lost a star because no dark mode"), so read the text, not the stars. Twist two: app reviews are more likely to be recent, which is an advantage — sort by newest and read the last three months, because an app's current state is what matters, and its history tells you the trajectory. Rising complaints about the same issue mean a decline; the same praise repeated for years means stability.
Search the reviews for the words that matter: "subscription," "data," "deleted," "battery," "crash." Those five searches surface the real issues faster than any sorting. And check the developer's replies: a developer who answers "thanks, we're aware, here's the timeline" is a company; a developer who argues with reviewers is a forecast.
Outside the store, thirty seconds of community search — the app name plus "alternative" or "disappointed" — tells you what the store page will not.
Fifth: the two-week trial, done deliberately
The final vetting step happens on your phone, and it is a matter of setting up the trial honestly. Give the app one deliberate session: create the real account (or decide not to), set up the real data, and use it for its actual job once. Then wait. The apps worth keeping announce themselves by being opened without promotion; the rest sit in a folder you never open. After two weeks, keep what earned a habit and delete the rest — including its account, its data export if you want it, and its permissions.
Delete thoroughly, because deletion is part of vetting: remove the app, then revoke its access in your account settings (connected accounts, shared permissions), and if you gave it anything sensitive, check its deletion options. Five minutes now prevents the slow accumulation of half-alive accounts that leak attention for years.
When to break these rules
The checklist has exceptions, and they are worth stating so the rules stay sharp. Try apps early when the category matters to you and the risk is contained: being an early user of something in your professional field is how you learn the category. Accept messier apps from developers you already trust — reputation compounds, and a known-good team deserves benefit of the doubt on a new release. And skip vetting entirely for apps you will use once (a parking app, a one-time scanner): the five-minute audit costs more than the app's whole lifecycle. Match the depth of vetting to the stakes: daily-use tools and anything holding your data get the full check; the rest get a glance.
The short version
Five checks, five minutes: who is behind it and are they alive; what it costs after the honeymoon; whether each permission has an obvious job; what the newest reviews and five keyword searches say; then a deliberate two-week trial with a thorough delete at the end. Most app mistakes are not bad luck — they are unvetted downloads, and the vetting is short enough to actually do. Your phone, your data, and your next three years of renewals will all be cleaner for it.