the research, read honestly

Do app blockers actually work?

21 April 2026 · 4 min read · by the makers of finit, which is not a blocker, so read on with that in mind

Mostly yes, while they're running. The best studies we have point the same way: adding friction between you and an app cuts how often you open it, sometimes dramatically. They also agree on the uncomfortable part, which the marketing tends to skip: the effect lives exactly as long as the intervention does, and it fades when the tool comes off.

Here is the evidence, caveats attached, including the ones inconvenient for us.

What the field data shows#

The most-cited result in this niche belongs to one sec, the app that makes you take a breath before Instagram opens. A study published in PNAS in 2023 followed roughly 280 users over six weeks and found they attempted to open their targeted apps 57% less often by the end, and abandoned about 36% of the openings mid-pause. The phone was already in hand, and they closed it anyway. Two things belong next to those numbers: it was a field study without a control group, not a randomized trial, and one of the authors is one sec's developer. To their credit, both facts are printed in the paper rather than hidden.

Their preregistered follow-up with around 500 participants is, for our money, the more interesting result. It dismantled the intervention into parts and asked which part does the work. Answer: the friction itself. The forced delay, and how easy the escape hatch is. A pure awareness message, the gentle "are you sure?" without any real delay, did roughly nothing in the moment.

What the randomized trials add#

A 2026 randomized trial in JMIR mHealth tested a blocking app (Wellspent) on 70 iPhone users for three weeks. Time on each person's most problematic app dropped by about 29 minutes a day versus control, a real effect. But how problematic the usage felt, and people's sense of control over it, didn't measurably change. The minutes moved; the relationship to the phone didn't.

The most instructive trial didn't use software at all. Researchers at the University for Continuing Education Krems randomized 111 students into a three-week commitment: keep smartphone use under two hours a day, checked with weekly self-monitoring. It worked: use fell by roughly 2.5 hours a day, and at the end of the trial the group scored better on depressive symptoms, sleep, stress, and well-being. Then everyone went back to their lives, and six weeks later only the sleep improvement was still statistically significant. The rest had washed out.

That washout is the real headline of this literature. Not "blockers don't work": they do, measurably, while active. The finding is that nothing sticks by itself.

Why the effect doesn't hold#

A blocker is a splint. Splints are useful, and nobody expects the muscle to be stronger the day the splint comes off. As we put it elsewhere on this site: the morning the lock comes off, you're the same person, with the same habits and no record of either. The dismantling study adds the design corollary: the strongest lever in any blocker is how costly the escape hatch is, and every blocker must have one, because it's your phone. Tools that make ignoring easy get ignored; tools that make it hard get deleted in a weak moment. Most people quietly cycle through both endings.

When a blocker is the right call#

Plenty of times. The evidence supports using them the way they work:

• For seasons, not identities. Exam month, a deadline sprint, the first weeks of breaking an autopilot loop. Defined windows, then reassess.
• Friction before force. A delay you must sit through changes behavior in the moment; a hard lock mostly changes when you uninstall. Start gentle, escalate only if needed.
• For sleep. The one effect that survived in the Krems trial. If a blocker guards nothing else, let it guard your nights.

If you want specific picks, we wrote a guide to the best app blockers for iPhone; we don't sell one, and nobody paid to be in it.

What no blocker can give you#

A memory. Every blocker experiment ends: the trial expires, the exam passes, the app annoys you off. When it does, you're left with the question the splint can't answer: did anything change? Your phone won't tell you; Apple deletes the detail after about four weeks, so by the time you wonder whether March's experiment worked, March is gone.

That's the gap finit is built for. It never blocks anything. It keeps the record: every app, every day, for years, on your iPhone and in your own private iCloud. What a long record gives you is legibility: which blocker worked, what the job change did, which weeks you'd want back. Run whatever experiments you like. finit is how you'll know what they did.