Attention Span Is a Training Effect
You don't have a focus problem. You have a training environment problem.
"I have the attention span of a goldfish" gets said like a diagnosis — a fixed trait, unlucky draw, nothing to be done. The research doesn't support the fatalism. Measured attention performance moves with conditions and practice, in both directions, on a timescale of weeks. Which means the honest way to think about focus is the way you'd think about anything trainable: current capacity, training conditions, progressive load, recovery.
By that frame, the average phone user is running an anti-training program of unusual quality.
What the anti-training program looks like
Training sustained attention would mean repeated, progressively longer intervals of single-tasking, with feedback. The feed-era default is the exact inverse: attention interrupted every few minutes (each check leaving residue on the task), a stimulation baseline ratcheted upward until ordinary work registers as under-stimulating, and recovery sabotaged at night by the device that caused the fatigue. Every rep shortens. It would be strange if capacity didn't fall — you built the program for it.
The encouraging part is the same arrow pointing the other way. In a randomized trial, two weeks of mindfulness training — structured practice at noticing and returning attention — improved working memory and GRE reading-comprehension scores while reducing mind-wandering (Mrazek et al., 2013). Two weeks. The system adapts fast when the reps change. Attention behaves like a training effect because, functionally, it is one.
The training program, itemized
Here's what a deliberate focus program needs, and what each piece is doing in HabitApp.
Protected reps. The unit of training is the uninterrupted block. The shield makes the block enforceable at its weakest point — the automatic reach — by turning the mid-block pickup into a priced, deliberate act. A rep you can't accidentally end is a rep that counts.
A log. No training program survives without one. The check-in is the log: did the block happen, yes or no, recorded. Recorded self-monitoring reliably improves the monitored behavior on its own (d ≈ 0.40) — the logbook is part of the stimulus, same as in the gym. And the pickups chart is the performance metric: watch reaches-per-day fall the way you'd watch a resting heart rate.
A coach that reads the log. Every two weeks, calibration reviews the actual record — which goals held, which collapsed — and proposes adjustments you approve or decline. That's progressive programming: load that adapts to demonstrated capacity instead of week-one optimism. A plan that never updates isn't a program, it's a poster.
Recovery. Attention trains against the night before. The optional pre-sleep window guards the recovery half of the program, because the strongest finding in the screen-time literature is that the pre-sleep hour, through sleep, is where the damage compounds.
The deload at the end. The reward tapers across the 90 days, on schedule, like a patch stepping down its dose. This is the part that makes it training rather than dependency: the point of scaffolding is the day it comes off. At day 90 the program ends and offers you the exit, with the blocks expected to stand on their own.
The honest paragraph
Required disclosure, as always on this blog: no randomized trial has been run on this app, and this article will not pretend otherwise. What's been tested are the components — the pause at the pickup (57% fewer openings in six weeks), recorded self-monitoring, implementation intentions, sleep protection, and the trainability of attention itself in controlled settings. The program assembles tested parts; the assembly is the bet you're evaluating.
But notice what the fatalist version asks you to believe instead: that attention is the one human capacity that only degrades, that responds to years of anti-training and would respond to nothing else. That's not a finding anywhere in the literature. That's the feed's marketing.
You trained it down. Same equipment, other direction.
Sources: Mrazek, M. D., et al. (2013), Psychological Science — two-week mindfulness training, working memory and GRE gains. Leroy (2009), attention residue. Self-monitoring reactivity, d ≈ 0.40. Gollwitzer & Sheeran (2006). Six-week RCT on pre-app pauses. Sleep and sustained attention: sleep-deprivation literature; see "The Hour Before Bed Is the Whole Ballgame."