What the data actually shows
Implementation intentions are among the better-supported findings in behaviour-change research. Gollwitzer's work, and a large body of studies and meta-analyses building on it, find that specifying the when, where, and how of an action substantially raises the odds of actually doing it compared with holding only a goal. The mechanism is that linking a concrete cue to a concrete response automates the start, so you don't have to re-decide in the moment.
The planning fallacy is equally well-established. Kahneman and Tversky described our systematic tendency to underestimate task duration even when we know similar tasks have run long before, and Buehler and colleagues documented it experimentally — people predict optimistic completion times and then routinely overshoot. Planning that ignores this tends to produce schedules that fall apart by mid-morning, while planning that pads estimates and looks at how long similar tasks really took holds up better.
But the evidence also cautions against rigidity. Highly detailed, minute-by-minute plans are brittle: a single disruption cascades through the whole schedule, and the research on goal pursuit suggests overly rigid plans can reduce follow-through and increase frustration when reality deviates. Flexible plans that fix intentions but leave room to adapt tend to outperform both no plan and tightly scripted ones.
Why this feels different from how it actually is
Planning often feels useless because people compare a rigid schedule against the messy reality of a real day and conclude that planning 'never works.' What actually failed is usually the brittle, optimistic plan, not planning as such. The version the evidence supports — concrete intentions plus honest time estimates plus flexibility — is rarely the version that disappointed them.
It also feels pointless because the planning fallacy is invisible from the inside. Each time a task runs over, it feels like a one-off interruption rather than a systematic pattern, so we keep making the same optimistic estimates and keep being surprised. The shortfall feels like bad luck, not predictable bias.
And there is the appeal of the perfect plan. A detailed, color-coded schedule feels productive to build and gives a satisfying sense of control, which is exactly why over-planning is tempting and why its collapse feels so demoralising. The feeling that you 'should' be able to run the day to the minute sets up the disappointment when you can't.
What the research says to do about it
Make plans specific in the way that matters: when, where, and how. Converting 'I'll exercise' into 'after I drop the kids at school, I'll walk for twenty minutes' is the core of an implementation intention and the single best-supported move for turning intention into action. Tie the action to a concrete cue you'll actually encounter.
Build in realistic time. Because we reliably underestimate duration, the evidence-based correction is to base estimates on how long similar tasks actually took before, rather than on how long you hope this one will take, and to pad accordingly. A plan that expects the day to take longer than the optimistic version is one that survives contact with reality.
Keep the plan rough and flexible. Fix the few intentions that matter most, leave slack for the unexpected, and treat the plan as a guide rather than a contract. The research pattern favours light, adaptable structure over either no structure or rigid scheduling — enough plan to direct the day, not so much that one interruption breaks it.
What the research says does not help
Vague goals without a concrete trigger do little on their own. 'I'll be more productive' or 'I'll read more' lacks the when-where-how that makes implementation intentions work, so it tends to dissolve the moment a competing option appears. Intent without a plan is the condition the research repeatedly finds to be weak.
Minute-by-minute scheduling is fragile and often counterproductive. Packing a day to the minute, with no slack, all but guarantees the plan breaks at the first disruption, and the cascade of falling behind can be more demoralising than not planning at all. Tight control is not the same as effective planning.
Optimistic estimating sabotages otherwise good plans. Assuming everything will go quickly — the planning fallacy in action — produces schedules that are wrong by design. Building a careful plan on hopeful time estimates rather than on how long tasks really take is one of the most common ways planning quietly fails.
Real numbers in context
The findings here are about direction and reliability rather than a single headline statistic. Across the implementation-intentions literature, specifying when, where, and how an action will happen consistently raises follow-through relative to holding a goal alone — a robust, repeatedly replicated effect rather than a one-off result. The practical takeaway is the format ('if cue, then action'), not a precise percentage that would vary by behaviour and study.
On the planning fallacy, the durable point is the bias itself: people systematically underestimate task duration, even for tasks they have done before, which is why padding estimates and anchoring to past actual times is the recommended correction. Taken together, the evidence supports planning — provided it is specific, realistic about time, and flexible — over both winging it and over-scheduling.