How to Set Fitness Goals Without Defaulting to SMART

On January 8, 2026, Exercise and Sports Science Australia published an expert statement addressing the use of SMART goals in exercise. For three decades, the answer to how to set a fitness goal has been the same five letters. Specific, measurable, achievable, relevant, time-bound. The researchers behind the statement, writing in Sports Medicine, catalogued problems with treating SMART as a universal recipe, including goals that ignore a person's actual skill level and a framework with nothing to say about what happens the week a goal gets missed.¹²

Whether you're just starting or already mid-program changes which goal type fits the evidence. This piece also covers what to do the day you break a streak, when moving your goal weight is a revision rather than quitting, and what to set the morning after you hit the number.

01How to set a fitness goal that matches your activity level

A specific, difficult target motivates someone who already has a routine and starts to backfire on someone who doesn't. That single mismatch explains a lot of abandoned fitness apps. The ESSA statement's clearest practical recommendation is to key goal type to current activity level rather than handing everyone the same SMART template.¹

A pilot randomized controlled trial from the same research group tested this directly in walking behavior. In a six-week program with 30 adults, open and specific goals produced similar increases in activity, while the open-goal group reported more positive experiences.³ This result supports matching the goal format to the person's starting context; it does not show that open goals outperform specific goals on distance in every population.

The timing of the goal matters as much as its shape. A once-a-week target such as "exercise three times this week" leaves six days with no signal telling you whether you're on pace, so the miss only shows up after it's too late to correct. The ESSA statement pulls apart the trials that tested this directly. Interventions using a daily goal, and interventions pairing a daily goal with a weekly one, raised physical activity. Interventions using a weekly goal by itself did not move the needle.¹ Set the goal at the grain of a day. Review it on a fixed weekly date.

02Where the write down your goals statistic comes from

The claim says that people who write down their goals are 42 percent more likely to achieve them. The underlying Matthews summary reports self-rated scores under several writing and accountability conditions, not a 42 percent result.⁵

The claim traces to a research summary by psychologist Gail Matthews of Dominican University of California, distributed as a PDF rather than published in a peer-reviewed journal. The summary also discusses a separate 1953 Yale story about written goals and earnings; that story is not the source of the study's reported scores.⁵

Matthews recruited 267 professionals from businesses and networking groups and retained 149 through four weeks of tracking a workplace goal, not a fitness goal. Participants rated their own progress on a 1-to-10 achievement scale under one of five conditions, ranging from simply thinking about the goal to writing it down, forming action commitments, and sending a friend weekly progress reports. The unwritten-goal group averaged 4.28. The group with the full stack, written goal, action commitments, and weekly reports to a friend, averaged 7.6, the highest of the five conditions and significantly ahead of every other group.⁵ That's a real and sizeable gap on a self-rated scale. It is not a 42 percent figure, and the summary itself was never submitted for peer review.

A number that has been through peer review exists, and it's smaller and more honest than the myth. Epton and colleagues pooled 141 papers, 384 effect sizes, and 16,523 participants and found that setting a goal produces a real but moderate effect on behavior change, d = 0.34.⁴ Use that number instead. And if you're choosing which version of goal-writing to bother with, Matthews' own data, whatever its methodological limits, still points the same direction as the peer-reviewed literature. The sharpest jump came from accountability to another person.

03What to do the day you break a fitness streak

Take a lifter with a 78-day ring-closing streak heading into a planned deload week. The reflexive move is to keep closing rings through the fatigue the deload exists to resolve, which defeats the point of programming rest at all. The better move is to name the break before it happens.

Silverman and Barasch tracked how consumers responded to streaks in progress. In their study, an intact streak predicted continued engagement at b = +0.25, while a broken streak predicted a drop at b = -1.01, roughly four times the size in the opposite direction.⁶ A condition that explicitly repaired the break, reframing it and resetting expectations, recovered engagement to 85.2 percent of baseline, against 68.7 percent for a break left unaddressed.⁶

That pairs with a separate finding about how habits actually form. Lally and colleagues tracked real people building daily habits over three months and found that missing a single opportunity to perform the behavior did not measurably disrupt how automatic it became.⁷ An isolated miss did not measurably change habit automaticity in that study.

Design the miss in advance rather than discovering it after the fact. Log a planned deload as planned rest on the calendar before the week starts, and switch the unit you're tracking from consecutive days to weeks that clear a minimum threshold, four or more training sessions, say, rather than seven unbroken days. That is a tracking protocol for preserving continuity; the streak study does not establish four sessions as a universal threshold.

04Why 10,000 steps a day was never a research target

The number traces to a 1960s Japanese pedometer marketing campaign built around a product name that translates roughly to "10,000-step meter," part of a walking boom that coincided with the Tokyo Olympics.¹¹ It was advertising copy, and it has outlived nearly every study that has actually measured where step-count benefits level off.

Ding and colleagues' 2025 dose-response meta-analysis found all-cause mortality risk at 7,000 steps a day running about half of what it is at 2,000 steps, HR 0.53 (95% CI 0.46 to 0.60), with most of the risk reduction already captured by that point.⁸ Beyond roughly 7,000 to 8,000 steps, additional walking keeps helping but the curve flattens noticeably.

The counterargument to just picking a smaller, more achievable round number is worth naming. Anson and Madras randomized participants to different assigned step goals and found that lowering the target from 10,000 to 5,000 actually reduced how much people walked, even though the group still cleared its new, lower number.¹² The goal itself shapes effort. Pick a number that matches your actual age band and current activity level from the table above, and let the six-decade-old marketing slogan retire for good.

05Who should set your fitness goal and who should hear it

Two well-designed studies point in different directions here, and naming the disagreement is more useful than smoothing it over.

Patel and colleagues randomized 500 adults across a 2x2 design, self-chosen versus assigned step goals, and immediate versus gradual increases toward that goal. Only one of the four combinations worked. Self-chosen and immediate. That group sustained an increase of 1,384 steps per day, holding at +1,391 through an 8-week follow-up period after the intervention ended. The other three combinations, including assigned-and-immediate and self-chosen-but-gradual, produced no meaningful change.¹³

Epton's meta-analysis of the wider goal-setting literature found the opposite emphasis. Across 384 effect sizes, the goals that performed best were difficult, public, and set for a group, the reverse profile from the easier, private, individually chosen goals that ENGAGE's winning arm relied on.⁴

These don't fully reconcile, and treating them as one finding would misrepresent both. The difference sits in what each study actually varied. ENGAGE isolated who picks the number and how fast you ramp toward it, inside a single step-count context. Epton's moderator analysis pooled goal difficulty, publicness, and group structure across a much wider range of behaviors and settings, without isolating who chose the number in the first place. The practical rule that survives both is a split one. When you're deciding whether to set your own number today or wait for someone to assign one, choose it yourself and start now instead of easing in. When you're deciding whether to tell anyone about it, tell a group rather than no one, and pick a harder number than feels comfortable, provided the choice of that number is still yours.

06When changing your goal weight is a revision not a quit

Foster and colleagues asked dieters starting a weight-loss program what weight they were hoping for. The average goal represented a 32 percent reduction in body weight. After losing an average of 16 kg, 47 percent of participants hadn't even reached the weight they'd described beforehand as merely "disappointing."¹⁴ Read alone, that looks like a clear warning against setting an ambitious number.

Linde's two larger cohorts read the opposite way. In a 302-person cohort, unrealistic goal weights didn't undermine how much weight people actually lost, with higher dream-weight losses tracking with greater loss at 18 months.¹⁵ In a separate 1,801-person cohort, women with less realistic goals lost more weight at 24 months than women with more modest ones.¹⁶ Both findings come from their own samples, and they don't resolve into one tidy answer about whether ambition helps or hurts.

The resolution that survives both studies is to separate the number from the plan. Keep the ambitious goal weight as direction. It did not predict worse outcomes in these cohorts, and for some people it may sustain effort longer than a conservative one would. Put the actual specificity into the weekly review instead of the target itself. Painter and colleagues found that dieters who weighed in three or more days a week and logged food three or more days a week, in at least 64 percent of weeks, lost significantly more weight than those who didn't hit that floor.¹⁷ That threshold comes from Painter's study population. Self-monitoring carries real weight in the broader literature too. Michie's meta-regression across 44,747 participants found self-monitoring combined with one other control-theory technique produced an effect of 0.42 versus 0.26 without it, and Harkin's meta-analysis of 138 studies and 19,951 participants put the monitoring effect at d+ = 0.40, though a later meta-review cautions that not every self-monitoring technique carries equal weight across contexts.¹⁹²⁰²¹

In the IDEA trial, dieters given a wearable activity tracker on top of a behavioral weight-loss program lost less at 24 months, 3.5 kg, than dieters given the identical program with no wearable, 5.9 kg.¹⁸ That result belongs to that intervention and does not establish that every wearable reduces weight loss. The review cadence is the behavior measured by the self-monitoring findings above; the gadget is optional.

Revising a goal weight mid-program is compatible with a plan that keeps its weekly review running. Keep the weigh-ins and food logs tied to that review.

07What goal to set the morning after you hit your target

A reader who steps on the scale and sees the number they picked eight months ago has two bad defaults waiting. One is coasting, since nothing on the tracker tells them what to do once the target is reached. The other is picking a new number out of habit before deciding whether weight is even the right thing to keep optimizing.

The goal type after reaching a scale target depends on current activity level, not the number alone. If the reader is still not training consistently, an open or learning goal remains the fit. If a routine is established, a specific, difficult performance goal may fit better.¹ Set the next goal the same day the current one is reached, before the motivation that got you here has a chance to fade. Choose it yourself and make it immediate rather than gradual, the combination that moved behavior in ENGAGE.¹³ Use a specific and difficult target only when the current routine supports it; otherwise keep the goal open or exploratory.¹³ Build the streak-repair rule in from day one, too, naming what a planned miss will look like before the first one happens instead of improvising a response after a bad week turns into a bad month.⁶

The goal that gets someone to a number and the goal that keeps them past it are different goals. Set the second one before the first one stops mattering.