Steps Home

Personalized Step Goals: What a Real‑World App Experiment Found

A skeptic-friendly way to set a step target that helps—without turning walking into a daily referendum on your willpower

By Steps Team

Photorealistic lifestyle photo of an adult adjusting a walking goal on a smartphone while standing on a quiet sidewalk at sunrise, casual clothing, natural light, no logos or text

If you’ve ever had a step goal, you’ve probably had the same thought at least once:

“Who picked this number?”

That’s the quiet flaw in most step targets. They’re either too low to matter (so you forget about them) or too high to be sustainable (so you resent them). And when a goal feels arbitrary, it’s easy to interpret a miss as a personal failure instead of what it usually is: a mismatch between the target and your actual week.

A new real‑world field study in a Dutch mobile health app tested something refreshingly practical: two different ways to build a personalized step goal—letting users set their own targets versus having an algorithm set targets based on recent step history. Both strategies increased steps, but not for the same kinds of users. That nuance is the whole point. (Liu et al., 2026; JMIR mHealth and uHealth)

The study, in plain language

The paper: Effectiveness of Step Goal Personalization Strategies on Physical Activity in a Mobile Health App: A Field Study (JMIR mHealth and uHealth, 2026). (Liu et al., 2026)

Here’s what makes this study unusually useful for regular humans:

  • It used real app users (not a tiny lab sample).
  • It compared personalization against a low default goal (2,000 steps/day, five days/week)—not an unrealistic 10,000-step mandate.
  • It tested two personalization routes:

    • 1) Personalized-by-you: you choose your own goal. 2) Personalized-by-the-algorithm: the app proposes goals based on your previous 4 weeks (a moving percentile method).

Design details (because skepticism is healthy): 5,800 users were split into two groups. One group got an email asking whether they were satisfied with their goal; people who were dissatisfied could choose one of the two personalization options. Another group received no email and stayed on the default. Because users self-selected into “personalize” or “don’t,” the authors used statistical methods (propensity score matching + difference‑in‑difference) to try to make comparisons fairer—but it’s still not the same as a fully randomized trial. (Liu et al., 2026)

What they found (numbers worth remembering)

Compared with users who kept the default goal, weekly steps increased by:

  • +3,793 steps/week for people who set their own goals (“personalized-by-you”).
  • +4,315 steps/week for people who used the algorithm (“personalized-by-the-algorithm”).

Those two averages are in the same ballpark.

But here’s the part that matters in real life: baseline activity level changed which strategy worked.

  • For low-active users, the algorithm strategy increased weekly steps (about +5,095 steps/week), while the self-set strategy did not show a meaningful change.
  • For high-active users, the self-set strategy increased weekly steps (about +4,266 steps/week), while the algorithm strategy did not show a meaningful change.
  • For medium-active users, both strategies helped.

In other words:

> The “best” personalization method depends on where you’re starting.

This is a rare moment where the research aligns with common sense: when you’re low-active, deciding a goal from scratch can feel like staring into a blank spreadsheet. When you’re already active, an algorithm can feel like it’s nagging you to do more of what you’re already doing—or it may not be ambitious in the way you find motivating.

What this does not prove (and why you should care)

A step goal is a tool. A tool can be useful without being magic.

This study suggests step-goal personalization can increase steps in the short term. It does not prove:

1) Long-term adherence. The authors themselves note the need for longer follow-up. (Liu et al., 2026) 2) Health outcomes. More steps is generally associated with better outcomes in many observational studies, but this paper focused on step count—not blood pressure, glucose, or longevity. 3) Causality as cleanly as an RCT. People opted into personalization, which can reflect motivation or readiness.

Still, as behavior-change evidence goes, this is the right kind of practical: it tests an intervention that you can actually replicate tomorrow.

The underrated insight: your goal should match your decision fatigue

Most advice about step goals fixates on the number. This study points to a quieter variable: who does the thinking.

  • If you’re low-active, the “thinking” part can be the hardest part—choosing a goal can trigger perfectionism (“What’s the right number?”) or avoidance (“I’ll set it later”).
  • If you’re already active, the “thinking” part can be motivating—self-setting turns into a commitment device (“I chose this; I’m the kind of person who does it”).

So the personalization strategy isn’t just about math. It’s about cognitive load.

A practical rule of thumb

Use this as a starting heuristic—not a moral judgment:

  • If you’re currently low-active: let an “outside brain” (an app, a coach, a simple formula) propose the next step. You can edit later.
  • If you’re already medium/high-active: set the goal yourself, but make it about consistency (days/week) rather than hero days.

How to set a personalized step goal (without making it weird)

Here are three options, in increasing order of involvement.

Option 1: The “algorithm in your head” (2 minutes)

If you don’t want an app to decide, you can approximate what many algorithms do: use your recent history.

1) Look at the last 7 days of steps. 2) Find your median day (the middle value). 3) Add a small bump:

  • Low-active: +300 to +700 steps/day
  • Medium-active: +500 to +1,000 steps/day
  • High-active: +0 to +800 steps/day (often the best lever here is where you place steps, not the total)

Why median? It ignores your outlier “amazing day” and your “flu day.” It’s a more honest picture.

Option 2: The “two-goal system” (recommended)

One reason step goals backfire is that they’re binary: hit/miss.

Try this instead:

  • Floor goal (non-negotiable): the minimum you’ll do even on a chaotic day.
  • Stretch goal (optional): what you do when the day cooperates.

Example:

  • Floor: 3,000 steps
  • Stretch: 6,000 steps

This approach reduces shame spirals while keeping the goal meaningful.

Option 3: The “days per week” lens

If you’re already active, make the goal about repeatability:

  • “I will hit my stretch goal 4 days this week.”

That’s a behavior target, not just a number. It’s also more robust to travel, weather, or workload.

Do this today (10–20 minutes): a step-goal reset that actually sticks

This is a mini-plan designed for busy, skeptical people. Total time: 10–20 minutes.

Step 1 (2 minutes): pick your mode

Choose one:

  • Mode A — low-active / decision-fatigued: use the “algorithm in your head.”
  • Mode B — medium/high-active: set a two-goal system (floor + stretch).

Write it down somewhere you’ll see it today (notes app is fine).

Step 2 (8–15 minutes): earn one clean win

Do a single, simple walk that matches your mode:

  • Mode A: 8–12 minutes at any comfortable pace. The goal is to start, not to perform.
  • Mode B: 10–15 minutes with a “brisk middle.”
  • 3 minutes easy
  • 4–7 minutes brisk (you can talk, but not sing)
  • 3 minutes easy

You’re not trying to set a record. You’re building evidence that the goal is compatible with your day.

Step 3 (1 minute): set tomorrow’s trigger

Pick one trigger you already do:

  • after your first coffee
  • after lunch
  • right before your evening shower

Decide: “When X happens, I walk for 8–12 minutes.”

That’s it.

A note on motivation: the goal is not the point

A personalized step goal is a way to reduce friction.

The point is not to win at step counting. The point is to make movement more likely to happen on the days when you’d otherwise default to sitting.

If you adopt one thing from this study, let it be this:

> If a goal isn’t working, change the goal—not your opinion of yourself.

Sources

Back to all posts