All or nothing is the worst way to build a fasting habit

Most fasting apps delete your streak if you miss a day. Some delete it if you do not open the app, whether or not you fasted.

This is bad habit design dressed up as motivation.

Ember Fasting was built on the opposite premise: fast above half your target and the streak holds.

What the habit research actually says

The most cited study on habit formation followed 96 people forming a new daily habit over 12 weeks (Lally, van Jaarsveld, Potts and Wardle, 2010, European Journal of Social Psychology). Two findings matter here.

First, habits took a median of 66 days to become automatic, with enormous individual variation, from 18 days to well beyond the study period. The popular "21 days" figure has no basis in this data.

Second, and more useful: missing a single day had no measurable effect on the eventual strength of the habit. The trajectory absorbed it. What mattered was resuming.

The decisive variable is what you do on day two, not whether day one was perfect.

That is the opposite of the streak design that dominates fitness and fasting apps, where a streak is binary. Hit the target or the counter resets to zero.

The what the hell effect

Behavioural scientists have a name for what happens next. After a perceived lapse, people frequently abandon the goal entirely rather than resume, an effect first documented in dieting research by Polivy and Herman and now widely replicated under the label "the what the hell effect."

A streak that resets to zero is an unusually clean trigger for it. You did not fail at fasting. You failed at a number on a screen, and the screen told you so.

The 50 percent threshold

Ember's streak threshold is half your protocol target. Fast eight hours on a 16:8, nine on an 18:6, and the streak holds.

This is not lowering the bar. Your target is still your target. This is about what happens when you do not quite clear it.

Someone on a 16:8 who fasted for 14 hours did not fail. They fasted for 14 hours. Insulin fell. Depending on the person, they may well have crossed the metabolic switch, which the literature places anywhere between hour 12 and hour 36 (Anton et al., 2018, Obesity). The only thing that did not happen was the extra two hours that would have made a calendar entry look tidy.

Resetting the streak punishes that day. Holding it reflects what actually happened.

Why streak deletion exists

Because it works in the short term. A user who feels they have broken something is, for a while, more likely to come back and start over. That shows up quickly in retention dashboards.

Whether it works over months is a different question, and I want to be careful here: I am not aware of published data comparing long term retention between apps with forgiving and unforgiving streaks. If it exists, I have not found it. What we have is the habit formation literature above, which is about people rather than apps, and it points clearly in one direction.

So this is a design conviction, argued from evidence about how habits form, not a claim about anybody's retention numbers.

What good streak design looks like

Three principles:

1. The streak should track the behaviour, not the app. If you fasted, it should hold whether or not you opened anything.
2. It should be forgiving of partial days. A 15 hour fast on a 16:8 is a slightly short success.
3. It should never use guilt. "You missed a day." "Your streak is at risk." These train avoidance, and avoidance ends habits.

Ember does the first two by default. The third was a writing rule we held throughout the app.

How to fast through a bad day

• If you cannot hit your target, hit half. A nine hour fast still counts, and insulin still falls.
• If you cannot hit half, eat normally and pick it up tomorrow. One day off is noise. Several in a row is a pattern.
• Do not compensate by fasting aggressively the next day. That is the what the hell effect wearing a disguise.

The version of consistency that compounds is unglamorous. Most days hit. Some days half hit. Very few days zero. Over a year, that is the only version that survives.

Ember treats a partial fast as a partial win. Break above 50 percent of your target and the streak holds. Habits compound; perfectionism does not. $14.99 a year, no account, no guilt trip when you miss a day.

This article is for general information and is not medical advice. Fasting is not appropriate for everyone. Speak to a clinician before fasting if you are pregnant or breastfeeding, live with diabetes, take medication affecting blood sugar, or have any history of disordered eating.