How to set goals that stick
Almost everyone can set a goal. The part that goes wrong is the month after. Crystal is an AI coach that keeps the goal in front of you once the enthusiasm has worn off, and asks about it on a schedule you agreed to in advance.
Goals stick when they are specific enough to fail at, small enough to start this week, and attached to a date you have already agreed to review them. Most goal setting stops at the first of those. Write the outcome as a sentence someone else could check, name the first action and when it happens, then put a weekly review in the calendar before anything else. A goal with no review cadence is a wish with a deadline on it.
Why most goal advice fails
The standard advice is a format. Make it specific, make it measurable, give it a date. That is useful, and it is also the easy part. A badly worded goal and a well worded goal both get abandoned in week three, for the same reasons, and the wording is not one of them.
What actually separates the goals that survive is unglamorous: there are few of them, the first step is small enough to do today, and something brings them back into view after the initial enthusiasm has gone.
It stops at the outcome
"Get fit" and "run a half marathon in May" are both outcomes. Neither tells you what happens on Tuesday. The plan is a separate piece of work and it is the piece people skip.
It encourages too many at once
January lists run to eight or ten items. Two or three is what most people can actually hold, and the rest quietly become a source of guilt rather than direction.
It has no review date
A goal nobody looks at is indistinguishable from a goal nobody has. The review is not admin around the goal, it is the mechanism that keeps it real.
It treats a missed week as a verdict
Most people quit after the first miss, not the tenth. A goal that has no written plan for what happens when you miss one is a goal built to be abandoned on its first bad week.
A goal is not a plan
A goal describes a state of the world that does not exist yet. A plan describes the actions you will take this week. They are different documents, they are written differently, and conflating them is the most common failure in the whole subject.
The test is simple. Read what you wrote and ask whether you could start it in the next twenty minutes. "Save six thousand dollars this year" fails the test. "Move four hundred into savings on the first of each month, starting Saturday" passes it. The first is the goal, the second is the plan, and you need both.
Writing the plan also exposes whether the goal is possible at all. Many goals die not because of willpower but because the arithmetic never worked, and the arithmetic only shows up once you write down the weekly action.
A goal without a review cadence is a wish
The review is the whole difference between a goal and a note you once made. It does not need to be long. Ten minutes a week, looking at what you said you would do and what happened, is enough to catch a drifting goal while it is still recoverable.
What a review is for is adjustment, not judgement. Three outcomes are legitimate every week: keep going, change the plan, or drop the goal because it is no longer the right one. Dropping a goal on purpose is a good result. Letting it fade is not.
In Crystal, goals and future snapshots sit in Progress, the forward-looking part of the product, on its Future page. Crystal holds the review cadence so that it is not one more thing to remember, and it brings back what you wrote when you set the goal rather than asking you to recall it.
Go deeper
Where to start.
How to set goals
This is the method, in order, with the parts most guides leave out: how many to set, what the first week looks like, and what you do on the week it goes wrong.
ReadGoal setting worksheet
Nine prompts, in order, designed to be answered on one side of paper. It takes about thirty minutes and the last two prompts are the ones that matter most.
ReadSMART goals
The best known goal framework there is, and a genuinely useful one, as long as you know which part of the problem it solves and which part it leaves to you.
ReadLetter to your future self
The hard part is not the idea, it is the blank page. Here is what is worth writing down, and what you will wish you had included when you read it back.
ReadCommon questions
What people ask about goals.
How do you set goals?
Write the outcome as a sentence someone else could check, pick a date, then write the first action and when it happens this week. Set a weekly review before you start. Those four things, in that order, are the whole method. Everything else is refinement.
How many goals should you have at once?
Two or three. Attention is the binding constraint, not ambition, and a list of eight goals usually functions as a list of zero. If more than three feel essential, rank them and let the lower ones wait for the next quarter.
What makes a goal a good goal?
It can be checked by someone other than you, it has a date, and it implies an action you could take this week. If a goal fails any of those three, the fix is usually to make it smaller rather than to reword it.
How often should you review your goals?
Weekly for the plan, quarterly for the goal itself. The weekly review asks what happened and what changes. The quarterly one asks whether the goal is still the right goal, which is a different question and gets missed when the only cadence is weekly.
What should you do when you miss a goal?
Treat it as information rather than failure. Ask whether the plan was wrong, the week was unusual, or the goal was never achievable, because the three have different fixes. The dangerous response is not the miss, it is quitting after one.
Crystal is the AI coach that understands you.
Personality looks inward, Relationships looks outward, Progress looks forward. Start with the part you need most.