How to track mood triggers without guessing causes

A trigger, as most trackers use the term, is something recorded near a mood change. That is a description of position in time, not of cause. Two things can appear together because one affects the other, because a third thing affects both, because you only notice one when the other happens, or by coincidence over a small number of days.

This guide keeps the useful part of trigger tracking—noticing what recurs—while refusing the step that makes it unreliable. Throughout, the working phrase iscandidate, not cause.

Record fewer candidates, more carefully

The instinct is to tag everything so nothing is missed. In practice a long tag list makes entries slower, so you log less often, and the entries you do make cluster on memorable days. A short list you actually complete beats a thorough one you abandon.

Start with three or four candidates for a fortnight. Useful ones are specific, observable without judgement, and things you could plausibly notice every day—“slept under six hours,” “ate lunch away from my desk,” “spoke to nobody before noon.” Vague candidates such as “stress” describe the mood rather than something near it, so they cannot function as a separate observation.

Log the absences too

This is the step most trigger tracking skips, and it is the one that does the work. If you only record a candidate on the days it appears, you cannot learn anything: every entry agrees with your hypothesis by construction.

So record the candidate’s state whether or not it happened. Fourteen days gives you four groups rather than one: candidate present with a lower score, present with a higher score, absent with a lower score, absent with a higher score. A candidate worth keeping shows up more in the first group than the third. A candidate that appears about as often on unremarkable days is telling you it is common, not that it matters.

A fourteen-day method

1. Choose three candidates and write down, in advance, what you expect each one to look like. Predicting first is what stops the result being fitted afterwards.
2. Log at roughly the same time each day: a score, and present-or-absent for each candidate. Same time matters, because time of day is itself a candidate.
3. Add one short note only when something stood out. Theemotion record templateis the longer form for those individual events.
4. Do not review until day fourteen. Reading a running total mid-way changes what you notice and what you bother to log.
5. On day fourteen, sort the days into the four groups above and count them. Counting beats remembering, which favours vivid days.
6. Keep at most one candidate for a second fortnight. Drop the rest, including ones that merely looked interesting.

What a fortnight cannot tell you

Fourteen entries is a small record. It can show that something recurs and is worth watching again. It cannot establish a cause, rule one out, or support a number—“my mood is twelve per cent lower when…” is a statistic that fourteen self-selected days cannot carry.

Reverse direction is the alternative people forget. A low stretch can reduce sleep, cancel plans, and shorten conversations. The candidate you recorded may be a consequence of the mood rather than something preceding it, and the order in your notes will not distinguish the two.

Two more limits worth writing down. Anything absent from your record cannot appear as a candidate, however much it mattered. And once you are watching for a candidate you notice it more, which inflates it relative to the fortnight before you started looking.

Turning a candidate into a next step

• Observation: “Of eleven logged days, the four lowest followed nights under six hours.”
• Boundary: “Eleven of fourteen days were logged; three had no sleep estimate.”
• Alternatives: “Late work shortened both sleep and the evening; or a harder week reduced sleep.”
• Next step: “Log sleep and a score for two more weeks, including the nights I sleep normally.”

Sleep is the most common first candidate and has its own pitfalls, covered intracking sleep and mood. For the weekly habit of reading entries without over-reading them, use theweekly check-in questions; to improve the inputs themselves, see thedaily questions.

Doing this in an app

Vibbrancy groups context-adjacent cards under a Drivers view in Insights. “Driver” there is a navigation label for cards that sit near your entries in time—it is not a finding that a factor drove a change. Which cards appear depends on the fields you used, the current plan, platform, permissions, and how much you have recorded; a card that is locked, empty, or still building progress means no supported result, not a negative result. TheInsights walkthroughcovers those states in detail.

Candidate lists are unusually revealing—people, places, health details, and habits, dated and in one file. Decide what you are willing to store before you build the list, and check what an export contains and what deletion removes. Theprivacy overviewand theprivacy policydescribe what is stored and where.

Nothing here diagnoses a condition or predicts what happens next. If what you notice is severe, persistent, or distressing, a qualified professional can help place it in context—and a short list of dated examples is more useful in that conversation than a conclusion. If hunting for triggers increases rumination or self-blame, stop the exercise. In immediate danger or crisis, contact an appropriate local emergency or crisis service rather than reviewing a log.