Does mood tracking actually work? A research review

"Does mood tracking work?" is the wrong question. The honest answer depends on what kind of tracking, attached to what kind of intervention, in whose hands. The research base is real, but it's narrower and more cautious than the marketing copy on most apps suggests.

What "mood tracking" actually means

Two things often get blurred under the same name. The first is self-monitoring: a person records their own emotional state, usually daily, for their own use, to spot patterns, plan around them, or bring the data to a therapist. The second is clinician or researcher data collection: structured sampling (often called ecological momentary assessment, or EMA) used to measure symptoms in studies and clinical care. Most of the high-quality evidence sits in the second category, and the implications for the first are partial.

The distinction matters because EMA has been validated as an assessment method. Trull and Ebner-Priemer's 2009 review in Psychological Assessment showed that repeated in-the-moment sampling captures emotional patterns that retrospective recall systematically distorts.³ That's a strong finding. It does not, by itself, mean the act of recording your mood makes you feel better. The two questions get answered separately.

What the trials show works

The clearest signal is in bipolar disorder. The MONARCA I trial (Faurholt-Jepsen et al., 2015, Psychological Medicine) randomised 78 patients to six months of daily smartphone self-monitoring versus standard care.² Self-reported depressive symptoms dropped significantly more in the tracking group. The catch: the trial's pre-specified primary outcome (clinician-rated depression and mania scores combined) did not reach significance, mostly because mania scores moved very little. So the headline is real but specific: in bipolar patients, daily structured self-monitoring reduced self-rated depressive symptoms over six months, while leaving manic symptoms largely untouched. Earlier work from the same group¹ showed smartphone-recorded behavioural signals (call patterns, activity) correlated with clinician-rated symptom severity, which is what enabled the trial in the first place.

For general public mental health, Bakker, Kazantzis, Rickwood and Rickard's 2018 RCT in Behaviour Research and Therapy compared three smartphone apps against a wait-list with 226 participants.⁴ The apps that included CBT-style components (mood tracking plus thought records, behavioural activation prompts, and emotion labelling) produced reliable, modest improvements in depression and anxiety scores at one month, with effects holding at three months. The trial's main contribution is showing that the components matter more than the tracking itself. Apps that asked users to do something with the data outperformed apps that mostly logged it.

What the trials show doesn't work, or is unclear

Tracking on its own, without prompts to act on the data, has weak evidence. Schueller, Aguilera and Mohr's 2017 review of ecological momentary interventions in Depression and Anxiety summarises the field bluntly:⁵ most studies show small effects, samples are small, follow-up is short, attrition is high, and unpublished null results almost certainly bias the literature upward. The interventions that worked best were the ones that combined assessment with active in-the-moment prompts (a coping skill, a behavioural suggestion) rather than logging alone.

There's also a basic logical limit. Naming a pattern is not the same as fixing it. Knowing that Sundays are bad and that mornings before work meetings are worse gives you information; the change still depends on what you do with the information. The behaviour-change literature has been clear on this for decades, and mood tracking apps don't escape it.

When tracking can backfire

Self-monitoring is not free of side effects. The literature on EMA and rumination notes that asking already-ruminative people to focus repeatedly on their internal state can extend, rather than break, the rumination cycle. Schueller's review explicitly flags this risk.⁵ Two related patterns:

Excessive checking. The same machinery that drives compulsive symptom-checking in health anxiety can attach itself to mood tracking. Several entries a day, comparison against past entries, dread of opening the app: at that point the tool is generating the problem.

Hopelessness from the data. A user who logs honestly for three months and sees a flat line of "low" can read that as a verdict rather than a baseline to work against. This is more likely when tracking is the only intervention.

Harari and colleagues' 2016 review of smartphone behavioural data in Perspectives on Psychological Science makes a quieter point worth keeping in mind: the act of being measured changes what gets measured.⁶ Daily tracking shapes attention, framing, and recall in ways that aren't always neutral.

When mood tracking is most likely to help

A short list, drawn from the trials above:

The user wants the data. Imposed tracking (required by a clinician, pushed by an app's notifications) produces worse adherence and weaker effects than user-driven tracking.

Active intervention is attached. CBT-style prompts (thought records, distortion tagging, behavioural activation suggestions), notes that get shared with a therapist, or in-the-moment coping skills. Bakker 2018 is the cleanest demonstration.⁴

Specific clinical contexts. Bipolar disorder, especially in remission and as an early-warning system; mood disorders during post-therapy maintenance; depression where behavioural activation is the active ingredient.

Granular emotion labelling rather than a 1–5 mood scale. Specific labels ("frustrated", "ashamed", "defeated") give the cognitive work something to grip, which is the bridge into cognitive distortion tagging and CBT-style thought records.

Colors covers the components with the best evidence — granular emotion labelling, CBT-style thought records, and between-session note-taking through the reframe flow — and skips the bits that consistently fail to move outcomes (gamified streaks, AI mood predictions). It is not a substitute for therapy, and the research is clear that tracking alone, without action, has weak effects. As a journal that prompts you to do the cognitive work and a record you can bring to a clinician, it sits in the part of the evidence base that holds up.

References

1. Faurholt-Jepsen, M., Frost, M., Vinberg, M., Christensen, E. M., Bardram, J. E., & Kessing, L. V. (2014). Smartphone data as objective measures of bipolar disorder symptoms. Psychiatry Research, 217(1–2), 124–127. doi:10.1016/j.psychres.2014.03.009
2. Faurholt-Jepsen, M., Frost, M., Ritz, C., et al. (2015). Daily electronic self-monitoring in bipolar disorder using smartphones — the MONARCA I trial. Psychological Medicine, 45(13), 2691–2704. doi:10.1017/S0033291715000410
3. Trull, T. J., & Ebner-Priemer, U. W. (2009). Using experience sampling methods/ecological momentary assessment (ESM/EMA) in clinical assessment and clinical research. Psychological Assessment, 21(4), 457–462. doi:10.1037/a0017653
4. Bakker, D., Kazantzis, N., Rickwood, D., & Rickard, N. (2018). A randomized controlled trial of three smartphone apps for enhancing public mental health. Behaviour Research and Therapy, 109, 75–83. doi:10.1016/j.brat.2018.08.003
5. Schueller, S. M., Aguilera, A., & Mohr, D. C. (2017). Ecological momentary interventions for depression and anxiety. Depression and Anxiety, 34(6), 540–545. doi:10.1002/da.22649
6. Harari, G. M., Lane, N. D., Wang, R., Crosier, B. S., Campbell, A. T., & Gosling, S. D. (2016). Using smartphones to collect behavioral data in psychological science. Perspectives on Psychological Science, 11(6), 838–854. doi:10.1177/1745691616650285