Design the record around what comes back out of it#
A food diary you will actually keep has three properties, and none of them is about willpower. It demands very few fields — the food, roughly how much, and when. It records a quantity rather than only a name, because a record with no amount in it cannot teach you anything later. And it returns something to you on the same day you write it, because a diary that only absorbs information is a chore with no payoff attached.
Those are setup decisions, made once, in about ten minutes. They are worth separating from the question of whether you can stick to a new routine, which is a different problem with its own literature and its own answers. The evidence below is specifically about the artifact: what the record is made of, what goes in it, and what it does back. Two randomized trials suggest those choices matter more than most people setting up a diary assume — and in one specific way that runs against the obvious expectation.
The medium bought adherence. Only the feedback bought results.#
The cleanest test of this is a three-arm trial in which 210 adults in a behavioural weight-loss programme were randomized to keep their food record on paper, on a handheld device, or on the same device with a daily tailored feedback message1.
| Arm | Median adherence over 6 months | Adherence at month 6 | Weight change at 6 months | Reached ≥5% loss |
|---|---|---|---|---|
| Paper record | 55% | 31% | −5.3% | 46% |
| Handheld device | 80% | 53% | −5.5% | 49% |
| Device + daily feedback | 90% | 60% | −7.3% | 63% |
Data: Burke et al., 2011; n=210, 91% retention.
The first column behaves exactly as you would expect: making the record easier to keep nearly doubled how much of it got kept, and the gap widened as the months passed. Paper had fallen to 31 percent of days by month six; the feedback arm was still at 60.
The second half of the table is the surprise. Going from 55 percent adherence to 80 percent — a very large change in how much of your eating is written down — moved mean weight loss from −5.3% to −5.5%, which is nothing. Only the arm that got something back each day pulled away, reaching a 5 percent loss in 63 percent of participants against 46 percent on paper (P=0.04). The trial's authors do not paper over the oddity; they write that "it was unclear why the improved adherence in the PDA without feedback group did not lead to a greater weight loss."
Filling the diary in more often did not, by itself, change the outcome. What changed the outcome was the diary answering.
That is the single most useful thing to know when setting one up. Convenience is worth buying — 31 percent of days is not a diary, it is a memory — but convenience alone was not the active ingredient. The setup question is not "which app is easiest," it is "what does this thing tell me today that I did not already know."
A record with no quantity in it cannot be learned from#
The second design choice is what each entry contains. It is tempting to strip a diary down to photographs, on the reasonable theory that the fastest capture is the one that survives. A small randomized pilot tested exactly that: 41 adults with overweight or obesity, all getting the same twice-weekly behavioural podcasts, randomized to track either with a calorie-tracking app or with a photo-based app2.
Both arms lost weight, and by essentially the same amount — −2.4 ± 0.9 kg for calorie tracking and −2.5 ± 0.9 kg for photos. Tracking frequency did not differ between them, and was low in both (under 30 percent of days). But the relationship inside each arm split: how often someone tracked predicted how much they lost among the calorie trackers (r = 0.70, P = 0.004) and did not reach significance among the photo trackers (r = 0.51, P = 0.06).
Hold that loosely — it is 41 people, a pilot, and a correlation of 0.51 in a sample that size is not a demonstration of absence. But the direction fits the mechanism, and the mechanism is not mysterious. A photograph is a perfect record of what and a silent one about how much. Doubling the number of photographs doubles your archive without adding a single number you can total, compare, or be surprised by. If the point of the diary is to make a quantity visible, the quantity has to be in it.
The practical version is not "weigh everything." It is that every entry needs an amount attached, even a bad one. "Pasta, big bowl" is a quantity. "Pasta" is not.
Reactivity is a design parameter, not a defect#
There is a well-documented property of food records that most guides mention as a warning and almost nobody uses on purpose. Keeping a record changes what gets eaten: reviews of the method describe reactivity as people "changing usual dietary patterns for ease of recording or social desirability to report foods perceived as 'healthy'"3.
Read the first half of that sentence again. For ease of recording. Your diary will quietly push you toward whatever it is least annoying to log, which for most setups means packaged, barcoded, single-ingredient, repeatable food. That drift is going to happen whether you plan it or not, so the setup decision is which direction you want it pointing.
A diary is not only measuring your eating. It is applying a small, constant force to it — in the direction of whatever is easiest to write down.
So load the easy path deliberately. Pre-enter the six or eight things you eat most, at the portions you actually use, before you log a single real day. If your saved entries are a chicken-and-rice bowl, Greek yoghurt, oats, and a standard dinner, the friction gradient runs toward those. If your only saved entries are the packaged snacks whose barcodes scanned cleanly, it runs there instead. Pre-loading the log is a miniature version of planning the week, which in 40,554 French adults tracked with higher food variety and better guideline adherence4.
The same source explains why research food records run three to four days rather than a month: "participant burden generally causes a decline in the quality of information recorded if more days are recorded." Professional dietary assessment does not ask for indefinite records because it knows the data rots. Your diary is asking for indefinite records, which means keeping per-entry burden genuinely low is not laziness — it is the only way the later entries stay worth reading.
The fields, and what to leave out#
| Field | Keep it? | Why |
|---|---|---|
| Food | Required | Nothing works without it |
| Amount (any resolution) | Required | The one field that makes totals and comparisons possible |
| Time | Required, but automatic | Free to capture; it is where weekly patterns show up |
| Photo | Optional, as an attachment | Excellent capture, poor record — it holds no quantity |
| Location or company | Optional | Cheap if it is a tap; the second-best predictor of a surprising day |
| Hunger or mood rating | Skip at setup | High per-entry cost, and it is a research instrument, not a habit |
| Full macro breakdown | Skip at setup | Add it later if a specific question demands it |
Three required fields is the whole design. Everything below the line in that table is something to add in month three if a question comes up that the log cannot answer — and most of those questions never come up.
Two more setup moves and you are finished. Set the diary to show you a running total during the day rather than only a verdict at midnight, because a number that arrives while you can still act on it is the version that returns something. And decide now what you will do with a blank day, because you will have them; a diary designed on the assumption of perfect completeness breaks the first time reality touches it, which is why how you handle the days you miss belongs in the setup rather than in the recovery. Then leave it alone for a week before you read anything into it — the reading is a separate skill with a separate cadence, covered in reviewing the log.
Done properly, this takes ten minutes and decides more about whether the diary survives March than any amount of resolve does. The rest is the standard counting workflow and starting small enough not to burn out.
FAQ#
Is a paper food diary as good as a phone app?#
For results, the trial evidence says roughly yes; for staying kept, clearly no. In a three-arm randomized trial, mean weight loss on paper (−5.3%) and on a handheld device (−5.5%) was statistically indistinguishable, but median adherence over six months was 55% on paper against 80% on the device, and paper had dropped to 31% of days by month six1. Paper works while you keep it. The device makes keeping it likelier.
What should a single diary entry actually contain?#
Three things: the food, an amount at whatever resolution you can manage, and the time. The amount is the one people drop and the one that makes the record useful — logging frequency predicted weight loss among calorie trackers (r = 0.70) but not among people recording photographs alone2. Everything else — macros, mood, location — is an optional field to add later, if a question arises that the basic record cannot answer.
Should I log before or after I eat?#
Before, when the choice is still open, and after when it is not — but the important thing is that the running total is visible during the day rather than assembled at midnight. The arm of the Burke trial that received a tailored message each day was the only one that pulled ahead on outcomes, despite the no-feedback arm logging nearly as diligently. A diary that reports back while you can still act is doing a different job from one that only records.
Sources#
- Burke LE, Conroy MB, Sereika SM, et al. The effect of electronic self-monitoring on weight loss and dietary intake: a randomized behavioral weight loss trial. Obesity (Silver Spring). 2011;19(2):338-344.
- Dunn CG, Turner-McGrievy GM, Wilcox S, Hutto B. Dietary Self-Monitoring Through Calorie Tracking but Not Through a Digital Photography App Is Associated with Significant Weight Loss: The 2SMART Pilot Study—A 6-Month Randomized Trial. J Acad Nutr Diet. 2019;119(9):1525-1532.
- Bailey RL. Overview of Dietary Assessment Methods for Measuring Intakes of Foods, Beverages, and Dietary Supplements in Research Studies. Curr Opin Biotechnol. 2021;70:91-96.
- Ducrot P, Méjean C, Aroumougame V, et al. Meal planning is associated with food variety, diet quality and body weight status in a large sample of French adults. Int J Behav Nutr Phys Act. 2017;14:12.
