Run an Energy Audit: Tracking What Drains You for One Week

An energy audit is seven days of timestamped self-observation: you rate your energy at fixed points, note what you were doing and who you were with, then read the week as data rather than memory. This guide covers what to log, how to score it so entries stay comparable, what a real filled-in log looks like, and how to separate genuine drains from assumed ones.

What is an energy audit, and what does one week of tracking reveal?

An energy audit is one week of timestamped self-observation in which you rate your energy several times a day, note what you were doing and who you were with, and then read the seven days as data rather than trusting your memory of them. The format is deliberately plain: three or four entries a day, one number, a handful of words, nothing more. Its value comes from the gap between what actually happened to your energy and what you remember happening. Memory compresses a day into its most intense moment and its ending, the distortion Barbara Fredrickson and Daniel Kahneman identified in 1993 and called the peak-end rule, and it discounts how long anything went on, which the same pair called duration neglect. So one bruising meeting at four in the afternoon can repaint an otherwise steady Tuesday as a wreck. A log written at the time does not do that. Seven days is the shortest window that contains a full work cycle and at least one weekend, which is enough to show whether your low points track the clock, the calendar, or particular people. What one week reveals is usually narrower than people expect. The finding is rarely that your job exhausts you. It is that Wednesday afternoons do, or that unstructured group time after six does, while the commitments you dreaded turned out to cost almost nothing.

Two unrelated traditions arrive at the same format. In the late 1920s and 1930s the British statistician L. H. C. Tippett, working at the Shirley Institute on problems in the cotton industry, developed what he called a snap-reading method: you could describe how a factory spent its hours by recording what was happening at scattered moments rather than timing everything. He set it out in the Journal of the Textile Institute in 1935, and it became known as work sampling. Decades later Mihaly Csikszentmihalyi and Reed Larson built the Experience Sampling Method, which paged people at intervals and asked what they were doing and how they felt. The first pager studies ran in Csikszentmihalyi's lab at the University of Chicago in the early 1970s, and the two of them set out the method's validity and reliability in the Journal of Nervous and Mental Disease in 1987. It became a standard way of studying how mood and attention move across an ordinary day. An energy audit is a domestic version of both, run by the subject on themselves. That last part introduces a known problem: self-monitoring changes the thing being monitored. Rating your energy at four in the afternoon makes you notice four in the afternoon, and some people find their scores drift upward during the audit week purely because attention is a mild intervention. Treat that as a cost you price in rather than a flaw that ruins the data, and compare days within the week against each other rather than against some imagined normal. Kahneman would not have called the log straightforwardly truer, either. His argument was that the remembering self, not the experiencing self, is the one that decides what you do next, so a week you recall as good carries consequences that a logged week does not. The audit takes the experiencing self's side deliberately. That makes it a good guide to what to change and a poor guide to what you will want.

Is an energy audit the same as a time audit?

No. A time audit records where your hours went; an energy audit records what those hours cost you. The two often disagree in useful ways. A task can eat ninety minutes and leave you fine, while a twelve-minute phone call flattens the rest of the afternoon. Run them together if you like, but the energy column is the one that changes decisions.

Do I need any special equipment or app to run one?

No. Csikszentmihalyi's original studies ran on radio pagers in the early 1970s because nothing better existed then, and the format outlived the hardware, which tells you the equipment was never the point. A pen and any paper you keep on you will do, with three or four phone alarms as the prompt. Apps that score your energy for you defeat the exercise, because assigning the number yourself is the measurement.

What if my week turns out to be unusual?

Log it anyway and mark the anomaly in the notes column. An unusual week still shows how you respond to disruption, which is worth knowing on its own. What you cannot do is generalize from it, so if a funeral, a deadline or a bout of flu lands mid-audit, finish the seven days, read them as a record of that week, then repeat the audit a month later for a comparison.

What should you log in an energy audit, and how many times a day?

Log four things at each entry: the time, an energy score, one line on what you were doing in the half hour before, and who you were with. Four entries a day is the working default, placed at waking, late morning, mid-afternoon and bedtime. That spacing catches the two transitions where most of the interesting variance lives, the morning climb and the afternoon dip, without turning the audit into a second job. Three entries works if four feels punishing. Two does not, because a morning and an evening reading cannot distinguish a day that sagged at two from a day that sagged at five, and that distinction is most of the point. Add three optional columns only if you will actually fill them: hours slept the previous night, caffeine and its timing, and any movement longer than ten minutes. Sleep is the largest confounder in any energy log, so record the number even if you record nothing else optional. Record clock times rather than a bare total, since seven hours ending at five in the morning and seven hours ending at eight produce different Tuesdays. Whatever set of columns you choose on day one, keep it unchanged through day seven. Adding a column on Thursday means the last four days cannot be compared with the first three, which is how one promising audit becomes two short ones.

The two hardest columns are the honest ones. Who you were with gets skipped because it feels like bookkeeping on your relationships, and it is exactly the column that produces the audit's most uncomfortable findings, since a repeated name beside repeated low scores is difficult to argue with. The free-text note gets skipped because writing feels like effort at the moment your energy is lowest, which is when the note matters most. Cap it at six words and it survives the week. Keep interpretation out of the entry itself. Write "call with M, 40 min" rather than "draining call with M", because a judgment written into the raw data can never be re-examined later. You are collecting; the reading happens on day eight. One more discipline: log the entry at the scheduled time even when nothing happened. Blank stretches of an ordinary Tuesday are the baseline that makes the bad Thursday legible, and audits that record only eventful moments produce a week that looks far worse than it actually was.

Should I log at fixed times or whenever something changes?

Fixed times. Event-triggered logging looks more efficient and produces unusable data, because you notice changes when they are dramatic and miss them when they are gradual, so the record skews toward drama. Fixed intervals sample the boring hours too. If a genuinely large shift happens between scheduled entries, add an extra line and mark it as unscheduled so it does not distort the averages.

How do I remember to log four times a day?

Attach each entry to something you already do without thinking: the first coffee, the walk back from lunch, closing the laptop, putting the phone on charge. Alarms on their own are easy to swipe away once the novelty wears off, and a dismissed alarm costs you the entry. The mechanics of anchoring a new behavior to an existing one, and what to do on the day you miss, are covered in James Clear: Key Ideas, Methods and How to Apply Them.

Do I log weekends the same way?

Yes. Shifting the times to match when you actually wake is fine, as long as the spacing stays roughly even. Weekend entries do the heavy lifting in the analysis, because they are your only reading of what happens when the calendar stops supplying structure for you. Do not assume in advance that the weekend holds your high scores, since some audits put the week's lowest entry on a Sunday afternoon.

How do you score your energy so the numbers stay comparable across the week?

Score energy on a fixed 1 to 10 scale with written anchors you define before day one, and never redefine them mid-week. The anchors are what make the number mean anything. Write down, in your own words, what a 2 feels like, what a 5 feels like and what an 8 feels like, then reread those definitions before your first few entries until the scale becomes automatic. A workable set: 1 to 2 is barely functional, capable of nothing beyond passive input. 3 to 4 is dragging, able to do familiar work slowly. 5 to 6 is ordinary, neither pulled down nor pushed along. 7 to 8 is capable and clear, the state you would choose for hard work. 9 to 10 is rare and slightly elevated, worth noticing precisely because it is rare. If ten points feel too fine-grained, use a five-point scale, which is easier to apply the same way twice. Consistency beats resolution here. Score the moment you are in rather than the hours behind you, and score before you look at what you wrote yesterday, since seeing a 4 primes you to write another 4.

Consider splitting the number in two. Physical energy and mental energy come apart often enough that a single score hides it: the state after a long walk is bodily tired and mentally bright, and the state after four hours of video calls is the reverse. Two columns, each 1 to 10, cost about three extra seconds per entry. Some people add a third for emotional charge. Beyond three the entry stops being quick and the audit dies by Wednesday. Scale drift is the other threat. By day five a 4 has quietly become what a 5 meant on Monday, usually because the week has been hard and your reference point moved with it. Rereading your written anchors at the start of each day is the cheap fix. Anchored self-report has a long working history outside energy tracking. The Swedish psychologist Gunnar Borg built his 6 to 20 scale of perceived exertion around this exact problem, and attached verbal descriptions at intervals along it, running from no exertion at all at 6 through somewhat hard at 13 to maximal exertion at 20, so that athletes and cardiac patients would use it the same way from one session to the next. The odd range was deliberate. Borg set it so that multiplying a rating by ten gives a rough heart rate for a healthy adult, which puts a 13 near 130 beats per minute and tethered the number to something outside the rater's head. That correspondence is approximate and it shifts with age, fitness and medication, but it gave the scale an external check. An energy audit has no such tether. Nothing external corroborates your 6, which is why the written anchors have to do all the work.

Is a 1 to 10 scale better than high, medium and low?

Marginally, and only because three levels drop most of an ordinary week into the middle bucket, where no trend can show. Ten levels move the problem the other way. On a day when you cannot tell a 6 from a 7, the extra resolution is invented rather than measured. George Miller's 1956 paper on the magical number seven put the limit at roughly seven categories along a single dimension before judgments blur, though what he was reviewing were other people's experiments on judging tones and salt concentrations rather than anything resembling self-report, so read it as an analogy and not a result about energy. Five points is a reasonable place to settle after a first audit.

What if I cannot tell what number to give?

Ask one concrete question instead of introspecting: what could I start right now without negotiating with myself? Nothing at all is a 1 or a 2. Email and tidying is a 4. The hardest thing on today's list is an 8. Behavioral anchors beat feeling-based ones because they stay stable across days when your mood is coloring everything you look at.

Should I score how I feel or how much I got done?

How you feel, with output kept in a separate column if you want it at all. Productivity and energy diverge in both directions: a grim, dragging morning can still produce work through sheer habit, and a bright afternoon can vanish into nothing much. Merging the two makes drains invisible, since the days you pushed hardest through low energy will look identical to genuinely good ones. Those are the days worth catching.

What does a filled-in energy audit log actually look like?

A filled-in entry is one line of about a dozen words: time, score, activity, company, short note. A real Wednesday from a four-entry audit reads something like this. 07:10, 5, woke before alarm, alone, slept six hours. 11:30, 7, drafting the budget, alone, best stretch so far. 15:00, 3, back-to-back calls since one, four colleagues, could not follow the last one. 22:15, 4, television, partner, never recovered after three. Four short lines, under a minute of writing, and already a shape: a strong late morning, a hard fall somewhere after one, and no recovery in the evening. Compare that with the Saturday from the same log. 09:40, 6, slow start, alone, slept eight hours. 12:30, 8, long walk, alone, best score of the week. 16:00, 6, shopping, partner, fine. 22:00, 5, late dinner with friends, six people, wanted to leave by nine. The Saturday entries do something the Wednesday ones cannot: they show the same person reaching an 8, which establishes that the 3 on Wednesday was situational rather than a fixed ceiling. Keep the format identical every day so the week can be read down a column.

Failed logs fail in recognizable ways. The most common is the retrospective fill, where four entries appear in one burst at eleven at night in a suspiciously smooth pattern, with the four numbers sitting within a point of each other. Real logs are jagged. Another is the essay, where the note column grows into paragraphs by day two and the audit is abandoned by day four. A third is the missing weekend, which quietly removes the only comparison that shows what your energy does without a calendar around it. A fourth is the interpreted entry: "draining meeting" instead of "planning meeting, 5 people". If you already know it was draining, you have skipped the audit and gone straight to the conclusion you arrived with. Gaps are not failures, and a missed entry marked as missed beats one invented an hour later. As a working rule, aim to fill at least twenty-four of the twenty-eight entries in a four-a-day week. Well below that, the analysis in the next section starts producing patterns that are artifacts of when you happened to remember rather than of the week itself. Experience sampling researchers meet the same problem at a larger scale: a 2019 review of response rates across such studies found compliance averaging around three-quarters of prompts and slipping further as the days went on.

How long should each entry take to write?

Twenty to thirty seconds. If an entry regularly takes longer, the format is too elaborate and it will not survive to Friday. The test is whether you can complete a line while still standing in a corridor. Anything requiring you to sit down, open an app and think will be skipped on precisely the low days that hold the most information.

Should I number the days or use dates?

Use both, plus the day of the week. Dates let you line the audit up against a calendar afterward, which is how you find out that the Thursday crash coincided with the weekly all-hands. Day numbers make the sequence easy to scan. Weekday names matter because several of the strongest patterns people find are weekday-shaped rather than date-shaped.

Can I keep the log on a spreadsheet?

Yes, and it earns its place on day eight rather than during the week. Sorting twenty-eight entries by score takes one click, whereas doing the same on paper means recopying the whole log. Set the columns before day one and leave them alone: timestamp, score, activity, company, note. The usual arrangement is paper during the day, then ten minutes on Sunday moving seven days into those columns, which doubles as a first read of the week.

How do you read the log once the seven days are over?

Read the log in four passes, and do not start until all seven days are in. First pass, draw the curve: days along the bottom, scores up the side, points joined. A shape appears within seconds that no amount of staring at rows will produce, usually a repeating daily arc plus one or two days that break it. Second pass, rank: list every entry from highest to lowest, then read the top five and the bottom five together and ignore the middle. Extremes carry most of the signal in a dataset this small. Third pass, cluster: write down what the bottom five have in common and what the top five have in common, in terms of time of day, activity type and who was present. Look for a repeated element rather than a single dramatic one. Fourth pass, name candidates: write no more than three suspected drains and three suspected restorers as plain sentences with the evidence attached, for instance, group calls longer than forty minutes appear in four of my five lowest entries. Three of each is the ceiling. A list of eleven candidates is a list you will never act on.

Two distinctions save you from the obvious mistakes. The first is level versus change. A 4 at nine in the evening on a day that started at 7 tells a different story from a 4 that has been flat since morning, so read the drop between consecutive entries as its own quantity rather than reading scores in isolation. Most drains show up as a delta rather than a level. The second is time of day versus content. Almost everyone has a mid-afternoon trough that arrives whether or not anything is happening, and the meeting scheduled inside that trough will be blamed for it. Check whether your low entries cluster by clock time across all seven days, weekend included. If they do, you are looking at a circadian pattern rather than a behavioral one, and the levers are physiological: light, sleep timing, and when the last caffeine landed. Caffeine has a half-life of roughly five hours in most healthy adults, with wide individual variation, so a three o'clock coffee is still half present at eight. The post-lunch dip is the useful test case. Sleep researchers, Timothy Monk among them, report that it appears in people who have eaten no lunch at all and who cannot see a clock, which points at body timing rather than the meal, and means the meeting you booked for two may have landed inside a trough it did not cause. If the lows cluster by activity instead, and the same clock times are fine on Saturday, the cause sits in your week rather than your body.

How many days does it take before a pattern is real?

Three occurrences is the working threshold for a seven-day audit, and even then it is a hypothesis rather than a finding. Something that happened once was probably just the day. Something that happened twice might be coincidence. Something that appears three times in seven days with the same shape each time is worth testing deliberately in a second audit. Single occurrences go in the notes, never in the conclusions.

What if the curve looks completely random?

Random-looking curves usually mean one of three things: the scale drifted, entries were filled in from memory, or a large hidden variable such as broken sleep is swamping everything else. Check the sleep column first. If the nights ranged from five hours to nine, that variation alone can account for a week of noise, and the audit needs rerunning across a more stable stretch.

Should I show the log to anyone else?

One person, chosen carefully. Someone who was present for part of the week will remember context you never wrote down, and on day eight that is worth a great deal. Avoid anyone whose name sits beside your lowest scores, since that conversation turns an analysis into a negotiation. Read them your conclusions first and the raw entries second, so they respond to the pattern rather than to one bad Wednesday.

How do you tell a real drain from something you only assume drains you?

A real drain shows up in the log on days you were not expecting it to; an assumed drain shows up only in your description of the week. The test is mechanical. Before you read the data, write down the three things you believe drain you most. Then check each one against the entries. Count how many of your lowest scores actually contain it, and count how many times it appeared without a low score following. That second number dissolves most assumptions. If the commute you have complained about for two years appears eight times in the week and sits beside a low score twice, it is not the drain worth attacking first. Three failure modes account for most false positives. Anticipation, where the dread before a task costs more than the task and the log shows the dip an hour early. Attribution, where a genuinely low afternoon gets pinned to the most memorable event inside it rather than the one that caused it. And contamination, where an activity is always scheduled next to the real culprit and inherits the blame for it, which is why the company column matters as much as the activity column.

The audit's more useful output is the drain nobody suspected. Look for entries with low scores and unremarkable content, the ones where the note says admin, or nothing much. Low-grade, undramatic activity is systematically underrated as a cost because it produces no story to remember afterward. Task switching, unfinished decisions carried around all day, and rooms with bad light do not feel like events, and they show up in the numbers anyway. Where a person's name recurs beside your lowest entries across several days and several settings, you have a different kind of finding, and it needs care: the first question is whether the pattern holds when the setting changes, since one difficult context can masquerade as one difficult person. Sort those entries by whether you chose to be there. A name that costs you two points in a meeting you were assigned to and nothing at all over lunch is telling you about the meeting. A name that costs you two points in both is telling you about the person. The audit's job stops at naming the pattern. Deciding what a repeated name means is separate work, and slower.

What if the thing I assumed was a drain turns out to be neutral?

Believe the log, then check it once. Assumptions about energy tend to come from a memorable bad instance rather than an average, so a task can carry a reputation earned on one afternoon two years ago. Run a second week with that activity flagged in advance, and if it still sits beside ordinary scores, retire the belief and stop spending planning effort on managing something that costs you nothing.

Can something drain me and still be worth doing?

Frequently, and the audit is no permission slip to remove everything expensive. Caring for a parent who is declining costs energy every week and is still the right call. So does the hard half of creative work. What the log gives you is the price, so you can schedule around it: put a known expensive commitment where a restorer can follow it, and stop stacking two of them in the same afternoon.

How do I test a suspected drain deliberately?

Pick one candidate and change a single variable for a week while logging exactly as before. Move the meeting from four o'clock to ten, or take the call standing and outdoors, or cut it from an hour to thirty minutes. One variable. Two changes leave you unable to say which one worked. Compare the entries immediately around that activity rather than your overall mood for the week.

What restores your energy, and how does the log prove it?

Restorers are the activities followed by a measurable rise in your score, and the log proves them by pairing the entry before with the entry after. That pairing is the whole method: a thing you enjoy is not automatically a restorer, and several genuinely restorative things are mildly unpleasant to begin. Read the deltas. Anything that reliably adds two points on three or more occasions is a real restorer for you. Anything that leaves the score flat is neutral, however good its reputation elsewhere. The candidates worth testing during an audit week are short and cheap: ten minutes outdoors in daylight, a nap under twenty-five minutes, a conversation with one specific person, a hot shower, food after a long gap, silence with no input at all. Test them one at a time and log the score immediately before and about thirty minutes after. Grounding and Earthing: Techniques, Claims and What Research Says covers one such candidate in detail, including where the evidence for it is thin, and treating it as a hypothesis to test in your own log rather than a settled fact is the honest way to use it. Scrolling frequently fails this test.

Rest and restoration come apart in the data more often than people expect. A pause stops the drain; a restorer moves the number up. Lying on the sofa with a phone usually registers as a pause, since the score at the end matches the score at the start, while ten minutes of walking outdoors often registers as a genuine climb. Neither result is universal, which is the reason for measuring yours rather than adopting somebody else's list. Watch the timing too. Some restorers act within minutes and some need an hour before the effect shows up in a score, so a restorer logged only immediately afterward can look like a failure when it is merely slow. Naps invert the problem. Sleep researchers put the usual onset of slow-wave sleep somewhere around the half-hour mark, which is the reason naps kept under about thirty minutes are the ones generally recommended for avoiding grogginess. Wake out of deeper sleep and you get sleep inertia, so the score taken ten minutes later can land below the one that sent you to lie down, then climb above it an hour on. Log both readings before concluding that napping does not work for you. The audit answers the slower question: across seven days, which things actually moved your scores, and what did each of them cost you in time?

How soon after a restorer should I log the score?

Log twice: immediately afterward, and again about thirty to sixty minutes later. The two readings answer different questions. The immediate one catches short-lived lifts that fade before they are any use, and the delayed one catches slower restorers that would otherwise look like failures. A restorer that still holds an hour later is the kind worth building a week around.

Does sleep count as a restorer in the log?

Sleep is better treated as a background variable than as one restorer among many, because it sets the ceiling every other score of the day sits under. Everything else gets measured underneath it. Record hours and a rough quality rating each morning, then read the whole day against them. A week where the sleep column swings by three hours will not produce clean findings about anything else.

What if nothing in my week raises the score?

A week with no upward deltas anywhere is itself a finding, and it usually has one of two causes. Either you logged the week without testing anything in it, so there were no candidates to rise from, or the lifts are real but smaller than your scale can resolve, which happens on a five-point scale where a genuine half-point gain rounds to nothing. Schedule three candidate restorers into a second week before concluding that nothing works. If that week also runs flat along the bottom of the scale, stop redesigning your calendar and see a clinician.

What do you change first once the audit is finished?

Change one thing, chosen from your list of three suspected drains, and pick the one with the best ratio of energy recovered to social cost. That second half is what stops audits ending in a plan nobody executes. Most people finish a week of data and attempt four changes at once, which makes the result uninterpretable and usually collapses inside a fortnight. One change, held for two weeks, logged the same way, gives you an answer you can trust. Rank your three candidates on two axes before choosing: how much energy the log says each drain costs, and how hard the change would be to make without a conversation you are dreading. The most expensive drain is often the hardest to move, so starting there produces a stalled audit and a sense that the whole week was wasted. Start instead with the largest drain you can change alone. Common first changes that need no negotiation: moving demanding work into the two-hour window where your scores are highest, adding a ten-minute recovery gap after whatever reliably costs two points, cutting one recurring commitment that appeared beside low scores and produced nothing you value, or ending the working day at a fixed time.

What to put into the freed time is a prioritization problem rather than a measurement one, and the urgent-versus-important sort and the weekly planning session in Stephen Covey: Key Ideas, Methods and How to Apply Them handle it better than any energy log can. Name the replacement before you remove the drain. The failure that catches people is subtler than a badly chosen change: the audit works, the drain goes, and inside a month the calendar has refilled the gap with something nobody measured, because by then the log had stopped. Then set the repeat. An energy audit is a periodic instrument rather than a permanent habit, and running it continuously turns it into a chore. Repeat it after any change worth verifying, after a job or household change, and otherwise roughly quarterly. Keep the same anchors and the same scale each time, or the weeks cannot be compared and repeating the audit buys you nothing.

How long should I hold a change before judging it?

Two weeks, with a repeat audit running through the second one. The first week after any change is contaminated by novelty, which can lift scores on its own and then fade. Judging at day three produces enthusiasm; judging at day fourteen produces information. If the change has not moved the entries around that activity by then, drop it and take the next candidate off your list.

What if the audit points at my job as a whole?

Break the finding down before acting on it, because my job is rarely what the entries actually say. Sort the low-scoring work entries by task type, by who was present and by time of day. Most jobs contain a small number of specific expensive elements surrounded by neutral ones, and the useful change is almost always to a meeting pattern, a shift boundary or one relationship rather than to the job itself.

Do I need to keep logging forever?

No, and there is a reason beyond tedium. A scale you use every day stops discriminating: by week three your 6 has quietly absorbed whatever the last fortnight felt like, and the anchors you wrote on day one no longer describe anything. Experience sampling studies usually run in short bursts, partly because compliance falls away the longer the signaling goes on. The audit works because it interrupts. Stop at day seven and act on one finding.

Frequently Asked Questions

How long does an energy audit take each day?

Under three minutes a day, spread across four entries of twenty to thirty seconds each, plus a minute in the evening if you add a short daily note. The analysis on day eight takes longer, around thirty to forty minutes for the four reading passes, and it is the only sitting the audit really demands. If the daily total creeps past five minutes, the format has grown too elaborate and it will be abandoned by Thursday. Cut the note column before you cut the number of entries.

Should I track my energy on paper or on my phone?

Paper for the entries, and a phone only for the reminders. A phone reached for at the exact moment you are lowest usually ends in twenty minutes of something else, and the entry you meant to write gets absorbed. A pocket notebook, an index card in a wallet, or a grid taped inside a cupboard door all work. Transfer the week into a spreadsheet on day eight if you want to sort by score, which is genuinely easier on a screen than on paper.

What if my energy is low in every single entry?

A week of uniformly low scores is a result, and it changes what you do next. First check the scale: if your written anchors put a normal day at 7 and everything came in at 3, the finding is real rather than an artifact. Persistent fatigue that rest does not touch, especially alongside low mood, disturbed sleep, breathlessness, weight change or new pain, is a reason to see a clinician and have the ordinary causes ruled out. Anemia, thyroid problems, sleep apnea and depression can all present this way, and no amount of calendar redesign addresses them.

Should I run the audit during a normal week or a busy one?

A normal week, which usually means the one you would describe as unremarkable rather than the one you dread. Extreme weeks produce vivid data that generalizes to nothing. Avoid weeks containing travel, illness, a deadline, houseguests or a holiday, and avoid the week immediately after any of those, since a recovery week is equally atypical. If your work runs on a monthly cycle with a heavy stretch and a quiet one, run the audit twice and compare the two rather than picking a winner.

Can I fill in the whole day from memory at bedtime?

No, and the resulting log is worse than no log. A bedtime reconstruction is written by the peak-end rule: whatever was sharpest and whatever came last set the tone for all four entries, so you get a smooth curve describing how you feel at ten at night rather than how you felt at eleven that morning. The distortion runs both ways. A good evening makes the whole day look fine, and a bad one darkens hours that were actually steady. If you genuinely missed an entry, leave it blank and mark it missed, which costs you one data point instead of corrupting four.

How often should I repeat an energy audit?

Quarterly is a sensible default, plus an extra run after any deliberate change you want to verify, and after anything that reshapes your weeks, such as a new job, a move, a new shift pattern or a new baby. Seasonal repeats have an extra use in places with large daylight swings, since a February week and a June week can differ enough to change which findings hold. Between audits, a single evening score costs almost nothing and gives the next full week a baseline to sit against.

Does a wearable or smartwatch replace an energy journal?

No. A wearable measures heart rate, movement and estimated sleep, and infers readiness from those; an energy journal records what you actually experienced and what you were doing at the time. The two disagree often, and the disagreements are informative, since a high readiness reading on a day you felt flat usually points at something in the content of the day rather than in the body. Use device data as an extra column beside your own scores, and when they conflict, trust your entry for how the day felt.