Money Clarity

    How to track my money: record automatically, review monthly

    Most people who search how to track my money have tried before, stopped somewhere around week three, and filed the failure under discipline. The arithmetic points somewhere else. In the illustrative household on this page, logging each payment as it happens means 101 entries a month at 25 seconds each: 42 minutes of effort, split into 101 interruptions, every one of them arriving at the moment you are busy paying for something. What the effort buys arrives once, at month end, and only if the month is still complete by then.

    Tracking is two jobs that most methods bundle into one. Recording is the list of what left the account, and it should cost you nothing: taken automatically from the bank's own records, or from a statement once a month. Reviewing is where decisions come from, and it should be short, on a fixed date, and ask three questions only: what repeated, what was new, and what was bigger than last month.

    Below: one household's month, four ways to record it priced in minutes and coverage, a simple model of which one lasts three months, and a 20-minute review template you can run on your own statement this weekend.

    Last reviewed 2026-09-28

    Why tracking stops in week three

    The technique

    Effort at the till, payoff on day 30

    People read a lapse as a flaw of character. It is a pricing problem. A method that charges 25 seconds at every payment and pays out once a month asks for about a hundred small sacrifices per reward, and the payments that most need recording tend to happen when you are least able to stop and type.

    Take a renting household with ₹92,000 of take-home pay, a car loan, a SIP, a term insurance policy, a credit card from a bank other than the salary bank, and the usual handful of mandates. Its month runs to 98 debits and ₹84,495, leaving ₹7,505. The figures are illustrative; the working carries over to one month of your own statement.

    The last column is the one that matters for tracking. Logging at the moment of payment assumes there is a moment. For seven of these debits there is none: the EMI, the SIP, the premium and four AutoPay mandates leave the account on their own schedule while you are asleep or at work, and nothing prompts you to write them down.

    Kind of debitDebitsRupeesShare of rupeesA moment of paying?
    Everyday UPI72₹15,12017.9%Yes
    Card swipes12₹14,40017.0%Yes
    Cash (two ATM withdrawals)2₹3,0003.6%Yes, 12 small payments
    NACH: EMI, SIP, premium3₹15,85018.8%No
    UPI AutoPay mandates4₹1,0761.3%No
    Transfers: rent, parents, house help3₹33,00039.1%Yes, once each
    Bills paid by UPI2₹2,0492.4%Yes
    Total98₹84,495
    Illustrative. EMI ₹9,400, SIP ₹5,000, premium ₹1,450. AutoPay ₹199, ₹149, ₹599 and ₹129. Rent ₹22,000, parents ₹8,000, house help ₹3,000. Electricity ₹1,650, mobile ₹399. The card row is this month's swipes; the ₹11,900 card bill paid this month settles last month's swipes and is left out so the card is not counted twice.
    • The seven debits with no moment of paying come to ₹16,926. A method built on noticing payments misses them by design, and they are exactly the ones that keep charging after you stop using what they pay for
    • A missed day costs more than a day. It leaves a hole in the month, a month with a hole feels not worth reviewing, and so the habit ends at the first busy week rather than fading slowly

    Two jobs: recording and reviewing

    The technique

    Recording costs nothing; reviewing costs 20 minutes

    Most methods spend all the effort on recording, where it produces a list, and none on reviewing, where it would produce a decision. Reversing that allocation is most of the fix, and it needs no extra willpower.

    Recording has one standard: every rupee that left, itemised, with no act required from you on the day. Two routes meet it. The first is automatic, from records the bank already produces: its SMS and email alerts, the history in its own app, or account data shared by your consent. The second is the statement from every account and card, downloaded once a month. Which automatic source sees which kind of payment is set out on the page about automatic expense trackers; the point here is that neither route asks anything of you at the till.

    Reviewing has a different standard: it should be short enough to never skip, on a fixed date, and made of questions that end in an action. What repeated points at things you can cancel or renegotiate. What was new points at trials that converted, payees you do not recognise and anything fraudulent. What was bigger than last month points at drift, the slow creep that no single payment reveals. A category chart answers none of the three; it tells you food is a large slice, which you knew.

    • If a method asks you to open it every day, recording has leaked into your time. Done properly, recording produces nothing you need to look at until review day
    • A review with no comparison month has nothing to say. Keep last month's three answers and read this month's against them

    How to track my money: four methods, costed

    Here are the four common ways to track expenses, applied to the same 98 debits. Every method gets the same 20-minute review, so the difference is entirely in recording.

    The manual log, typed at the moment of paying, needs 101 entries: the 72 UPI payments, 12 swipes, 12 cash payments, three transfers and two bills. At 25 seconds each that is 42 minutes, 62.1 with the review. It is the only method that itemises cash, but it sees ₹67,569, 80.0 percent of the month, because the NACH and AutoPay debits never pass through your hands.

    The weekly catch-up, copying rows from the salary account's app into a sheet, costs 4.33 sessions a month at four minutes of setup each, 17.3 minutes, plus 87 rows at 12 seconds, 17.4 minutes: 54.7 in all. It sees every NACH and AutoPay debit, but the card from another bank appears only as last month's ₹11,900 bill, so the ₹14,400 of this month's swipes goes unitemised. Coverage: ₹70,095, 83.0 percent.

    Monthly statements, one from the account and one from the card, take two downloads at four minutes each and 98 rows at ten seconds to label: 24.3 minutes, 44.3 with the review. All ₹84,495 is there. The cost is lateness: a charge you find in a statement has already gone.

    Automatic recording from the bank's records costs no entries at all. What remains is naming the nine rows the source could not name, at 30 seconds each, and a minute noting what the cash bought: 5.5 minutes, 25.5 with the review, all ₹84,495 seen on the day it leaves.

    MethodRecordingReviewMinutes a monthItemised coverage
    Manual log at payment42.12062.1₹67,569 (80.0%)
    Weekly bank-app catch-up34.72054.7₹70,095 (83.0%)
    Monthly statements24.32044.3₹84,495, all
    Automatic from bank records5.52025.5₹84,495, all
    Assumed timings: 25 seconds per manual entry; 4 minutes of setup per weekly session and 12 seconds per row copied; 4 minutes per statement download including the password; 10 seconds per row labelled; 9 unnamed rows at 30 seconds. Cash counts as itemised only in the manual log; every other method sees the ₹3,000 withdrawn.
    • Over a year the manual log costs 745 minutes and automatic recording 306, a difference of 7.3 hours. And 240 of the 306 are review, the only minutes in either method that produce a decision
    • The weekly catch-up feels like a sensible compromise and costs almost as much as daily logging, 54.7 minutes against 62.1, while seeing less of the card
    • Monthly statements are automatic recording without an app: full coverage, one sitting. If you track spending this way, download on a fixed date, because a statement you meant to download is a month with no record

    Which tracking method lasts three months

    The technique

    Count the asks, not the minutes

    People choose a method by how little time it takes. Whether it lasts depends on how many separate times it asks something of you, because each ask is a chance to skip and not start again.

    There is no dependable survey figure for how long tracking habits last, so here is a model you can argue with. Give every ask a chance of being the one that ends the habit: 1 percent for a 25-second entry, 5 percent for a weekly session or a 20-minute review, 10 percent for a statement sitting of 44 minutes. These are assumptions, deliberately kind to the manual log; 1 percent means only one entry in a hundred is the one you skip and never come back from.

    The chance of surviving is the chance of getting past every ask. Over three months the manual log asks 303 times for entries and three times for a review. The weekly catch-up asks 16 times, and the other two methods ask three times each.

    MethodAsks in three monthsChance per askSurvives three months
    Manual log at payment303 entries + 3 reviews1% and 5%4.1%
    Weekly bank-app catch-up13 sessions + 3 reviews5%44.0%
    Monthly statements3 sittings10%72.9%
    Automatic from bank records3 reviews5%85.7%
    Illustrative model, not measured rates. Survival is (1 minus the chance per ask) multiplied once for every ask; replace the chances with your own guesses and recompute.
    • Halve the manual risk to 0.5 percent per entry and three-month survival rises only to 18.8 percent. The trouble is the 303, not the 1 percent
    • For the manual log to last as often as monthly statements, you could skip no more than one entry in 1,869. That is a standard nobody's month meets
    • The ranking holds for any sensible inputs. A method that asks three times a quarter outlasts one that asks three hundred times, even when each of its asks is bigger

    A 20-minute monthly review template

    The technique

    Three questions, each ending in an action

    A review that starts with a chart ends with an observation. Asking what repeated, what was new and what grew points at particular lines, and a particular line can be cancelled, queried or watched.

    Same date every month, the weekend after salary, with last month's answers open. Five steps, timed.

    Two minutes on the total: ₹92,000 in, ₹84,495 out, ₹7,505 left. If the account balance disagrees by more than a few hundred rupees, the record is missing something. Find it before anything else.

    Six minutes on what repeated: everything that also appeared last month, or three or more times this month. Here that is the committed block of EMI, SIP, premium, mandates, transfers and bills, ₹51,975 or 56.5 percent of take-home, plus three repeat merchants: 11 quick-commerce orders averaging ₹320, 14 cab rides averaging ₹190 and 9 café visits at ₹160, together ₹7,620 in 34 payments. For each line, one question: would I set this up again today? The four mandates alone are ₹12,912 a year.

    Five minutes on what was new: a ₹129 mandate from a trial that converted, ₹2,400 to a payee you do not recognise at first glance, and a ₹1,899 annual charge on the card. Every unfamiliar payee gets named or queried with the bank; new is where forgotten trials and fraud both surface.

    Five minutes on what was bigger: flag any repeated line up by 20 percent and by ₹500. Quick-commerce went from ₹2,240 to ₹3,520, up 57.1 percent. Card swipes went from ₹11,900 to ₹14,400, up 21.0 percent. Electricity rose 39.8 percent, but by ₹470, under the rupee bar, and it is seasonal.

    Two minutes on one decision, written down: cancel the converted trial, and watch quick-commerce next month.

    This month's review sheet
    Out against in
    ₹84,495 of ₹92,000
    Repeated: committed block
    ₹51,975
    Repeated: three merchants, 34 payments
    ₹7,620
    New: three items
    ₹4,428
    Bigger and flagged: quick-commerce
    +₹1,280
    Bigger and flagged: card swipes
    +₹2,500
    Decision: cancel the ₹129 mandate
    ₹1,548 a year

    Flag rule: up 20 percent and up ₹500 on last month. Cabs rose ₹110 and café spending fell ₹160, so neither is flagged.

    • Both thresholds matter. Twenty percent alone flags a ₹40 rise on a small line; ₹500 alone flags ordinary noise on the rent. Together they keep the list to two or three lines
    • If the quick-commerce rise holds, ₹1,280 a month is ₹15,360 a year. One month is a data point; the review exists so that the second month is noticed too

    Money tracking mistakes that end the habit

    Setting a budget before recording a complete month. Limits drawn from memory are limits drawn from the payments you remember, which are the large ones. Record one full month first; the budget then starts from what happened.

    Checking daily. Two minutes a day is 60 minutes a month, more than the whole automatic method with its review, and a single day has nothing to compare against. Daily checking feels like diligence and produces anxiety rather than decisions.

    Backfilling a lapse. Ten days off the manual log leaves 33.7 entries to rebuild from memory, and memory keeps the ₹1,899 charges and drops the ₹120 ones. When a month has a hole, do not reconstruct it; take that month from the statement instead.

    Categorising every row. Labelling 98 debits is 98 small decisions, and the three review questions need names, not categories. Most of the review happens on a dozen lines.

    Counting the card twice. Keep this month's swipes and drop the bill payment, the ₹11,900 row here, or the month looks ₹11,900 more expensive than it was.

    • Each of these mistakes adds effort to recording, which is the half that should cost nothing. When a method starts to feel like work, look for which of them has crept in before blaming yourself

    When tracking spending will not fix the month

    Tracking shows where money goes. It does not change what is already committed, and in this household ₹51,975, 56.5 percent of take-home, is committed before any choice is made. The chosen spending, UPI, card and cash, is ₹32,520, or 35.3 percent. Cut it by a tenth and the month gains ₹3,252. If the shortfall is larger than that, tracking has done its job by showing so, and the fix lies in the rent, the loan or the income. How committed debits stack up across the days of the month is traced on the page about where a salary goes.

    If most of your money leaves as cash, no automatic method sees what it bought. A notebook for the larger cash payments does more than any app.

    If you will not do the review, do not bother with the recording. An automatic record nobody reads costs nothing and changes nothing; the 20 minutes are the part that pays.

    And if the month already ends with money left and nothing in the three questions surprises you two months running, the review can move to once a quarter. Tracking is a tool for finding decisions, not a duty.

    • A month that is short by more than a tenth of chosen spending is a structural problem. Tracking will describe it accurately and leave it exactly where it was

    How Unyfy helps with tracking your money

    Unyfy does the recording half so that it costs nothing. It reads your bank and card transaction emails and, on Android, your bank's transactional SMS, so UPI payments, card swipes, NACH debits and AutoPay charges arrive without an entry from you, and a debit seen by both SMS and email is counted once. Each UPI handle is matched against a merchant database of about 10,000 entries, so everyday payments show as merchant names rather than phone numbers. It also parses statement PDFs from Axis, HDFC, ICICI, Kotak and Federal Bank, which covers the once-a-month statement route as well.

    For the review it prepares two of the three questions. On Pro, the Fixed Expenses screen forecasts what the coming month is already committed to in EMIs, SIPs, rent, bills, subscriptions and the card bill, with what is paid and what is left so far this month, and its Subscriptions list shows each recurring subscription with its amount and whether it is due or paid. What was bigger than last month, and the decision at the end, stay with you.

    It never asks for your bank password or UPI PIN, and every payment is one you authorise. Install Unyfy on Android, or use the web app at app.unyfy.co.in on an iPhone.

    Common questions

    How to track my money if I have given up before?

    Stop recording by hand and keep only a monthly review. Logging each payment asks for effort about a hundred times a month and pays back once, which is why it stops around week three. Take the record from the bank's alerts or a monthly statement, then spend 20 minutes on a fixed date asking what repeated, what was new and what was bigger than last month.

    How to track expenses without logging every payment?

    Use a record the bank already makes. Its SMS and email alerts, its app history and its monthly statement all list every debit without any typing. In the worked household, automatic recording costs 5.5 minutes a month for naming a few unclear rows, against 42 minutes for logging 101 payments, and it also sees the ₹16,926 of EMI, SIP and mandate debits that manual logging misses.

    How often should I review my spending?

    Once a month, on a fixed date, for about 20 minutes. A weekly check costs almost as much as daily logging for little extra insight, and a daily one has nothing to compare against. If two reviews in a row find nothing new and nothing bigger, once a quarter is enough until something changes.

    How do I keep track of money I spend in cash?

    Every automatic method sees the ATM withdrawal, not what the cash bought. If cash is a small share of your month, 3.6 percent in the worked household, a minute's note at review time is enough. If most of your spending is cash, a notebook for the larger payments is more useful than any app.

    Is a monthly bank statement enough to track spending?

    Yes, if you download one from every account and card. Statements carry every debit, so coverage is complete; the cost is about 24 minutes of labelling and lateness, because a charge found in a statement has already gone. UPI rows often show only a handle, so some payees need a moment of recall.

    Does money tracking need my bank password or UPI PIN?

    No. Tracking reads what has already happened, from alerts, emails, statements or data shared by your consent, and none of those needs a login or a PIN. A PIN authorises payments; a tracker asking for it wants something tracking does not need. Unyfy holds no bank password or UPI PIN and cannot move money.

    Tracking fails at the till, not in the character. Take recording out of your day, automatically from the bank's records or from a statement once a month, and give the review 20 minutes on a fixed date with three questions. In the worked household that is 25.5 minutes a month instead of 62.1, every rupee seen instead of 80.0 percent, and a method that has to survive three asks a quarter instead of 306. Informational page, not financial advice. The household, the timings and the survival chances are illustrative assumptions; your own statements govern what left your account, not this page.

    Our Partners

    Banks and NBFCs we compare

    Unyfy compares offers from these lenders and earns a commission if you take one. The comparison is shown first, and it can tell you not to switch.

    HDFC Bank logo
    ICICI Bank logo
    Axis Bank
    State Bank of India logo
    IDFC First Bank logo
    Kotak Mahindra logo
    IndusInd Bank logo
    Yes Bank logo
    Bajaj Finserv logo
    Tata Capital logo