How to Budget

What a budget is for

A budget is a spending plan you make in advance, while you are calm, so that you are not making every financial decision alone in a shop at 6pm on an empty stomach. That is the whole idea. It is a set of decisions moved earlier, to a moment when you can think.

The wrong definition is what makes people quit, so it helps to say plainly what a budget is not. It is not a scoring system for your character, and it is not a punishment schedule you can fail. When a category runs over, what you have learned is that the plan was wrong about that category. That is information, not a verdict.

There is one more thing it is not, and this one is expensive: a budget is not a record of what you spent. That is tracking, and it is a separate job. Confusing the two is why so many attempts die. Making the plan takes an hour, once. Keeping the record takes a few minutes every day, forever, and it is the part that quietly ends most budgets somewhere in the second week.

Both parts matter, because a plan you never compare against reality is just a wish with a spreadsheet around it. But they are different problems, and only one of them needs to stay yours.

The evidence that people are willing to do the first part and not the second is fairly clear. In a YouGov survey of 2,087 British adults conducted in February 2026, 51 percent said they had a budget for the year, up from 46 percent the year before. Of those, 39 percent were running it on spreadsheets or similar manual methods against 9 percent using a budgeting app, and 36 percent were using no tool at all. Meanwhile, when Debt.com asked non-budgeters in its 2026 survey why they had not started, the leading answer was that budgeting takes too much time, at 34 percent. And having a budget is not the same as living inside one: Old Mutual’s 2025 Savings and Investment Monitor found that 69 percent of employed South African consumers overshoot their monthly budget.

The plan is not the hard part. The upkeep is.

The four methods

There are four budgeting methods, and everything else you will read about is a variation on one of them. Each one below comes with its actual origin, the math where math exists, and the situation it genuinely suits.

Zero-based budgeting

The rule: assign every unit of income a job until nothing is left unassigned. Income minus all assignments equals zero. The zero does not mean you spent everything; savings is one of the jobs.

The name comes from the corporate world, where Peter Pyhrr developed zero-base budgeting at Texas Instruments in the 1960s and wrote it up in the Harvard Business Review in late 1970. Jimmy Carter, then Governor of Georgia, brought Pyhrr in to apply it to the state budget in 1973. The corporate technique is about forcing every department to justify its whole budget from scratch each cycle instead of inheriting last year’s. The personal-finance version, popularised much later, borrowed the name for a different idea: not re-justification, but completeness. Nothing is allowed to be unallocated.

Here is the math on a monthly take-home of $3,200:

Left over: $0. That is the method working.

What it is good at: it is the only method that makes leftover money impossible, which matters if your money tends to evaporate without any single purchase you can point to. It is also the most honest about trade-offs, because every increase has to come out of a named category.

What it costs: it is the highest-maintenance method by a wide margin. It assumes a stable income you can allocate in advance, and it needs the record kept accurately or the allocations become fiction within a fortnight. If your income arrives in irregular lumps, you are re-doing the whole allocation every time money lands.

50/30/20

The rule: 50 percent of income to must-haves, 30 percent to wants, 20 percent to savings and debt payoff.

This one is worth getting right, because the version that circulates is a slightly corrupted copy. It comes from All Your Worth: The Ultimate Lifetime Money Plan, published in 2005 by Elizabeth Warren and Amelia Warren Tyagi, where it is called the Balanced Money Formula. Three details from the original are routinely lost:

It runs on after-tax income, not gross. The percentages apply to what actually reaches your account. Applying them to a gross salary produces a budget that is short by exactly the amount of your tax bill, which is a common and painful mistake.

The buckets are must-haves, wants, and savings. The popular “needs” framing is a later paraphrase. The distinction is not academic: a must-have is a commitment you cannot exit quickly, like rent, utilities, insurance, and transport to work. A cheap thing you buy every week is a want, however routine it feels.

Credit card minimums do not belong in the 50 percent. The original puts minimum payments on your obligations in must-haves but explicitly carves out credit card debt, which belongs in the 20 percent alongside savings, to be attacked rather than maintained. Most explainers lump all minimums into needs, and the result is a budget that treats a revolving balance as a fixed cost of living, which is precisely how it becomes one.

The math on a monthly take-home of £2,400:

What it is good at: speed and durability. Three numbers, no per-category upkeep, and it survives a chaotic month because it is coarse. If you have never held a budget for more than six weeks, this is the sensible place to start.

What it costs: precision, and in many places, realism. The 50 percent ceiling on must-haves was written for a US household in 2005, and housing has not stayed where it was. If your rent alone consumes 45 percent of take-home, the rule as written is arithmetically unavailable to you, a critique John Hancock lays out in detail. The fix is not to abandon it but to renumber it: 60/20/20 or 65/15/20 is still a working budget. The ratios are a starting proposal, and the 20 percent is the one worth defending.

The envelope system

The rule: convert the month’s variable spending into physical cash, split it into labelled envelopes, and when an envelope is empty, that category is finished until next month.

It has no clearly documented origin, which tells you something: it is old, folk-transmitted, and was standard household practice long before anyone wrote it down. It came back in a large way through cash-stuffing videos, which have collected billions of views.

The interesting part is that the mechanism has some evidence behind it. Paying by card appears to raise what people are willing to spend compared with paying cash. Prelec and Simester found the effect in auction experiments published in Marketing Letters in 2001, where the abstract reports it may run as high as 100 percent. A stronger and more recent test comes from Runnemark, Hedman and Xiao, whose 2015 incentivized experiment used debit cards rather than credit, so the result cannot be explained by borrowing, and controlled for how much cash people had on them. Willingness to pay was still significantly higher on card.

Read those honestly, though. Both are laboratory studies with student participants, the 2001 one predates contactless payment entirely, and neither tracked real household spending over months. What they support is a modest claim: cash makes spending feel more like something, and that friction is the whole point of the envelope.

What it is good at: it enforces itself. No app, no reconciliation, no discipline required at the moment of purchase, because an empty envelope is not an argument you can have with yourself. For a specific leaking category, it is the bluntest and most effective instrument available.

What it costs: almost everything else. It does not work for online purchases, subscriptions, or direct debits, which for most people is now the majority of spending. Carrying substantial cash carries obvious risk. And it produces no record at all, so at the end of the month you know you stayed inside the envelope and nothing else about where the money went.

A reasonable compromise: run envelopes for the two or three categories that genuinely leak, and a normal budget for everything else.

Pay yourself first

The rule: move savings out on payday, before any spending happens, and live on what remains.

The idea is a century old. It is the core lesson of The Richest Man in Babylon, assembled by George S. Clason from pamphlets he wrote in the early 1920s and published as a book in 1926, where the character Arkad puts it as a tenth of everything you earn being yours to keep.

Strictly, it is not a full budget, since it says nothing about how the remaining money gets spent. It is one rule that happens to do most of the work, and behavioural research is unusually kind to it. Madrian and Shea’s study of automatic 401(k) enrolment found that switching the default so employees were enrolled unless they opted out more than doubled participation among new hires. The same study found the catch: around three quarters of auto-enrolled employees stayed at the default 3 percent contribution rate and most changed nothing at all, so the default that gets you in also anchors you low. Thaler and Benartzi’s Save More Tomorrow programme addressed exactly that by committing people in advance to raise their contribution when their pay rose, lifting average saving rates from 3.5 percent to 13.6 percent over 40 months.

The lesson transfers directly: automate the transfer, and then separately schedule the increase, because the automation will hold your rate steady for years if you let it.

On a monthly take-home of €2,600, Clason’s tenth is €260, moved on payday. What is left, €2,340, is now the whole budget, and you can be considerably more relaxed about how it gets spent.

What it is good at: it protects the one thing every other method protects last. Savings stops being what survives the month and becomes a fixed cost like rent.

What it costs: it is silent on everything else, so if your problem is that spending is chaotic rather than that saving never happens, this alone will not fix it. And it fails badly if the transfer leaves you short mid-month and you move the money back, which teaches you the rate was wrong rather than the method.

How to pick one

The usual advice is to pick the method you can stick to, which is true and completely useless, because you cannot know that in advance.

A better question is what actually goes wrong with your money. Answer that, then pick the method built for that particular failure rather than the most rigorous one or the one your friend uses.

If money disappears without any purchase you can point to, the problem is unallocated money, and zero-based budgeting is the direct answer. It is the only method that makes “left over” impossible.

If you have never held a budget past six weeks, the problem is upkeep, not accuracy, and 50/30/20 is the answer. Three numbers survive a bad month in a way that eleven categories do not. Get your renumbered ratios from what your housing actually costs, not from the original 50.

If one or two categories leak while everything else behaves, you do not need a new budget. You need envelopes on those categories, or a hard cap on the card you use for them, and nothing else changes.

If your income is fine but nothing is ever saved, the problem is order of operations, and pay yourself first solves it on payday in about four minutes of setup.

If your income arrives in irregular lumps, none of the four works as written. Budget on your lowest realistic month rather than your average, hold everything above that in a buffer, and pay yourself a fixed amount from the buffer each month as though it were a salary. That is a large enough topic that it deserves its own treatment, and the tax estimator is the more urgent tool if some of that income is untaxed when it arrives.

One combination is worth naming because it is what most durable budgets actually look like after a year: pay yourself first on payday, coarse 50/30/20-style ratios for the rest, and envelopes or a hard cap on the one category that has always been the problem. Nobody designed that. It is what remains after the parts that were too much work fall away.

Setting the amounts

This is where budgets are usually lost, and it happens before the month even starts.

Almost everyone sets category amounts by guessing, and the guess is optimistic, because a budget is written in the same mood as a New Year’s resolution. You put £300 for food because £300 sounds like a reasonable amount for a person to spend on food. Then you spend £417, as you did last month and the month before, and the budget has now failed at something that was never achievable.

The fix is unglamorous: get the amounts from your history.

Take your last three months of bank and card statements. For each category, take the three totals and average them. That average is your honest baseline. It is what you actually spend, in your actual life, at current prices.

Groceries, over three months, in naira:

Now the part that people get backwards. Do not set the budget to ₦80,000 because you want to spend less. Set it to ₦95,000, or even ₦100,000, and hold it for one month. You are not trying to cut yet. You are trying to build a budget that matches reality closely enough to be worth looking at. A budget you meet is a tool. A budget you miss every month is just a monthly reminder that you are bad at this, and people stop opening those.

Cut in month two, from a number you have proven you can hit, by an amount you have proven is available.

Two adjustments improve the baseline considerably.

Handle the exceptional expenses separately. This is the single most useful finding in the budgeting literature. Across seven studies, Sussman and Alter found that people predict their ordinary recurring spending fairly accurately and systematically underestimate their exceptional spending: the birthday gift, the car repair, the wedding, the vet. Each one feels like a one-off, so it never gets grouped with the others, and collectively they arrive with a frequency nobody plans for. The NYU summary puts it neatly: the exception is the rule.

So count them. Go through a full year and list every expense that felt like a one-off. If there were twelve of them averaging £70, that is £840 a year, or £70 a month, and it belongs in your budget as a standing category rather than as a series of surprises. This one adjustment fixes more broken budgets than any amount of additional discipline.

Find what is repeating before you set anything. Recurring charges are the part of the baseline you are least likely to remember, because they never involve a decision. The subscription tracker totals up what you already know about, and it usually surfaces at least one you had forgotten.

When it breaks

Your budget will break at some point, and probably sooner than the method promised. The boiler goes, or the friend’s wedding lands in the same month as the car, and a budget that assumed an uneventful month turns out to have been written for a life nobody has.

What matters is what happens in the four days after the overspend, because that is when budgets are actually abandoned. Not at the overspend itself, which is just a number, but at the moment you decide there is no point looking now.

So decide the rule in advance, while nothing has gone wrong.

The rule: the overage comes from wants. Say food was set at €400 and by the 18th you have spent €310, leaving €90 for twelve days at €7.50 a day, which is not going to happen. Do not pretend it will. Move €80 out of the wants category now, today, and food is now €480 with the rest of the month funded. You have not failed. You have moved money between two categories, which is the thing budgets are for.

That single rule, decided in advance, is worth more than any amount of resolve, because it converts an emotional event into a two-minute administrative one. The reason it has to be decided in advance is that at the moment of the overspend, the honest options feel like confession or denial, and denial is much easier.

Three more things that keep a broken month from ending the budget:

Do not rebuild from scratch mid-month. Adjust two numbers and continue. Starting over on the 1st is how a bad week becomes a lost quarter, because “I will restart next month” means eleven days of not tracking, which means a backlog, which means you never restart.

Do not backfill the whole gap. If you have not logged anything for two weeks, do not sit down to reconstruct it all from statements. Log today, and let the two weeks go. A budget with a hole in it that is running is worth more than a complete budget that stopped.

Treat a category that breaks three months running as a wrong number, not a weak will. Three consecutive overspends is a measurement telling you the baseline was optimistic. Raise it and take the difference from somewhere you actually can.

Reviewing it

The standard advice is a weekly budget review, and it is quietly responsible for a lot of quit budgets. A Sunday evening ritual of collecting receipts, categorising transactions, and reconciling totals is an hour of unpaid administration a week, which is 52 hours a year, and it is the piece nobody keeps past March. It is also the exact thing 34 percent of non-budgeters in the Debt.com survey were pointing at when they said budgeting takes too much time.

The review itself is not the problem. The review is genuinely useful, and it takes about four minutes. The problem is the data entry that has to happen before the review can start.

Which means the useful question is not how to be more disciplined about Sunday. It is how to arrive at Sunday with the record already made.

Three ways to get there, in increasing order of how much they help:

Log at the moment of spending, not later. Four seconds at the till beats forty minutes on Sunday, because you are never reconstructing anything and there is never a backlog to face.

Photograph receipts instead of keeping them. The receipt scanner reads the total off a photograph, which turns a pile of paper into a number without any typing.

Hand the record-keeping over entirely. This is what Auritrack is built for, and it is why the earlier distinction between the plan and the record matters so much. You describe what happened in the language you would use with a person, and the AI reads the amount, the vendor, and the date, then files it under the right category. Spent ₦4,500 on fuel yesterday. Paid £12 for lunch near work. It understands “last Friday” and “three days ago”, so logging late costs you nothing, and several transactions can go in one message on the walk home. When it files something wrong, and occasionally it will, you correct it in another sentence.

Then the review runs in the other direction. Instead of building a summary, you ask for one: Am I within my budget this month? or How much did I spend on food? and you get an answer rather than an assignment. Budgets in the app show progress per category and warn you as you approach a limit, which means the mid-month overspend from the previous section announces itself while there is still time to move the €80, rather than after.

Two things worth being straight about. There are no live bank connections yet, so getting existing history in means uploading a statement as PDF, CSV, or Excel, which the AI parses into transactions you review before anything saves. And the free tier covers manual tracking, budgets, and storage; the AI features, which are the part that removes the labour, run on pay-as-you-go Auricoins that do not expire. If you want the mechanics before deciding, the AI bookkeeping page walks through what happens to a single transaction, and tracking expenses without spreadsheets covers the same ground for anyone whose last system was a sheet that stopped in week two.

However you arrive at it, here is the whole review, once a month, four minutes:

1. Which categories went over, and by how much?
2. Was that a one-off, or the third month running? (Third month running means change the number.)
3. Did the savings transfer actually happen?
4. Is there anything recurring that I am no longer using?

That is the entire review. You are not reconciling anything, and you are not doing it on a Sunday.

Three months from now

It is the 22nd, which used to be a particular kind of day.

You check, because checking is now a thing you do rather than a thing you avoid. Food is at £340 of £400 with nine days left, which is tight but not a problem, and the £60 to cover it is already spoken for by the rule you decided back in the first week. Two categories are comfortably under. The savings transfer went out on payday, as it has for three months, and you did not have to decide anything about it.

You have not reconciled anything. You did not sit down on Sunday. What you did instead was say what you spent, roughly when you spent it, in about four seconds each time, and on the days you forgot, you said it two days later and nothing was lost.

The change is not that you spend less, though you probably do, mostly on things you turned out not to want. The change is that the 22nd is no longer a day you have a feeling about. You know the numbers, so there is nothing left for the dread to attach to.

Where to start

Not on the 1st, and not with a system. Today, with one of these.

One. Open the free budget planner. No account, runs in the browser. Put in your take-home pay and get a draft split you can drag until it looks like your actual life. Ten minutes, and you will leave with three numbers rather than a resolution.

Two. Find what is already repeating. The subscription tracker totals up the recurring charges you know about, and the exercise usually surfaces one you had genuinely forgotten. Cancelling it funds most of a savings transfer on its own.

Three. If debt is the part that makes the budget feel pointless, work out the actual payoff date first with the debt payoff planner. A date is easier to budget toward than an amount. If it is a savings target instead, the savings goal calculator does the same job from the other direction.

Four. If the draft budget looks livable, the piece that has always broken is the record-keeping, and that is the part to hand over. Create an account on the web app, or get it on Google Play or the App Store, and log exactly one expense by typing it like a text message. One expense. Not a month, not a migration, not a fresh start on Monday.

The budget was never the hard part. You proved that the last three times you made one.