How to Make a Monthly Budget That Sticks in 2026 (6 Steps)

Most monthly budgets are written on the first and abandoned around the twelfth. The plan was usually fine. It was just built on numbers somebody guessed, checked against nothing, and then broke on the first unplanned dinner with friends who suggested a second bottle.

Build it the other way round. Last month's real spending, one number to watch, and a decision made in advance about what happens when you go over. Six steps, and a worked example at the end.

Why budgets built from guesses fail in the second week

Ask someone what they spend on groceries. You will get a figure. Ask their expense log and you will get a different one, usually 20 to 40 percent higher.

That gap is not dishonesty. Memory rounds down, drops the small stuff and forgets the second trip to the shop on Thursday, which means a budget set on the remembered number is already over by the middle of the month, at which point the obvious conclusion is that budgeting does not work for you personally.

It does. It just cannot start from recall. The Bureau of Labor Statistics runs a national survey that asks households to keep a written diary of every purchase for two weeks, and they do that precisely because asking people what they spent produces fiction. Your budget needs the same foundation, and on an iPhone it costs a few seconds per purchase.

Step 1: start from last month's real numbers

Before setting a single limit, track every expense for one month. Not most. All of them, including the parking meter, the birthday card and the app you forgot renews every September.

The only way anyone survives a full month of this is if logging is faster than remembering to log. Four ways to keep it cheap: say what you spent into the mic, snap the receipt, let Apple Pay taps log themselves, or take three taps on the keypad. Which one hardly matters. Using one every single time does.

At the end of the month, open Insights and read spending by tag. Write the totals down before you judge any of them. Judging comes later, and it goes better with the numbers already on paper.

Step 2: use 50/30/20 as a starting point, not a law

The 50/30/20 rule puts roughly half of take-home pay on needs, 30 percent on wants and 20 percent on savings or debt repayment above the minimums. Good reference line. Poor rule. Your needs share is set by your rent and your city, not by a ratio someone published.

Sort last month's tags into the three piles. Needs: rent, utilities, groceries, transport to work, insurance, minimum debt payments, anything that cannot wait. Wants: everything optional. Savings: what was left, or what should have been. Then compare, and let the 50/30/20 budget calculator do the arithmetic on your income.

Needs at 65 percent is not a failure. It is a number. Set wants and savings out of the 35 percent that is actually there, and treat the needs side as a separate, slower project, because those costs take months to move rather than weeks. Fixed and variable expenses is the distinction doing the work here: fixed costs are a project, variable costs are a habit, and a monthly budget manages the habit.

Step 3: give the flexible part one allowance

This is where most guides go wrong. They tell you to set a limit for every category. $120 eating out, $60 coffee, $80 clothes, $40 books, $150 going out, on down fifteen lines, each one a decision to make and a number to hold in your head, and by week three nobody remembers a single one of them.

Give all your discretionary spending one figure instead. Say $560. That is the number. Dinner, a jacket, a concert, a present for your sister, all of it comes out of the same $560 and you pick the mix as you go.

One number you can actually know at any moment. Knowing it is the part that changes behaviour, not the categorising.

In Cash Book that is a monthly budget sitting under the hero figure on Home, reading "$310 left" or "over". You see it every time you log anything, which is the entire design. Weekly suits some people better, and the same allowance can be set that way; the daily spending allowance calculator cuts a monthly number down to a daily one if you want an even smaller unit to think in.

Step 4: per-tag limits only where money leaks

One allowance covers most people most of the time. But there are usually two or three tags that behave badly, run over every month, and respond well to a visible ceiling. Eating out and subscriptions, generally. Sometimes online shopping. For one person I know it is taxis, exclusively after 11pm.

Pick the two or three tag totals that surprised you last month and put a budget on those alone. Cash Book draws each as a thin progress bar under the tag, and it stays quiet until you cross the line, when it turns red. Everything else stays inside the allowance without a fence of its own.

Two things make this work. The tags have to exist as one tag, so if eating out is scattered across "food", "lunch" and "restaurants" then no limit can possibly hold, and half an hour tidying tags and categories is the prerequisite. And the limit has to sit a little under last month rather than half of it. $310 down to $220 is a habit change. $310 down to $100 is a resolution, and resolutions do not survive contact with a Friday.

Step 5: one number, once a week

A budget nobody looks at is a document. A budget checked daily becomes a chore, and chores get dropped by February. Weekly, same day, one number: what is left of the allowance.

That is the whole review.

Quarter of the way through the month with three quarters of the allowance left? Fine. Halfway through with a quarter left? The next two weeks are quieter, and the useful part is that you know it now rather than on the 28th when it is a fact rather than a choice.

Once a month spend five more minutes in Insights, where the change against last month is broken out by tag. That is where you catch subscriptions drifting up by $14, or groceries falling after you started planning meals on Sundays. Weekly for the number, monthly for the decisions.

Step 6: what to do when you go over

You will go over. Car repair, a wedding, a week where everything went sideways at once. Whether the budget survives depends entirely on what happens next, so decide the policy now, while nothing is on fire.

When the allowance runs out early, the rest of the month gets smaller, not cancelled. Count the weeks left, divide whatever is left between them, hold that line. If nothing is left, the remaining weeks are needs-only, which is grim for two weeks and genuinely damaging for two months, so keep it to two weeks.

When a real one-off lands, move money from the savings share once, deliberately, and log it as a transfer so it is visible afterwards. The rule is that it has to be a decision you can point at later, not a quiet drift. Three months running, though, and it is not a one-off. It is a need, or it belongs in a sinking fund collecting a twelfth of itself every month.

When one tag keeps blowing its limit, the limit is probably wrong. Either raise it to something honest and shrink the allowance to match, or go at the habit directly with the guide to impulse buying. Both are respectable. Insisting the limit is $100 while the spending is $250 is the only option that reliably fails.

And when the whole month goes? Do not start over. The month is data.

A worked example on $3,200 net income

Take-home pay of $3,200 a month, and a full month of tracking produced the middle column. The 50/30/20 reference line would be $1,600 needs, $960 wants, $640 savings. Real life arrived differently, as it does, so the plan in the third column starts from what happened and walks toward the split instead of leaping at it.

Needs are 67.5 percent rather than 50, because rent is what rent is, so the plan trims groceries by $30 and leaves the rest of that side alone. The flexible half becomes one $560 allowance with two limits inside it: eating out from $310 to $220, which is a cut you can feel but not one that cancels your social life, and subscriptions down $14 after an audit found a service nobody had opened since spring.

Then there is that $71 marked unaccounted. Cash, almost certainly, withdrawn and dispersed into a month with no record of where. It is also the single best argument for giving cash its own account in the app.

Savings more than doubles here, and nothing fixed was touched to do it. That is the part worth sitting with: the money was already there, it was just invisible.

Month three will adjust all of this again. Normal.

What this is based on

These steps follow public budgeting guidance from the Consumer Financial Protection Bureau and MoneyHelper, both of which start from tracked income and spending before any limit gets set. The gap between remembered and recorded spending is the reason the Bureau of Labor Statistics' Consumer Expenditure Survey collects a purchase diary instead of trusting recall. And the Federal Reserve's annual SHED report documents how thin the margin is for a great many households, which is why the savings share here is protected rather than treated as whatever happens to be left.

Get Cash Book on the App StoreFree for 7 days. $19.99 once for lifetime.

Get Cash Book on the App StoreFree for 7 days. $19.99 once for lifetime.