How Much to Set Aside for Taxes From Every Invoice

The intuitive approach is to work out what you owe when a payment is due, and pay it. That fails in a specific, predictable way.

Money in the operating account looks spendable. Not through irresponsibility — through arithmetic. You look at the balance, see it covers the equipment you need, and buy it. The balance was never really that size, but nothing in the account told you so.

Income is lumpy and bills are not. A good month feels like a good month. It is only later, aggregated with the other good months, that it becomes a larger tax liability than you planned for.

Deadlines cluster badly. Payment dates often land near other obligations, and in some systems a payment on account for next year arrives alongside the balance for last year. Meeting that from current cash flow requires either a great month or a loan.

Recovering is expensive. Once you are behind, you are paying last year's tax out of this year's income while also accruing this year's. That hole takes real time to climb out of, sometimes with interest and penalties attached.

Per-invoice set-aside removes the failure mode instead of asking you to be disciplined about it. The money is not there to spend, so you do not spend it. That is a systems fix, not a willpower fix — the same reason the ten-minute Sunday routine works better than intending to review your finances.

You need one number: the share of each payment that moves to the tax account. Getting it approximately right matters much more than getting it exactly right.

Start from your effective rate, not a headline rate. Marginal rates apply to the top slice of income, not all of it. What you want is total expected liability divided by total expected income. In most systems that effective rate is meaningfully lower than the top bracket — and in many it includes social or self-employment contributions that people forget entirely.

Include everything that gets paid to the state. Income tax is often not the whole picture. Social contributions, health contributions, and self-employment levies can rival or exceed the income tax component depending on where you are. If you are registered for a consumption tax like VAT or GST, that money is collected on behalf of the state as well and needs its own set-aside — it is not income at any point.

Base it on profit, not revenue, then apply it to revenue. Tax is generally on profit, but you receive revenue. Estimate liability from expected profit, then express it as a percentage of expected revenue so you can apply it to each payment as it lands.

Add a margin. Whatever the estimate says, a few extra percentage points cost you nothing — the surplus stays yours — and cover the year being better than forecast, an expense you cannot deduct, or a rule you misread.

Get the estimate from a real source. Most tax authorities publish calculators. Most accountants will give you a working percentage in one short conversation, and that conversation is cheap compared to being wrong for twelve months.

Write the number down along with the assumptions behind it. In eight months you will want to know whether it assumed your equipment purchase.

Numbers here are illustrative — they demonstrate the arithmetic, not a rate that applies to you.

Suppose you expect €80,000 of revenue and €20,000 of deductible business costs, so €60,000 of profit. Suppose your accountant estimates total liability — income tax plus contributions — at €18,000.

As a share of revenue: 18,000 ÷ 80,000 = 22.5%. Round up to 25% for margin.

Now every payment gets the same treatment:

The transfer happens the same day the payment lands. No monthly calculation, no decision to make.

Two things this makes visible. First, your real income per invoice: a €6,500 payment is €4,875 of business money, and after business costs and your buffer, less again. Pricing conversations get more honest once you see that consistently.

Second, the surplus. If liability turns out to be €18,000 and you set aside €20,000, you finish with €2,000 unallocated. That is not wasted — it becomes buffer, a pension contribution, or next year's opening balance.

If you are registered for VAT or an equivalent, handle it as a separate set-aside at the applicable rate. Mixing it into one percentage makes both numbers unreadable, and consumption tax was never yours to begin with.

The account matters almost as much as the percentage.

Separate from operating money. This is the whole point. A "reserved" portion of your main balance is not reserved. It is a note to yourself, and notes lose to a balance you can see.

Accessible on your payment dates, not before. Some friction is good; a fixed-term product that matures after your deadline is bad friction. Instant-access savings usually strikes the right balance.

Earning something if it can. Money sitting for months is worth putting somewhere with interest, especially the portion covering a payment far out. Do not chase yield with money you owe — capital safety and timing beat a slightly better rate. If you want to see what the difference actually amounts to over a year, the compound interest calculator will tell you quickly.

In its own currency where relevant. If you owe tax in one currency and earn in another, holding the set-aside in the currency you will pay avoids an unpleasant exchange-rate surprise near the deadline.

Visible in your books. The transfer should show up as a transfer between accounts you track, not vanish from view. If your business lives in its own workspace, the tax account belongs there too — it is business money, not personal savings, and mixing it distorts both pictures.

Practical detail: name the account something unambiguous. "Savings 2" invites raiding. "TAX — DO NOT SPEND" is harder to argue with at 11pm.

A percentage set in January and never revisited will be wrong by December, because your year will not match your forecast.

Review it quarterly — often enough to catch drift, rarely enough to avoid churn. Four questions:

Is revenue tracking the forecast? Substantially ahead may push you into higher brackets, meaning your effective rate rises and the percentage should too. Substantially behind may mean you are over-reserving, which is comfortable but ties up cash.

Have costs changed? A major equipment purchase or a new large recurring cost changes profit and therefore liability. So does losing a cost you had budgeted.

Have the rules changed? Rates, thresholds, and allowances get adjusted, sometimes mid-year. This is the one to ask your accountant about rather than researching yourself.

Is the balance on track? Compare what is in the tax account against your latest liability estimate for the year so far. If it is short, raise the percentage now rather than making a large catch-up transfer later.

Adjust in small steps. Going from 25% to 27% is manageable. Discovering in November that you need 40% for the rest of the year is not.

A note for genuinely volatile income: the percentage method handles volatility well precisely because it scales automatically. A quiet quarter sets aside less because you earned less. That is correct behaviour, and it is why per-invoice beats fixed monthly transfers for irregular earners — the same reasoning behind holding a business buffer sized with the emergency fund calculator.

The ones that show up most often, roughly in order of cost.

Forgetting social or self-employment contributions. People plan for income tax and get blindsided by contributions that can be comparable in size. This is the single most common expensive surprise.

Treating consumption tax as revenue. VAT or GST collected from clients is not income. Spending it and finding out at filing is a bad month.

Not accounting for payments on account. In some systems your first bill includes both last year's balance and an advance toward next year — potentially far more than you reserved for.

Setting aside from profit but transferring from revenue, or vice versa. Mixing the bases produces a number that is wrong in a direction you will not notice for months.

Borrowing from the tax account "temporarily." It is never temporary. If you genuinely must, write down the amount and the repayment date, and treat it as debt.

Skipping the transfer when money is tight. Understandable, and it converts a cash flow problem into a tax problem, which is a worse category of problem.

Assuming last year's percentage still applies. Your income, costs, and the rules all move.

Doing none of this because the rules are confusing. Setting aside a rough 25% imperfectly beats setting aside nothing perfectly. Start with an approximation and refine it with an accountant.

A habit that depends on remembering will fail eventually. Reduce it to something mechanical.

Attach it to a trigger, not a date. The trigger is "a client payment landed," not "it is the 1st." Payments do not arrive on schedule; the response should still be immediate.

Use standing transfers where the income is predictable. For retainer clients paying the same amount monthly, a standing transfer for the set-aside portion removes the decision entirely.

Put it in your weekly check. Minute nine of the ten-minute Sunday routine exists for this. Any payment received since last week gets its set-aside moved before anything else is considered available.

Track the target as a goal. Watching the tax account fill toward your estimated liability turns an abstract obligation into something with visible progress — and tells you at a glance whether your percentage is holding.

Keep it findable on your phone. The moment a payment notification arrives is the ideal moment to make the transfer. The iOS and Android apps make that a thirty-second action instead of a task for later.

Do the draw after the set-aside, never before. Tax money leaves first, business costs second, your personal draw third — the order described in separating personal and business expenses.

None of this is difficult. It is one transfer, triggered by one event, into one clearly named account. The reason it is worth writing eight sections about is that the people who do it find tax season boring, and boring is the correct thing for tax season to be.