Do You Need an Accountant or an Accounting System?

This question usually arrives at the same moment. You have raised something, or revenue has become real enough that the shoebox of receipts is no longer funny, and somebody has told you that you need to sort your finances out.

So you ask around, and you get two answers depending on who you ask. Other founders say get software. Your corporate secretary says use their accounting service. Both are partly right and both are answering a slightly different question than the one you asked.

Start with what is legally required

In most of Southeast Asia the statutory obligations are not optional and not automatable.

A Singapore company files an annual return with ACRA and a tax computation with IRAS, and needs a corporate secretary within six months of incorporation. A Malaysian company files with SSM and LHDN, and SST registration bites above the threshold. A Philippine company deals with BIR on a monthly and quarterly rhythm that is genuinely onerous. An Indonesian PT PMA must keep its records in Bahasa Indonesia and in rupiah, with supporting documents retained for ten years.

None of that is a software decision. You need a person with a licence, and trying to save money here is a well-documented way to spend far more later. Verify the current requirements for your own entity, because they change.

So the answer to half the question is settled before you start. You need an accountant. The real question is what else you need.

What only a person can do

Judgment on ambiguous transactions. Is this a capital expense or an operating one. Does this contract create deferred revenue. Should this founder loan sit in equity or liabilities. Automation is confidently wrong on exactly these, and confidently wrong is worse than slow.

Anything with a signature and a licence attached. Statutory filings, tax positions, audit support.

Absorbing your context. A good accountant knows that your December spike is one annual customer and not a trend. No system knows that unless someone tells it.

Being accountable. When a filing is wrong, a person can be held responsible. Software cannot.

What only a system can do

Be current. A person closes your books once a month. A system with bank feeds is roughly current every day. That difference is the whole game when you are deciding something on the eleventh.

Remember perfectly. Vendor memory that books the same supplier the same way every single time, without anyone re-deciding, is something humans are measurably worse at than software. Consistency is where manual bookkeeping actually fails, more than accuracy.

Work at 2am on a Sunday. When you want to know your cash position before a call, you do not want to email anyone.

Connect backwards to forwards. This is the one people miss. A person produces last month. A system can hold last month and next year in the same structure, so your actuals land against your plan without anybody copying numbers between two files.

The order that wastes least money

Most founders buy in this order: accountant, then spreadsheet, then eventually software, then eventually a model. That order costs the most, because the spreadsheet phase creates a model that immediately diverges from the books and then quietly stops being trusted.

A better order, roughly by stage.

Under about thirty transactions a month, no investors. An accountant for statutory work and a spreadsheet you update monthly. That is genuinely sufficient. Do not let anyone sell you a stack you cannot yet fill.

Thirty to a few hundred transactions, or you have taken outside money. Now add the system. The trigger is not revenue, it is that you have stopped being able to hold the state of the business in your head, or that somebody external has started asking for numbers on a schedule.

Once anyone is asking you what happens next. A board, an investor, or you at 2am wondering whether you can afford the hire. This is where a ledger alone stops being enough, because every question is forward-looking and a ledger only looks back.

The thing you can defer for a year

You do not need a fractional CFO at pre-seed. You need the artefacts a CFO would produce, which is a different and much cheaper thing.

Those artefacts are a forecast you believe, a monthly comparison of plan against actual, and a one-page summary you can send an investor without rebuilding it each time. A good fractional CFO produces those. So does a system, at roughly a hundredth of the cost, provided you are willing to think about your own business.

What a CFO adds on top is judgment about what the numbers mean and credibility in a room. Both are real and both matter more later than they do now. At pre-seed, with two products and eleven customers, you are the person who understands the business best. You are not short of judgment. You are short of a structure to put it in.

The short version

You need an accountant, and you are not getting out of it.

You need a system once you cannot hold the business in your head, which happens sooner than most founders expect.

You do not need a CFO yet, but you do need the things a CFO would make, and you need them before your first board meeting rather than the week of it.

And the connection between the three matters more than any of them individually. An accountant producing books nobody reads, a system nobody reconciles, and a model that diverged in month two is a common and expensive place to end up.

See how the pieces connect