Job knowledge test
Accounting and bookkeeping skills test
Working accounting judgement: coding a transaction to the right account, reconciling a month that does not tie, and knowing which discrepancies to fix and which to flag.
- Time to complete
- About 18 minutes
- Format
- Twelve coded transactions, one reconciliation, one written explanation
- Test family
- Job knowledge
- Included in
- All plans
What this test can see, and what an interview cannot
Accounting job titles say less about the work than almost any other function. A bookkeeper at a ten-person company closes the month, runs payroll and argues with the bank. A bookkeeper at a two-hundred-person company codes invoices and does nothing else. The title tells you where somebody sat, not what they can do.
The interview does not close the gap, because the standard questions have standard answers. Ask a candidate to explain debits and credits and you find out whether they have read the same page as everyone else. What you need to know is harder to rehearse: when a number does not tie out, do they notice, and do they know whether it is theirs to fix or theirs to raise. Every item here is a piece of real work with the documents attached, so the candidate does the work instead of describing how they would approach it.
The scenarios a candidate works through
An invoice that does not match the purchase order
The invoice bills three hundred units. The PO says two hundred and eighty, and so does the receiving note. Nothing has been paid yet.
What it separates Whether they open the third document before coding anything, and whether a twenty-unit gap reads as a supplier error to raise or a difference to absorb.
A cost that could be capitalised
Four thousand dollars covering a server, its installation and twelve months of support, arriving as one invoice with one line.
What it separates Whether they split the line at all, and whether they can say which part belongs on the balance sheet and which part is this year's expense.
A bank line with nothing behind it
A payment leaves the account on the last day of the month and appears nowhere in the ledger.
What it separates Whether the instinct is to go and find the source document, or to post an adjustment that makes the columns agree.
The same invoice twice
Two entries from one supplier in the same week, identical amounts, invoice numbers one digit apart.
What it separates Whether they catch it, and whether the first move is to stop the second payment or to email the supplier and wait.
An aged receivable nobody has chased
Ninety days out, a customer who still orders every month, and a sales manager who would rather it was left alone.
What it separates How they weigh the ledger against the relationship, and whether they escalate with a number attached or a complaint attached.
A payroll change that moves the gross
An employee adds a pre-tax deduction mid-period and asks why their take-home moved by less than the deduction.
What it separates Whether they can explain the arithmetic to a non-accountant without hand-waving or reciting a table at them.
How an answer to a conflict scenario is scored
Items are scored on the output, not the route taken to it. Where a single total would hide something, the report does not show one: the result breaks into four subscores, for bookkeeping mechanics, accounts payable and receivable, payroll, and reconciliation.
That split is the part worth reading. The expensive mistake in accounting hiring is rarely the candidate who is weak at all of it, because a CV screen already catches them. It is the candidate who is strong at three of the four and has never touched the fourth, which averages into a middling total and looks like mediocrity. On the reconciliation task, three things carry the weight:
- Whether it balances. Checked first, and it is pass or fail. A reconciliation that does not tie is not a partial-credit situation, however sensible the working looks.
- What was done to make it balance. Finding the missing document scores at the top. Posting a plug to force agreement scores at the bottom, and scores there even when the plug is small, because the size of the first one is never the point.
- What was flagged rather than fixed. Candidates are told they may raise anything with a named person. Raising the duplicate payment and the PO gap scores; quietly correcting either one does not.
The written explanation on the payroll item is scored by your team against a rubric shown beside the answer, the same criteria for every candidate. It is the one human-marked piece on purpose: an accounting hire in a small company spends much of the week explaining a number to somebody who does not want to hear it, and nothing automated measures that.
What separates a strong result from a weak one
| Reading the report for | A strong candidate | A weak one |
|---|---|---|
| Coding | Asks what an account is for before choosing it, and splits a mixed line. | Codes it wherever the last similar invoice went. |
| Reconciliation | Chases the missing document and leaves the difference visible until it turns up. | Posts an adjustment and moves on. |
| Controls | Stops the second payment first, then contacts the supplier. | Emails the supplier and waits for a reply. |
| Escalation | Raises the two items that need a decision from somebody else. | Raises everything, or silently fixes everything. |
| Payroll | Gets the arithmetic right and can explain it in plain words. | Right number, no explanation a non-accountant could follow. |
| Pressure from a colleague | Chases the ninety-day balance with the figure in front of the sales manager. | Drops it because somebody senior would rather they did. |
Roles it suits
- Bookkeeper, including the office manager who also keeps the books
- The closest fit, and the only role where all four subscores matter at once. When the books live in a spreadsheet rather than a ledger system, pair it with the Excel skills test, which covers the mechanics this one assumes are already there.
- Accounts payable clerk
- Weight the invoice matching, the duplicate payment and the coding items; the payroll subscore barely applies. If the job is high-volume invoice entry rather than judgement about invoices, the data entry accuracy test predicts more of the actual day.
- Accounts receivable and collections
- The aged-receivable item is the one that counts, and it is one item, so read this as a partial signal. It says nothing about how the candidate sounds on a collections call.
- Payroll clerk
- Two items and the written explanation touch payroll: enough to rule somebody out, not enough to rule them in. Payroll rules are jurisdictional and this test is not, so add items for the rules you run under. For keypad throughput, the 10-key numeric entry test is the faster measure.
- Staff accountant and accounting assistant
- A fair screen for the transactional half of the role. It covers nothing about the close, consolidation, or anything with a reporting deadline attached, and those are usually what the job gets judged on after month three.
How long it takes, and where it goes in the process
Eighteen minutes: twelve coded items at roughly a minute each, five on the reconciliation, and a couple on the written explanation. The clock runs per section rather than per item, so a candidate who reads the whole invoice before answering is not punished for it.
It sits after the application and before the phone screen. Eighteen minutes is short enough to send to a whole shortlist rather than the three CVs that read best, and that is where it earns its place: accounting candidates who screen badly on paper are disproportionately career changers and people who came up inside a small company without the title.
What it will not tell you
It is software-neutral on purpose, so it will not tell you whether somebody knows QuickBooks, Xero or NetSuite. It tests whether they understand what the software is doing underneath, which is the part that transfers between systems.
That is a trade-off rather than a gap. If the role is mostly inside one system and somebody has to be productive in week one, a product-specific test predicts more. The argument for the neutral version is that tool knowledge is the fastest thing to teach and the least durable to hire for: a candidate who knows the accounting picks up your ledger in a fortnight, and one who knows only your ledger is stuck the day you outgrow it.
Two further boundaries. Nothing here touches tax. And a credential is not a skill: this test does not verify that anyone is a CPA or a certified bookkeeper, and a good score should never be read as confirming a claim on a CV. Check credentials with the body that issued them.
Every test has a boundary like this. A test used outside it stops predicting anything, which is the usual reason a hiring team loses faith in testing altogether.