What Happens When Your Billing System and Your Books Disagree?

Your bookkeeper spends month-end reconciling two records of the same money, and most of the gap turns out to be definitions. What counts as billed, which date governs, and whether a write-down reaches the books all differ between systems, costing days to close and confidence in trust balances.

Where do the month-end hours go?

Picture the close from the bookkeeper’s chair. The billing system’s report gives one total for the month, and QuickBooks Online gives another. Neither is obviously wrong, so the work begins: list the invoices in one system, list the postings in the other, match them line by line, and write a reason next to every line that fails to match.

The matching is mechanical. The cost sits in the reasons. Each difference has to be traced to a cause before anyone can set it aside, and tracing takes the same effort whether the cause turns out to be a mistake or a definition. A bookkeeper who has learned that most gaps are definitional still cannot skip the trace, because the exceptions are the ones that matter. Meanwhile the close waits, and everything downstream of it (partner reporting, trust reconciliation) inherits the delay. The hours show up on the calendar, and the attention they consume shows up nowhere.

Whether a billing tool sends data to QuickBooks Online in one direction or two is a vendor-evaluation question, and two live posts cover it: why sync direction matters for trust accounting and a look at legal billing tools and two-way QuickBooks Online sync. This post takes the operational side of the same problem: what month-end looks like after the tool is chosen, when two systems each describe the same work in their own terms.

Why do two accurate records disagree?

Each system was built to answer a different question. A billing system follows the life of the work: recorded, reviewed, invoiced, paid. QuickBooks Online follows the accounting consequence: when a transaction posts, and to which account. Three definitional forks produce most of the gaps.

What counts as billed. In a billing system, work moves through distinct states: recorded, placed on a draft invoice, approved, sent. In the books, an invoice exists when it posts. A report that counts approved invoices and a report that counts posted invoices will disagree on any invoice caught between those states.

Which date governs. The work date, invoice date, posting date, and payment date can each place the same dollar in a different month. At a period boundary, one system files an invoice under the month that just ended while the other files it under the month that is starting. The gap then reverses a month later, which looks like an error correcting itself and earns another round of tracing.

Whether a write-down is an accounting event. Time written down before an invoice exists usually never touches the books, because unbilled time is typically absent from the ledger. A discount or credit on an issued invoice does touch them, as a reduction of income. The billing system’s realization view sees both; QuickBooks Online sees one. For hourly firms, whose leak runs through realization, this fork decides which system is measuring the leak. Our post on how write-downs, discounts, and credits work on an invoice covers the invoice side.

A worked example: the $3,000 gap

Take a hypothetical firm with invented numbers. The billing system’s month-end report shows $50,000 billed. QuickBooks Online shows $47,000 of income for the same month. The bookkeeper has a $3,000 gap to explain.

Two causes account for all of it. An invoice for $2,000 was finalized on the last day of the month and posted to the books on the first day of the next, so the billing report counts it and the books do not yet. And the firm agreed a $1,000 discount with a client on an invoice that had already been sent; the books recorded the discount as a reduction of income, while the billing report still shows the invoice at its original amount. Start from $50,000, subtract the $2,000 timing difference and the $1,000 discount, and the result is $47,000.

Neither report is wrong. Both are accurate under their own definitions. The bookkeeper still cannot close until every dollar of the gap has an explanation, and next month the $2,000 reverses in the opposite direction — the books will show more than the billing report — so the trace starts again.

The same forks reach trust. Suppose the gap sits on a client’s balance: the billing system shows one figure for what the client has on deposit, and the books’ client trust liability shows another, because a payment was applied on one date in one system and the next in the other. Before the person reconciling trust can tie the ledger to the bank, someone has to establish which balance is current. Each unexplained difference lowers confidence in the number the firm reports as client money. (A client trust liability account is where the books hold that balance.) Reconciling client trust is the firm’s obligation under its jurisdiction’s rules, and no software certifies that it has been done correctly. What a firm can control is how many records it has to explain.

LeanLaw’s Core plan includes two-way QuickBooks Online sync, which addresses the gap created when a change made in one system never reaches the other. It leaves untouched the questions of what “billed” means, which date your close uses, and how the firm treats a write-down. The frame we argue for is a billing layer built on QuickBooks Online, not synced to it, so the books stay the single ledger. Even then, the definitions are decisions the firm makes once, writes down, and applies in both places.

What this depends on

• Who does the books (in-house, or an outside bookkeeper) and which basis the close uses, cash or accrual, since basis changes which date governs.
• How the chart of accounts is built: whether discounts and write-offs post as a reduction of income or as an expense, and how income is split by practice area.
• Which operating model the firm runs. Hourly firms carry unbilled time and write-downs, flat-fee firms bill on milestones, and contingent firms issue no invoice until a matter resolves, so each produces different gaps.
• Which LeanLaw plan the firm is on, since the published plans describe the QuickBooks Online connection differently: two-way in Core and native in Contingency.

How many hours did your last month-end take?

Pull last month’s close and answer two questions. How many hours did the close take? And how many of those went to reconciling two records of the same thing? The second number is the cost of the disagreement.

Then take each difference your bookkeeper traced and ask which definition would make both records correct. Where a definition explains it, the definition belongs in a written policy that both systems follow. Where no definition explains it, you have found an error worth fixing at its source. Which of your definitions are written down today?