Chart of accounts
Mapped, merged where you want it merged, and approved row by row before anything loads. Dead accounts get retired rather than carried forward for another decade.
Migration
A botched migration does not cause an outage. It causes wrong numbers that nobody notices until a year-end review, which is far worse. So we treat reconciliation as a gate rather than a report — the migration cannot advance past a period that does not agree to the penny.
184 consecutive days tied · variance $0.00
The pipeline
A gate is not a checkpoint someone signs off under deadline pressure. It is a condition the pipeline enforces: a variance sends the work back to mapping, and is never waived.
What moves
Mapped, merged where you want it merged, and approved row by row before anything loads. Dead accounts get retired rather than carried forward for another decade.
Deduplicated against the twelve spellings of the same company name, with open balances tied to their control accounts.
Unpaid invoices and bills with their original dates, terms, and ageing intact — because ageing that resets at cutover is worse than useless.
One to two years of transactional history as standard, which is what an auditor expects and what keeps the migration verifiable.
Cost, accumulated depreciation, method, and remaining life — not just net book value, which is the shortcut that breaks next year.
Year-to-date vendor payment totals so January filing does not require the old system to still be running.
The standard scope is opening balances plus one to two years of transactional detail. This is not us being lazy about the rest — it is the scope that auditors expect and the one that stays verifiable. A full-history migration is available, quoted separately, and we will usually try to talk you out of it.
The reason is that historical fidelity decays. Transactions from six years ago were entered under a chart of accounts that has since been restructured, by people who have left, under conventions nobody documented. Moving them creates a body of data that looks authoritative and reconciles to nothing. Keeping the old system available read-only for the statutory retention period costs a fraction as much and is what most auditors actually prefer.
We never rewrite a period your accountant has signed. Prior-year balances arrive as opening balances, and the signed statements remain the record for those years. If a prior period genuinely needs restating, that is an accounting decision your CPA makes deliberately — not a side effect of a data load.
Every migration produces a per-period, per-account variance report showing that the loaded balances agree with the source. Early runs show variances, and that is the process working — timing differences, unmapped accounts, and classification disagreements surface in week one instead of detonating on a cutover weekend.
You keep that report. It is what you hand your auditor when they ask how you know the conversion was complete, and it is the single artefact that makes the difference between a migration you can defend and one you have to explain.
If you have been on Connect, this is much less of an event. The shadow ledger has already been reconciling against your books daily for months, so the migration is not a load and a leap — it is turning off the write-back on a ledger that has already demonstrated it agrees. We will not sell a cutover until three consecutive closed months have tied at zero variance, which is a gate on our revenue as well as on your risk.
Source systems
Desktop is a genuinely different piece of work from QuickBooks Online — QBXML, the SDK, IIF files, and direct company-file access rather than a modern API. It is quoted separately and takes longer. Any vendor treating them as one line item has not done a Desktop conversion.
Questions
Source system, entity count, volume, and history depth. You get a fixed price and a timeline, not a range.