Most founders who want to switch bookkeepers wait months longer than they should, and the reason is almost always the same. Not loyalty, and not price. It is the fear that something will fall through the gap between the old provider and the new one, and that it will surface during a raise.
That fear is reasonable and the risk is manageable. Switching is a defined process with one technique at its centre, and the thing that gates the timeline is access to your systems rather than the calendar.
Get your accounting file before anything else
Before you give notice, before you sign with anyone new, establish what you can actually take with you and take a copy of it.
If your books live in an accounting file you hold the subscription to, this is straightforward. If they live inside a provider's own platform, what you can extract is often report-level exports rather than a ledger a new provider can continue working in, and that difference changes the whole project. Bench's December 2024 shutdown made this concrete for a lot of companies, who discovered the answer at the worst possible moment.
Run a parallel close
This is the technique that removes most of the risk, and it is worth insisting on. For one month, both the outgoing and incoming provider close the same period independently. You then compare the two closes.
The point is not redundancy. It is that differences between two independent closes are exactly where historic errors live. If the old books were wrong, this is the month you find out, while someone is still accountable for them, rather than eighteen months later in a data room.
Do not sign off the transition until the two closes reconcile and every difference has an explanation. An unexplained variance carried into your opening balances does not go away. It becomes the number a buyer's accountant asks about.
The order of operations
- Take a copy of your accounting file and confirm what you own versus what is exportable.
- Choose the cutover month. Month end is cleanest, and avoid the month you are closing a round.
- Give notice, but keep the outgoing provider engaged through the parallel close rather than ending on the announcement.
- Hand over access: bank and card feeds, payroll, billing or invoicing systems, expense tools, and any accountant-level access to the ledger.
- Run both closes for the cutover month and reconcile them line by line.
- Resolve differences and fix opening balances properly rather than plugging them.
- Confirm who files what for the current tax year, so nothing sits in the gap between two providers.
Step seven is the one that bites after the fact. Tax filings and payroll returns often straddle a transition, and both providers can reasonably assume the other is handling them. Put it in writing.
How long it takes
Onboarding time is mostly waiting for access rather than work. Zinance is live in about seven business days on this pattern, with the parallel close included. The parts that stretch a timeline are predictable.
| What slows it down | How to avoid it |
|---|---|
| Bank feed access needs the account owner | Line up the person with admin rights before day one |
| Outgoing provider is slow to release the file | Request it in writing before giving notice |
| Historic books are behind | Agree who completes the catch-up, and price it, up front |
| Cutover collides with a raise or year end | Move the cutover, not the raise |
When not to switch
Two situations are worth waiting out. If you are inside a live fundraise, the disruption is not worth it, so finish the round and move immediately after. And if the problem is a person rather than the provider, ask for a different team first, because a switch is a large action to take on a solvable complaint.
There is one exception. If you cannot get a straight answer about whether you can take your books with you, that is a reason to move sooner rather than later, and it does not improve with time.
What to ask before you sign with anyone new
The transition is also the moment you have the most leverage, so use it to settle the questions that are awkward to raise later.
- Will you run a parallel close for the cutover month, and is it included?
- Where will my ledger live, and do I hold the subscription? This decides whether the next move is a handover or a migration.
- Who completes any catch-up work on historic months, and at what price?
- Which filings are you taking responsibility for in the current tax year?
- What happens to my file and my access on the day I give notice?
The four kinds of difference a parallel close finds
Reconcile them line by line assumes you know what you are looking at when a difference appears. In practice they fall into four categories, and the category tells you how seriously to take it.
- Timing differences. The same transaction recorded in different periods, usually around month-end cutoff. Benign in isolation, but a pattern of them means the old close had a soft cutoff, which matters if anyone is comparing your months.
- Classification differences. The same amount in different accounts. Often a genuine judgement call, sometimes a default nobody chose. Resolve it by deciding the treatment going forward and applying it consistently, rather than by preferring whoever is newer.
- Basis differences. One close accrues something the other does not: prepaid expenses, accrued payroll, deferred revenue. These are the important ones. They usually mean the old books were cash-basis in substance regardless of what the software was set to.
- Genuine errors. A missing transaction, a duplicated one, a reconciliation that never balanced and was plugged. Rare in well-run books and the entire reason to run the exercise.
Categories three and four are the ones to resolve before signing off. Categories one and two can often be accepted with a written note on the treatment adopted, provided the note exists. What you must not do with any of them is adjust the opening balance to make the difference disappear without recording why, because that converts a known issue into an unexplained one.
The first ninety days with the new provider
The transition is not finished when the parallel close reconciles. Three months of specific checks establish whether the new arrangement is actually better or merely newer.
- Month one: measure the close, do not just receive it. Note the date you got a complete package. That date is your baseline, and if it is later than what you were promised, raise it now rather than accepting a new normal.
- Month one: check the bank feeds are yours. Confirm connections are configured under credentials you control rather than the provider's. This is the failure that produces silent gaps later.
- Month two: test a judgement question. Send something genuinely ambiguous, an unusual contract or a transaction you are unsure about, and see whether you get a decision with reasoning or a question back. This is the whole difference between a recording service and a bookkeeping one.
- Month two: confirm the tax handoff happened. Follow up on the written confirmation of who files what. Assume nothing until you have seen it acknowledged by both sides.
- Month three: ask for something backwards-looking. Request a report you have not asked for before, such as gross margin by product or spend by department. Whether it can be produced tells you if the chart of accounts is actually structured for your business or merely tidy.
If month three produces a report you could not have got from the old provider, the switch was worth making. If it produces the same statements slightly faster, you changed vendors rather than solving the problem, and the underlying issue was probably structural rather than about who was doing the work.
Where Zinance fits
Zinance onboards in about seven business days with a parallel close as standard, working inside your own QuickBooks file so the ledger stays yours throughout and the next transition, whenever it comes and whoever it is to, is a handover rather than a migration. For what changes once you are funded, see bookkeeping for funded startups.
Before you switch, it is worth being clear on what you are switching to. The best bookkeeping services for VC-backed startups compares the market, and what changes in bookkeeping after you raise covers why the requirements move once you have a board.