Have you really outgrown your accounting software? A test to run before you migrate
Most finance teams that want a new ledger are short of hours, not features. An afternoon exercise that shows which one you have, before you commit to a migration.
Mark Holleman6 min read
Working on this yourself? Talk to Mark, our co-founder. We’re here to solve it with you.
On this page
You have genuinely outgrown your accounting software when it refuses to do something your business needs, not when doing it takes too many hours. In most growing companies the second is the real problem, and hours can be won back without replacing the software your accountant already works in.
Picture the 9th of the month at a 40-person company on Exact Online. The finance lead is still matching last month's Mollie payouts. Nine supplier invoices sit in three people's inboxes waiting for approval. The founder wants margin per project and gets a promise for Friday. Somebody says the obvious thing: we need a bigger system.
Sometimes that is right. Often the new system arrives a year later and the 9th of the month looks exactly the same. This post gives you a way to tell the two situations apart.
Is the software refusing, or is the work just slow?
That is the only question that matters before you plan anything.
Software that refuses is a hard limit. You ask for something and there is no way to do it, however much time you spend. Work that is slow is a different thing. The ledger can hold the transaction, but getting it there correctly, with the right account, VAT code and document, costs a person too many hours.
Both feel the same at month-end. They need very different answers. A new ledger removes refusals. It does nothing for slow work, because the matching, chasing and checking come with you to the new system.
How do you find out where the hours go?
Run a short exercise with whoever does your books. It takes an afternoon.
- List the recurring finance tasks that take the most time each month.
- Put a realistic hour count next to each one.
- For each task, ask whether the ledger blocks it or whether it is simply done by hand.
- For the manual tasks, ask whether there is a clear source and a rule you could write down. Matching a payout to its sales, routing an invoice to the budget owner and chasing a missing receipt all qualify. Those are agent work.
- Whatever is manual but has no clear rule yet is process work. It needs a decision from you first.
- Add up the hours in three groups: agent work, process work, and what the ledger cannot do.
A worked example
Here is an illustrative result for the company above. The numbers are invented and add up.
- Matching bank lines and Mollie payouts, 16 hours. Agent work.
- Routing supplier invoices for approval, 9 hours. Agent work.
- Chasing receipts for card spend, 7 hours. Agent work.
- Booking accruals and prepayment releases, 6 hours. Agent work.
- Correcting VAT codes before the return, 5 hours. Agent work.
- Building the monthly management report by hand, 12 hours. Process work, because nobody has agreed what goes in it.
- Answering ad-hoc questions from the founder, 6 hours. Process work, because the answers depend on definitions nobody wrote down.
- Reporting margin per project, 8 hours. The ledger cannot do it, because the current setup has no project dimension.
That is 69 hours a month. Agent work accounts for 16 + 9 + 7 + 6 + 5 = 43 hours, process work for 12 + 6 = 18 hours, and the ledger limit for 8 hours. A migration would address 8 of the 69 hours, after a project of several months. Even then, check first whether your ledger offers the missing dimension in another edition or as a setting.
When is the ledger genuinely the limit?
Real refusals look like this:
- A dimension you need does not exist. You must report by project, cost centre or location, and there is no field, add-on or edition that provides it.
- It cannot produce a filing you owe. For example, a UK company needs software compatible with Making Tax Digital to submit its VAT returns. If yours cannot, nothing else matters.
- Support is ending. The vendor is retiring the product, or your accountant no longer works with it and you do not want to change accountants.
- A hard ceiling. The plan you are on caps users, administrations or transaction volume, and there is no higher plan.
If one of these is true and it costs you serious time, plan a migration. If none is, the answer to your bottleneck lies elsewhere.
What are the signs that the work around the ledger is the problem?
Look at what fills the days of your finance team, not at the software itself:
- Receipts are chased by hand, with the same reminders going out every month.
- Payouts from Stripe or Mollie and card transactions are matched days after they happen.
- Supplier invoices wait in personal inboxes, and nobody can say which are approved.
- The same supplier is booked to different accounts depending on who did it.
- Between closes, nobody quite trusts the numbers, so every question becomes a reconciliation first.
Every item on that list has a clear source and a rule. None of them requires a different ledger.
What does a migration really cost?
A ledger migration is more than a new subscription. Count the full cost before you commit:
- History. You need opening balances at a clean cut-off date, and usually some transaction history for comparisons. Mapping an old chart of accounts to a new one is where errors hide.
- Record keeping. The old administration has to stay available. In the Netherlands, the Belastingdienst requires you to keep your basic records for seven years, and in the UK a company generally keeps its accounting records for six years. Cancelling the old subscription does not end that obligation.
- Your accountant. Many firms have their checks and filing routines built around specific software. A switch can mean a new firm, new fees or a transition year.
- Filings. VAT returns and annual filings have to keep working through the switch.
- Integrations. Banks, payment providers, payroll and invoicing all need reconnecting and testing.
- Running in parallel. For at least one period, ideally a full quarter, you check the new books against the old ones.
Check the record-keeping and filing points for your own situation with your accountant.
Which route should you take?
The exercise points you to one of three routes, and you can combine them.
Settle the rules
Process work needs decisions, not software. Agree what the management report contains, write down how recurring items are booked, fix who approves which invoices and set dates for the close. Do this on any route, because agents and new ledgers both run on the rules you give them.
Put agents on the ledger you have
This is what we built Thred Layer for. Agents connect to your existing accounting software, banks, inboxes and payment providers and take on the agent work from your list: matching bank lines and payouts, preparing invoice approvals, chasing missing documents, booking accruals and answering questions from live numbers. Thred keeps a full copy of your ledger in two-way sync, and finished entries are written back to your accounting software. Your accountant keeps working where they always have.
You decide per agent how much freedom it gets. Some ask you first, some run autonomously within limits you set, and anything that needs a person lands in one inbox. Agents read by default and write only with your permission, as described on our security page.
Migrate when the software refuses
If the ledger column of your exercise is large, migrate, and use the task list as your requirements. Choose a clean cut-off, map the chart of accounts carefully, run both systems side by side and keep the old administration accessible. If you move to a ledger built for agents, such as Thred's own Thred Ledger, our team moves your opening balances and history, connects your banks and systems, and runs alongside your current setup until you are confident. Starting with Thred Layer does not close that door.
The point of the exercise
Founders and finance leads want the same thing either way: books that are correct in real time and a close that does not grow with the company. Replacing the ledger is one way to get there. Taking the agent work off your team's desk is usually the faster one.
To see agents working on the ledger you already have, look at how Thred works for businesses or book a demo.
Frequently asked questions
How do I know if I need a new ledger or better processes?
List your recurring finance tasks with their hours and mark each one as agent work, process work or something the ledger cannot do. If most hours are manual work with a clear source and rule, settle the rules and automate first. If the ledger blocks work you must do, plan a migration.
How long do I need to keep my old accounting records after switching software?
In the Netherlands, basic records generally have to be kept for seven years. A UK company generally keeps its accounting records for six years from the end of the financial year they relate to. Check the exact obligations for your company with your accountant.
Can I automate bookkeeping without changing my accounting software?
Yes. Agents can connect to your existing ledger, banks, inboxes and payment providers, do the matching, approvals and chasing, and write finished entries back. Your accountant keeps working in the software they know.
What makes a ledger migration go wrong?
Usually a messy chart of accounts mapping, an unclear cut-off date, integrations that were not retested, and no parallel run. Running both systems side by side for at least one period catches most errors.
Sources
Read next
- Switching accounting software without losing your history: what to migrate, what to archive, and when not to switch
- Getting messy exports into your books: how to clean CSVs from banks, webshops and payroll reliably
- EU data residency for AI in finance: where models run, what gets logged and how long it stays
Where Thred comes in
Finance agents running on top of your existing ledger. You set the rules, approve the work and handle the exceptions. No migration.



