Best Way to Import Bank Data Into Multiple QuickBooks Files
No account needed for your first 3 conversions. We never store your bank login.
Short answer: for a firm running several QuickBooks company files, the most reliable route is to download each account's activity as a CSV, convert it to a QBO file, and upload that file to the one company it belongs to. Connected bank feeds are fine for a single live client, but they break down across a portfolio because a fresh connection commonly reaches only about 90 days and every file has to be connected, maintained and re-authenticated separately.
The choice matters more than it looks. Most firms drift into whatever worked for their first client and then discover in January, with a dozen files open and a year of history to post, that the method does not scale. Here is what the three realistic routes actually cost you across multiple entities.
| Route | Works well when | Where it fails on a multi file portfolio |
|---|---|---|
| Connected bank feed per company file | One or two live clients you close every month, on banks with a stable connection | A new connection commonly pulls only about 90 days, so it cannot cover prior year catch up. Each file connects separately and each connection can drop. |
| Manual entry or journal entries | A handful of transactions, or an adjustment you need full control over | Time scales with transaction count, and journal entries bypass the bank review queue, so nothing is matched or duplicate checked. |
| Convert the bank CSV to a QBO file and upload | Catch up work, prior year history, Desktop clients, and any month the feed missed | One file per upload, so it is a per account pass rather than one button for the whole portfolio. |
Last updated September 2026.
Why the bank feed stops being the answer at scale
A connected feed is genuinely good at one thing: keeping a live file current from today forward. The trouble is that almost nothing a firm gets paid for is that. Onboarding a client means posting history the feed never saw, and a freshly connected account commonly reaches back only about 90 days. Take on a client in January who needs the prior calendar year and the feed covers roughly a quarter of the job on its best day.
There is a second problem that only shows up once you have ten files. Connections are per company file, so ten clients means ten sets of credentials, ten sets of re-authentication prompts, and ten separate things that can quietly stop updating. Bank feed error codes are their own small vocabulary: 102 and 105 mean the bank site is down, 185 means the bank wants more than a user ID and password, and 324 means the account cannot be found. None of these are hard to fix individually. Ten of them landing in the same week is a different experience.
The deadline underneath all of this is that Direct Connect is being retired industry wide on October 30, 2026, with new enrollments already closed since April 30, 2026. Individual banks moved earlier: Bank of America ended its OFX service on September 30, 2025, so its downloads are CSV now. Your bank's own notice outranks the industry date, and it is worth checking each client's bank rather than assuming.
Why manual entry costs more than firms expect
Typing transactions is defensible for a dozen rows. It stops being defensible somewhere around a hundred, and it is genuinely risky as a portfolio habit because manually posted journal entries do not go through the bank transaction review queue. That queue is where QuickBooks matches transactions against what is already in the register and where duplicate protection lives. Bypass it and you lose both.
Firms usually reach for manual entry when the import route has failed and the deadline is close. That is exactly when the error rate is highest, and a mistyped sign on a credit card account is the kind of thing that surfaces weeks later as a reconciliation difference nobody can place.
Why conversion is the route that holds up
Converting the bank's CSV into a QBO file gives you the one thing the other two routes cannot: history in whatever depth the bank will hand over, in a format that still goes through the review queue with duplicate protection intact. Every transaction in a QBO file carries a FITID, an identifier unique within the account. QuickBooks skips FITIDs it has already seen and remembers them even after the transactions are deleted. A plain CSV has no transaction identifier at all, which is why the converted file is the safer thing to upload when you are working across many accounts and cannot remember what you posted last Tuesday.
It is also the only route that works for Desktop clients at all. No version of QuickBooks Desktop, Enterprise included, imports a bank CSV into the register. Bank data goes in as a QBO Web Connect file or as IIF, and IIF posts straight to the ledger with no review and no undo. If any part of your client base is on Desktop, conversion is not a preference, it is the mechanism.
How much history can you actually get from the bank?
More than the feed gives you, and it varies by institution, so it is worth knowing the reach before you promise a client a full year. Chase activity downloads reach about 24 months, and the file you want is Account Activity rather than Statements. American Express reaches recent activity plus the past six billing statements. TD online banking shows roughly 18 to 24 months. Citi offers no QBO or QFX download at all, only a CSV covering about 90 days.
Wells Fargo Business Online is a useful example of why the download menu matters. Its Download Account Activity option defaults to 90 days but can offer up to 18 months, and it will give you QuickBooks Web Connect, Quicken or a spreadsheet. The Statements section, by contrast, is PDF only. Firms that go looking in Statements conclude Wells Fargo has no CSV, when the CSV was in the other menu the whole time.
What breaks when you run the import across several files
Three things, and all of them are routing rather than formatting. The first is uploading a correct file to the wrong company, which is easy because every bank names its export something forgettable and your downloads folder fills with six files called the same thing. Rename each download to carry the entity, the account and the period before you convert, and the QuickBooks confirmation screen turns into a real check.
The second is the per upload ceiling. QuickBooks Online accepts 1,000 transactions and 350 KB per upload, and that 350 KB cap applies to QBO, QFX and OFX files as well as CSV. A busy operating account can pass 1,000 transactions inside a quarter and a year of catch up almost always does, so split by month or quarter rather than by row count. Cutting at row 1,000 leaves a seam in the middle of a month that nobody can reconcile later.
The third is signs on credit card accounts. A single Amount column means positive is money in and negative is money out, using a minus sign rather than parentheses, but credit card accounts flip the meaning: a purchase is negative and a payment is positive. Get this backwards on one client's card and the P&L is wrong in a way that looks plausible until someone reads it.
Which QuickBooks product should each client sit on?
For write up, tax only and year end clients, QuickBooks Ledger is the product Intuit built for after the fact work and it is billed per company file, though it is sold only to firms with a QuickBooks Online Accountant subscription. It drops invoicing, receipt capture, bill pay, inventory and sales tax tracking. Check sales tax before you place a client there, because anyone with post Wayfair nexus obligations needs a plan that tracks it.
One structural point that applies to every QuickBooks product: you cannot import bank transactions into a subaccount. Import to the parent and reclassify. Worth knowing before you design a tidy parent and child chart of accounts for a multi entity client, because rebuilding it after transactions are posted is considerably more work.
A close routine that survives a busy month
Work one entity at a time and finish it before opening the next. Download every account for that client, rename each file with entity, account and period, convert each one, upload it to the company file you have open, and reconcile before you move on. It feels slower than batching by task and it is reliably faster in practice, because the expensive part of a multi file close is not the uploading, it is working out which file you were holding when the phone rang.
Count accounts rather than clients when you size the work. A client with a checking account and two cards is three downloads a month, not one. Ten clients averaging two and a half accounts is roughly 25 conversions in a normal month and two or three times that in January. Once each entity is imported and reconciled, the deliverable is usually a set of statements per entity, and turning a finished bookkeeping export into a board-ready P&L, balance sheet and cash flow statement is a separate step worth automating too.
The full workflow, including what to do when a file lands in the wrong company, is covered in the guide to running a CSV to QBO converter across multiple QuickBooks company files. For firms placing year end clients, the QuickBooks Ledger bank import workflow covers what the Ledger feed will and will not reach, and the QuickBooks CSV import limit explains how to split a file that will not upload.