Catch up prior-period books with Tally

Back work bookkeeping is Tally's managed program for bringing a company's past periods current. It is the tool for a company that has months or years of transactions that were never booked: you give Tally the historical exports, it analyzes and categorizes them, previews financials that must balance and tie, and posts them into the books only after you approve.

Back work is a supervised program, not a one-shot import. Tally holds the whole job as a single durable thread for the company, so if a step stops it resumes from where it was rather than starting over, and the books do not change until a verified financial preview is approved.

Back Work Versus Starting New Books

Both programs are run through Tally and share the same financial-preview and verification machinery, so it helps to be clear on which one you want.

• Start new books with Tally builds the initial set of books for a company — a fresh chart, registers, and the first import. Reach for it when the company is new to LedgerHQ.
• Back work bookkeeping brings prior periods current. It can target an existing company and add historical activity to it, and it protects against re-importing a period it has already booked. Reach for it when the books exist but a stretch of history is missing.

If you are unsure, describe the situation to Tally and let it confirm which program fits before it starts.

Start The Program

Open the full-page Tally workspace and tell Tally to catch up the company by its exact name. Tally first checks the name so the program binds to the right company: it either finds the verified active company, offers an available name when there is a collision, or creates the company using LedgerHQ's standard setup, including the standard chart of accounts. One company has one active back work program at a time.

Intake accepts one CSV or XLSX transaction export per account — checking, savings, credit card, loan, and so on — and you can add several at once, including a ZIP of those files. Add every account's history you want brought current.

Registering And Reading The Exports

When files arrive, Tally runs a deterministic scan of the exports. It normalizes the rows, groups them by account and by month, checks each account's running balance to confirm the history is complete and in order, and flags duplicates and anything that needs a look. Every intake problem has to be resolved before the program moves on — the scan is how Back work proves it is working from complete, trustworthy evidence rather than a partial file.

If an export overlaps a period Back work has already imported for that account, Tally stops and asks whether it is the same account or a different one, so the same month is never booked twice. Use ordinary corrections, or supply a non-overlapping period, rather than re-importing an overlap.

Confirm Accounts And Answer Categorization

From the normalized evidence Tally proposes each account's setup — its name, type, opened date, and opening and closing balances read from the statement running balances — and marks the covered and missing months. Where the export proves a fact, Tally uses it; where it cannot, it asks you to confirm. You confirm the account type and enter a balance only for the facts the file could not establish on its own.

Tally then matches transfers between the company's own accounts and researches how the remaining activity should be categorized, grouping transactions by vendor and pattern against the company chart of accounts. It reuses the company's confirmed rules, and its own category suggestions are held to the same firm AI posting confidence threshold used everywhere else in bank feeds.

High-confidence, unambiguous groups are handled internally. Only the groups that are genuinely uncertain reach you for review, where you confirm the treatment before the program seals its plan.

The Financials Must Tie Before You See Them

Back work does not hand you an import and ask you to trust it. Before it completes, it rebuilds the posted books into a verification snapshot and runs eight live checks, all of which must pass:

• Exact company — the program is bound to the intended company.
• Registers mapped — every source account has a register.
• Source counts — every source row is accounted for.
• Transactions complete — the imported rows match the evidence.
• Transfers settled — matched transfers net to zero.
• Trial Balance — a zero-difference trial balance.
• Register balances — each account's roll-forward ties.
• Balance Sheet — the balance sheet balances.

Only when all eight pass does Tally deliver the final review — the verification table plus the resulting Income Statement, Balance Sheet, and register roll-forwards — and mark the program complete. Every source row is reported as imported and posted, parked for ordinary review, excluded, or a duplicate, so nothing is silently dropped.

After The Catch-Up

Once the final review is delivered, the back work program is complete and closed. From there the company runs its ordinary workflow: connect Plaid, code current activity, collect statements, reconcile, and produce reports. Rows Tally parked during the catch-up stay outside the books and are listed in the final review for ordinary follow-up.

Corrections to a completed back work period are made with normal recoding, not by reopening the program. If Tally reports a repeated external failure during a catch-up, the program stays saved with a next action — fix the file, account, or provider problem and ask Tally to continue the active program.