Bank Feed Reconciliation: Common Pitfalls for Canadian Firms

Bank feed reliability issues

Bank feeds are only as good as the connections that deliver them. Understanding the underlying technology helps explain why things go wrong.

Canadian bank connectivity

Most accounting software connects to Canadian banks through aggregators like Yodlee, Plaid, or Finicity. These services "scrape" data from online banking—essentially logging in as the user and copying transaction data.

This approach has inherent limitations:

• Authentication changes: When banks update login processes, feeds break until aggregators update their scrapers
• Rate limiting: Banks limit login frequency, which can delay transaction availability
• Security blocks: Unusual login patterns (from aggregator servers) sometimes trigger security holds
• Multi-factor authentication: When banks require MFA, connections need manual re-authentication

Open Banking coming (slowly)

Open Banking regulations will eventually provide official bank APIs for data sharing. The federal government has committed to a framework, but implementation remains years away. For now, scraping-based connections are the reality.

Auto-categorization gone wrong

Auto-categorization rules save time when they work, but create problems when they misfire. The issues compound because errors can repeat for months before anyone notices.

Common categorization errors

The memo field problem

Bank transaction descriptions are often cryptic or truncated. "POS PURCHASE 1234567" tells you nothing. Without additional context (receipts, invoices), categorization becomes guesswork.

This is why receipt capture integration matters—it provides the context that bank descriptions lack.

Timing and cut-off problems

Reconciliation depends on transactions appearing in the right period. Several timing issues complicate this.

Processing delays

A transaction on December 31 might not post until January 2. For year-end reporting:

• Credit card transactions often take 1-2 business days to post
• Pre-authorized payments post on the next business day if scheduled for a weekend
• E-transfers can be delayed by banking system maintenance

Statement date vs. transaction date

The date shown in online banking isn't always the economic date of the transaction. Pending transactions show estimated dates that change when they clear. This affects:

• GST/HST reporting periods
• Fiscal year-end cut-offs
• Cash flow analysis accuracy

Feed synchronization gaps

Bank feeds don't always capture every transaction. Connection interruptions, authentication issues, or aggregator problems can create gaps. If you're not reconciling to statements, these gaps go unnoticed.

The duplicate transaction problem

Duplicates are one of the most common and frustrating bank feed issues. They occur for several reasons.

Manual entry plus feed

The classic scenario: someone enters a transaction manually (perhaps from a receipt), then the bank feed imports the same transaction. Without careful matching, both stay in the books.

Connection reset duplicates

When a bank feed connection is re-established after a disruption, the system sometimes re-imports historical transactions. If the matching algorithm doesn't catch them, you get duplicates.

Credit card statement duplicates

Some businesses import credit card transactions from the feed AND from a monthly statement upload. This creates systematic duplicates that can be hard to catch.

Prevention strategies

• Import only: Don't manually enter transactions that will come through the feed
• Review before matching: Look for duplicates before accepting suggested matches
• Regular reconciliation: Monthly reconciliation to statements catches duplicates
• Duplicate detection reports: Run reports looking for same-amount transactions on same dates

Intercompany and transfer handling

Transfers between accounts create reconciliation challenges that bank feeds don't solve automatically.

Internal transfers

A transfer from one account to another shows as a withdrawal in one and a deposit in the other. If auto-categorization treats these as expenses and revenue, the books are wrong.

Best practice: Create specific rules for transfers that:

• Match transfers between known accounts
• Use a transfers clearing account or direct account-to-account entries
• Flag unmatched transfers for review

Intercompany transactions

For clients with multiple entities, transfers between companies need proper intercompany accounting. Bank feeds show the cash movement but don't understand the relationship between entities.

Owner draws and contributions

Transfers to/from shareholder accounts are often miscategorized. The bank sees a transfer; the rules might call it "Office Expense" or leave it uncategorized. These require manual review.

Best practices for clean reconciliation

Weekly vs. monthly processing

Processing bank feeds weekly rather than monthly offers several advantages:

• Transaction context is fresh (easier to remember what purchases were for)
• Issues are caught sooner (before they compound)
• Month-end close is faster (most work is already done)
• Cash flow visibility is current

Rule maintenance

Review categorization rules quarterly:

• Delete rules for vendors no longer used
• Update rules for vendor name changes
• Tighten over-broad rules
• Add rules for new recurring vendors

Reconciliation checklist

At each month-end:

1. Confirm all bank feeds are current (check last sync date)
2. Process all pending transactions
3. Run duplicate detection
4. Review uncategorized items
5. Compare GL balance to bank statement balance
6. Investigate and resolve any differences
7. Document reconciling items

Receipt matching integration

Bank feeds tell you money moved. Receipts tell you why. Matching these together:

• Validates categorization
• Provides documentation for tax purposes
• Flags transactions without supporting documentation
• Enables GST/HST ITC claims