A billing reconciliation that skips steps is one that leaves money behind. This checklist walks through every phase, from pulling your data to updating your PSA and issuing corrected invoices. Run it monthly and nothing falls through the cracks.
Step 1: Data Preparation
Before you run any matching, make sure your source data is clean and aligned. Garbage in, garbage out.
- Confirm both exports cover the same billing period, matching start and end dates.
- Export vendor billing from Microsoft Partner Center or your distributor (Pax8, Ingram) as a CSV at the line-item level, not a summary.
- Export your PSA billing report from ConnectWise, Autotask, or Kaseya. Include company name, product, quantity, and unit price.
- Verify file sizes are reasonable. Empty or truncated files will break your match.
- Open both files and confirm column headers are present and correctly labelled.
- Check for encoding issues. UTF-8 is recommended. Watch for special characters in company names.
Step 2: Normalization Checks
Mismatched names and SKUs are among the most common reasons reconciliation fails. Standardize before you match.
- Confirm client names use consistent formatting across both files, or verify your tool handles normalization automatically.
- Confirm product and SKU names are mapped. Microsoft renames products regularly; make sure your alias map is current.
- Remove or handle blank rows and null values in quantity fields.
- Confirm currency consistency, especially if you bill in both CAD and USD.
Step 3: The Matching Process
This is where leakage becomes visible. Run matching in order and log your results at each stage.
- Run exact matching first. Record how many rows matched cleanly.
- Run fuzzy matching on the unmatched rows. Set your threshold based on data quality, typically 0.65 to 0.75 similarity.
- Manually review any fuzzy matches with confidence scores below 0.75 before taking action.
- Flag all unmatched vendor rows, those with no PSA counterpart. These are your highest-priority leakage candidates.
- Flag all orphaned PSA rows, those billed in your PSA with no vendor counterpart. These are potential overbilling risks.
Step 4: Discrepancy Review
Every flagged row needs a decision. Document your reasoning so you can defend it and improve next month.
- For each quantity mismatch: determine whether the delta reflects a legitimate seat change or a billing gap.
- For each unmatched vendor row: confirm the client is active and the seat is in use. If both are true, it is unbilled revenue.
- For each orphaned PSA row: check whether the subscription was cancelled at the vendor level. If yes, stop billing the client immediately.
- Document your decision for every flagged row: update, write off, or escalate for further investigation.
Step 5: PSA Updates and Invoicing
Reconciliation is only useful if it changes something. Close the loop with your PSA and billing team.
- Update PSA agreement lines for every confirmed mismatch.
- Generate supplemental invoices for underbilled amounts in the current billing cycle.
- Issue credits for any confirmed overbilling.
- Notify affected clients of billing adjustments with clear supporting data attached.
- Log a note in your ticketing system for each correction, including the reconciliation date and the delta amount.
Step 6: Scheduling and Documentation
The MSPs that recover the most revenue are the ones who run this every month without exception.
- Set a recurring calendar reminder for next month's reconciliation. The last Friday of the month works well for most teams.
- Archive both CSV exports with the reconciliation date in the filename.
- Record your leakage ratio for the month: dollars recovered divided by total billing, multiplied by 100.
- Note any recurring issues for process improvement. If the same clients or SKUs flag every month, that is a fixable problem.
Frequently Asked Questions
What if my PSA does not export all the columns I need?
Check your PSA's report builder. Most platforms let you customize exports. If a column is unavailable via report, an API export or custom query can often fill the gap. For missing unit prices, pull them from your price list and add them manually.
How long should a full reconciliation take?
With an automated tool and clean data, a mid-sized MSP can typically complete reconciliation in 15 to 30 minutes. Manual spreadsheet work on the same dataset usually takes 3 to 6 hours. The first run is always the slowest. Subsequent months go faster as your data quality improves.
Who should own this process?
Assign a single owner, typically a finance or operations manager. That person runs the reconciliation, coordinates with technical and account management teams on flagged discrepancies, and maintains the monthly schedule. Shared ownership usually means no one does it.
Stop Losing Revenue to Manual Reconciliation
Leakage Finder automates the comparison between your vendor billing and PSA, flags discrepancies instantly, and gives your team a clear path to recovery every month. If you are still running this in spreadsheets, you are leaving money on the table.
See how Leakage Finder works and start your first reconciliation today.