Data Ownership Audit Checklist

Every software subscription you pay for holds some of your business: customers, orders, pricing, job history, procedures. The question this audit answers is simple. If you canceled a system tomorrow, would that data leave with you, or is it staying with the tool?

You can run the whole audit yourself in an afternoon. No vendor calls, no technical skills beyond exporting a file and opening it.

Step 1: List your systems

Write down every system that holds business records you would miss. Typical list for a founder-operated business:

• Accounting or bookkeeping software
• CRM, or wherever customer history actually lives
• Quoting, invoicing, or point-of-sale tools
• Scheduling, dispatch, or field service software
• Inventory or order management
• Email marketing or contact lists
• Payroll and HR records
• Any industry-specific platform the business runs on

For each one, note what it holds and how painful losing that history would be: high, medium, or low. Do the high-pain systems first.

Step 2: Run the exit test on each system

The exit test is one question asked honestly: can I get everything out, in a form another tool could read, without asking the vendor for help?

For each system:

1. Find the export function yourself. Look under Settings, Reports, or Data. If you need a support ticket just to find it, write that down. It is a finding.
2. Export your core records. Customers, transactions, history, whatever the system is for.
3. Open the file in a spreadsheet and actually look at it. Do not trust that the export worked. Check it.
4. Score the system: Green (complete export, readable format, tested), Yellow (partial export, or complete but you have never tested it until today), Red (no real export, proprietary format, or export requires the vendor’s cooperation).

Step 3: What to look for in the export

An export can succeed as a download and still fail as an exit. Check five things:

• Completeness. Count records in the file against records in the system. Vendors often export the main table and quietly drop the rest.
• Attachments and files. Photos, signed documents, PDFs on a job or customer record usually do not ride along with a data export. Where are they?
• History, not just current state. A customer list is easy. The order history, notes, and activity log are the asset. Are they in the file?
• Format. CSV or another open, machine-readable format is good. A proprietary backup file only the same vendor can open is not an export, it is a hostage photo.
• Relationships. In the system, orders connect to customers and jobs connect to sites. In the export, do those connections survive, or did you get disconnected lists you would have to rebuild by hand?

Red flags

Any of these on a system that holds high-pain data deserves attention this quarter, not someday:

• No self-service export at all. Getting your data requires a support ticket, a fee, or a “data services” engagement.
• Export limits: row caps, date-range caps, or one table at a time on a system holding years of history.
• The contract is silent on what happens to your data when you leave, or gives the vendor a short deletion window with no return obligation.
• The export leaves out the fields your team actually uses, especially custom fields and notes.
• Attachments are unreachable except one record at a time, by hand.
• The vendor charges to unlock “your” API or full backup.

What to do with the results

• Greens: put a repeat of the exit test on the calendar, quarterly or before every renewal. Exports rot as vendors change things.
• Yellows: schedule the fix. Usually that is one working session to find the missing pieces and set an export routine.
• Reds on low-pain systems: decide consciously to accept the risk, and write that decision down.
• Reds on high-pain systems: this is now a real project. Get a current export, however imperfect, today, so the worst case has a floor. Then plan the structured copy of that data in storage you control.

The goal is not to abandon cloud software. It is to be able to switch any single vendor without losing your history. Not painlessly, but possibly. That possibility is negotiating power at every renewal, and it is the difference between owning your data and renting access to it.