What is a data inventory, and how do I make one?

A data inventory is a one-page list of the systems your company uses, the date range each one still holds, roughly how much is in each, what kinds of records they contain and who can export them. It describes your data without containing any, so you can prepare it before anyone asks for a sample or an export.

What goes in a data inventory?

One row per system, with five columns:

• System and plan. For example, Gmail on Google Workspace Business Standard, or Slack Pro. The plan matters because retention defaults and export tools differ by plan.
• Date range retained. The oldest record you can open today, not the year you signed up.
• Approximate volume. Mailboxes, channels, jobs, invoices or tickets. Round numbers are enough.
• Record types. Email threads, chat, quotes, photos, call recordings, invoices.
• Who can export. The person who holds the admin rights the export needs.

Providers use different names for this. Polyshares says its intake “asks for record counts, not files or exports”. Corpus calls it a data manifest and puts it at day 0, with no credentials or exports (Corpus). Troveo’s guide puts inventory first in its six-step workflow (Troveo). In each case it is a description of what exists, made before anything leaves your systems.

Why is our retained history shorter than our company’s age?

Because systems delete or hide records on their own schedules, and companies change systems. Defaults worth checking, as of October 2026:

• Slack Free shows only the last 90 days and, since August 26, 2024, may delete messages and files older than a year on a rolling basis (Slack). Paid plans keep everything unless an owner sets a shorter policy (Slack).
• Google Workspace without Vault rules keeps mail and Drive files until someone deletes them (Google); deleted items are purged from trash after 30 days, and an administrator can still recover them for 25 days after that (Google).
• Microsoft 365 keeps mail until a user deletes it; once removed from Deleted Items, it stays recoverable for 14 days by default (Microsoft).
• ServiceTitan is not obliged to keep your data beyond a 60-day export period after the subscription ends (ServiceTitan).
• QuickBooks Online gives read-only access for one year after you cancel (Intuit).

The other usual gaps are mailboxes deleted when people left and years stranded in a system you replaced.

How do I check what each system still holds?

1. List every system people do real work in, including ones you no longer pay for.
2. Note each one’s plan from its billing page.
3. Open the oldest record you can find in each. That date goes in the inventory, not the sign-up date.
4. Check whether an admin has set retention or auto-delete rules.
5. Count former employees’ accounts that still exist, and note whether they are archived or deleted.
6. Take volumes from the admin console or a report, not from an export.

Who has to be involved in an export?

Whoever holds the admin rights, and those differ by system. Google Workspace’s organization-wide export needs a super admin account at least 30 days old with 2-Step Verification on (Google). Slack’s standard export covers public channels only (Slack); on Business+, exporting private channels and direct messages needs an application approved by the Primary Owner (Slack). Microsoft 365’s bulk export runs through Purview eDiscovery and needs eDiscovery roles (Microsoft).

If your IT is outsourced, the name in this column is often your IT provider’s. That tells a buyer early who must be involved before any records move.

What does a finished inventory look like?

For example, a hypothetical 45-person HVAC contractor founded in 2009 might produce this:

System	Records	Retained	Approx. volume	Who can export
Google Workspace, Business Standard	Email, quotes, job photos in Drive	2014 onward; older mail was not migrated	52 mailboxes, 9 of them former staff	Office manager (super admin)
ServiceTitan	Jobs, invoices, call recordings	2017 onward	About 40,000 jobs	Operations manager
QuickBooks Online	Transactions, bills, invoices	2019 onward, plus a QuickBooks Desktop file for 2009 to 2018	About 60,000 transactions	Owner, with the outside bookkeeper
Slack, Free until March 2025, then Pro	Dispatch and team chat	About March 2024 onward	14 channels	Owner

Seventeen years in business, but between two and twelve years of history in each system still in use; only the old QuickBooks Desktop file reaches back to the founding year. That sentence matters more to a buyer than the founding year. The table also shows where the depth sits: twelve years of email and about nine years of job records, while chat begins in 2024 because the Free plan’s one-year limit removed older messages before the upgrade.

What should we avoid while making the inventory?

• Exporting everything before anyone asks. An export copies personal data onto laptops and drives before you know what scope a buyer wants. Several providers say they need no files at this stage: Avelence’s first step involves “No system connections, passwords, or API keys”, and Telegraph Lab (affiliated with this site) says a business “can get an offer without estimating file sizes or preparing an export”.
• Letting history lapse while you decide. Shortening retention, deleting a departed employee’s mailbox or cancelling an old subscription can end the record for good. Appen’s FAQ warns that “once systems are decommissioned and licences lapse, the record often becomes unrecoverable”.
• Rounding up. Writing “since 2009” when the oldest email is from 2014 will surface in the sample and move the offer.

Once a buyer has read the inventory, the next request is usually a data sample. How the inventory feeds a price is covered in how data licensing offers work, and the guide puts both in sequence.

Use the retention check to compare your software plan’s retention and export rules, then carry the history you have checked into your application.