Your software inventory is only as durable as the vendor holding it

A software management platform shut down this month. Its customers' data was destroyed the same week.

Productiv was an established vendor in the SaaS management category. It connected to a company’s identity, finance, and contract systems, discovered the applications in use, and became the system of record for what a company owned, who used it, and what it cost.

Its own notice states that Productiv “officially ceased operations on August 6, 2026,” that “all customer access to the Productiv platform was terminated on August 5, 2026,” and that “all production systems, data stores, and backups have been permanently and securely destroyed. No customer data has been retained.”[1] The notice directs creditor-related matters and service-interruption claims to a third-party case-management portal.

No reason for the closure is stated, and this article does not speculate about one. The operating fact is enough on its own: whatever inventory those customers had built inside that platform is not recoverable from the vendor.

This is not an argument that SaaS management platforms are a bad idea. They do real work a spreadsheet cannot. It is an argument about a question most buyers never ask.

Short answer

Before a system becomes your record of what you own, ask what you would still have if it stopped existing tomorrow.

For most operational tools the answer does not matter much. If a scheduling app disappears you book meetings another way by Friday. The answer matters when the system holds a record rather than performing a task: your inventory, your owners, your decisions, your history.

The practical version is one line: keep the source of truth in a form you could still read if the vendor were gone.

Discovered data is derived data

An integration-fed inventory is not a document you wrote. It is a view the vendor computes for you from systems it connects to on your behalf.

That distinction is invisible while everything works, and it is the whole story when it does not. Two things are lost when such a platform closes, and they are not equally replaceable:

• The discovered layer regenerates. Applications, seats, and spend can be rediscovered by another tool, because the source systems (your identity provider, your accounting system, your card statements) still exist and still hold the underlying facts.
• The authored layer does not. The owner you assigned. The renewal you decided to let lapse, and why. The tool you reviewed twice and kept both times. The note explaining that a duplicate looks redundant but is not, because one instance is tied to a customer-facing workflow. Nobody rediscovers a judgment. It was reasoning, and reasoning only exists where somebody wrote it down.

Companies tend to insure the first layer and forget the second, because the first is what the product demo shows. The second is the part that took a year to accumulate.

The exit test

Five questions, worth asking of any system that holds an operating record. None of them require a vendor to answer.

1. Where does the record actually live?

In a file you hold, or in a database you access? Both are legitimate. Only one survives the vendor.

2. Could you export it today?

Not “is there an export feature.” Run it. Once. The gap between a documented capability and a file on your disk is where this kind of loss happens.

3. Is the export usable on its own?

Open it. If it is an archive of internal identifiers that only mean something inside the product, it is a receipt, not a record.

4. What is derived, and what did you author?

Separate the two explicitly. Then protect the authored half harder, because it is the half nothing else can reconstruct.

5. How many days would a rebuild take?

Give it a number and write the number down. “We would manage” is not a plan, and an unstated assumption is not a decision.

What to do this week

If you run any system of record for software, spend, or vendors:

1. Export it. Today, not at renewal.
2. Open the export and confirm a person could read it without the product.
3. Store it where it survives both the vendor and the departure of whoever set it up.
4. Write down what you would lose that no export contains. That list is your real exposure.

None of that is a purchase. It is an afternoon, and it is worth more than any vendor comparison you could run in the same time.

What this note does not claim

It does not claim the SaaS management category is unsafe, that connected tooling is a mistake, or that any other vendor is at risk. One company closed. No reason was published, and none is assumed here.

It does not claim a company that follows the exit test will save money. Durability and savings are different outcomes, and conflating them would be the same error this note is warning about.

Competing vendors have published migration guidance for affected customers.[2][3] If you are one of those customers, those pages are more useful to you today than anything argued here.