Product rationalization after software acquisitions: which products to sunset

What is software product rationalization?#

Software product rationalization is the periodic review in which an acquirer decides what to do with each product it owns: invest, maintain, merge into another product, sell or sunset. For buy-and-hold software groups, it usually follows a run of acquisitions that left overlapping products serving the same customers.

The review works best as a scoring exercise with fixed criteria and named evidence sources. Without one, decisions follow the loudest general manager, the latest outage or whichever product the group CEO used most recently.

The scoring matrix#

The scoring matrix rates every product on six criteria using qualitative bands such as strong, adequate and weak, with evidence from finance, product, engineering, legal and support. Record the evidence behind each rating so the next review can see what changed.

Weight the criteria to fit the group's strategy, but keep the weights fixed for a review cycle so products are compared fairly. A group focused on cash generation will weight margin heavily; one preparing a platform sale may weight overlap and technical risk.

From scores to decisions#

Scores point to one of five decisions, and a weak score rarely means immediate shutdown. A product with flat growth but healthy margin and loyal customers may be worth maintaining for a long time; a product with heavy overlap may serve customers better merged into a sibling.

Record each decision with a review date. A key customer's renewal, a security finding or a market shift can move a product between categories, and the matrix makes that movement visible.

Contract obligations set the sunset clock#

Contract obligations set the timing of any sunset, because customers on multi-year terms, SLAs or escrow arrangements cannot simply be switched off. Legal should produce a list of every active contract with end dates, notice requirements and data return duties before anyone announces a date.

Where a few large customers hold long terms, the choices are to wait, negotiate an early migration with an incentive, or keep a minimal service running until those contracts end. Each choice has a cost that belongs in the margin score.

• 1Remaining subscription terms and auto-renewal dates.
• 2End-of-life or discontinuation clauses and required notice.
• 3SLA and uptime commitments during the wind-down.
• 4Source code escrow release conditions.
• 5Data return formats and deletion certification duties.
• 6Reseller, partner and integration agreements tied to the product.

Score the records, not only the product#

A product's records deserve their own score because they can outlive the product. Support tickets linked to engineering issues, code reviews, release notes and onboarding history document real software work, and they are usually deleted with the infrastructure unless someone decides otherwise.

Separate customer content, which follows customer contracts, from the company's own operating records. The second group can feed internal knowledge bases, documentation for a successor product, exit diligence or, where rights allow, licensing. The SourceX Enterprise Data Value Framework is one way to assess them, weighing drivers such as human-generated signal, domain expertise, recency, data cleanliness and rights against preparation cost and privacy burden.

• 1Which systems hold support, issue and code review history, and on which subscription plan?
• 2How much history is still accessible, and was any lost in past migrations?
• 3Do tickets link to issues, commits or releases, or stand alone?
• 4Which records contain customer content or customer code?
• 5Who holds admin access to export everything before shutdown?

Running the sunset itself#

Running a sunset well means giving customers a read-only period and a clear export route before access ends. Large vendors follow the same pattern: Autodesk first put BIM 360 Team and A360 Team files into read-only mode, letting customers copy them to Autodesk Docs or download them, and fully retired both products at the end of 2023. Atlassian, retiring the Bitbucket Cloud issue tracker and wiki, announced a removal date of August 20, 2026 and offered two routes for issues: a zip export or a move into Jira.

For an acquired product, the sequence is announcement, migration offer, read-only period, final export and decommissioning. Archive the company's own records before infrastructure shutdown, and keep an index of what was retained, returned and deleted, with the contract basis for each.

Give one person ownership of the sunset runbook, from the first customer notice to the final deletion certificate. Most failures at this stage are handoff failures: a cloud account closed by finance before engineering finished the export, or a help desk plan downgraded and its history lost.

Illustrative: three overlapping products in one vertical#

Illustrative: a fictional vertical software group owns three products for commercial cleaning companies after successive acquisitions: a cloud scheduling platform, an older desktop job costing system and a small customer feedback tool. The group scores all three in one review.

The cloud platform scores strongly on revenue and margin. The desktop system has loyal customers but high technical risk and heavy overlap, so it is set to merge into the cloud platform as contracts renew. The feedback tool scores weakly on everything except its records: a long history of customer feedback linked to support resolutions.

The group sunsets the feedback tool, gives customers a read-only export period, and archives its internal support and engineering history with an index before shutting down hosting. The decision on further use of that archive is made separately.

How SourceX supports rationalization decisions#

SourceX can supply the record-value input to a rationalization review. A metadata-only fit check indicates whether a product's support, issue and code review history is a licensing candidate, without moving files, so the group can score that criterion with evidence rather than guesswork.

If the group proceeds, the SourceX five-step transaction, Supply, Rights, Preparation, Approval and Delivery, runs separately for that product's records, and the operating company approves every step. Data is licensed, not sold, so the group keeps ownership after the product itself is gone.