Price the Portfolio: Application Overlap Belongs in the Deal Model, Not the Cleanup Plan
Photo: executives reviewing merger documents and laptop in conference room, via collegefootballnetwork.com
In most acquisition models, technology appears twice: as a synergy line item and as a footnote about integration risk. That placement is the problem. Application overlap is not a downstream logistics question. It is a set of cash flows, contractual obligations, and revenue dependencies that either support the purchase price or undermine it — and it is measurable during diligence with far more precision than most deal teams assume.
Diligence Stops at the Balance Sheet
Commercial diligence maps customers and churn. Financial diligence maps quality of earnings. Technical diligence, when it happens at all, tends to answer a narrow question: does the platform work, and will it keep working? Rarely does anyone ask the question that determines integration cost — how much of what this company runs is duplicated inside our own estate, and what would it take to stop paying for both?
The result is a familiar pattern. Thirty days after close, a working group convenes, opens two architecture diagrams, and finds two ERPs, three CRMs acquired through prior deals, four overlapping data warehouses, and a customer master that disagrees with itself across every one of them. The integration budget is then written in a meeting where nobody has negotiating leverage left.
Where Overlap Actually Hides
Overlap is rarely visible in a system inventory, because inventories list systems rather than functions. Duplication lives in the spaces between them:
- Multiple instances of the same platform, often from earlier acquisitions, each with its own chart of accounts, tax configuration, and reporting hierarchy.
- Shadow subscriptions purchased by business units on departmental cards, invisible to IT and outside any master agreement.
- Customer data replicated across three systems of record, with no agreed definition of which one wins when they conflict.
- Embedded licenses inside larger contracts — analytics seats bundled into a data platform, workflow engines bundled into a service desk — that cannot be separated, transferred, or retired independently.
- Integration middleware built twice, once on each side, connecting the same pair of functions through different routes.
The last item is the one most often missed and most expensive to unwind, because middleware accumulates business logic that nobody owns and everybody depends on.
Three Numbers That Belong in the Model
Consolidation value. For each duplicated function, estimate the run-rate cost of the instance being retired — licenses, infrastructure, support staff, and the integration glue holding it in place. Then subtract the cost of migration, which is almost always larger than expected and almost never zero. The difference, discounted across a realistic timeline, is what consolidation is actually worth. Present it as a range, not a point estimate, and be explicit about the assumption that drives the spread: how long the two organizations can tolerate parallel operations.
Contract exposure. Read the change-of-control, assignment, and termination provisions in every material technology agreement. Some require notice; some require consent; some trigger renegotiation; a few allow the vendor to terminate. Quantify the exposure as the incremental cost of the worst plausible outcome — forced migration, early termination fees, or re-pricing at renewal without competitive tension. This number frequently exceeds the projected consolidation savings, which is precisely why it belongs in the model rather than in a post-close issue log.
Migration risk, expressed as revenue. The useful question is not how long migration takes but which revenue depends on the system being migrated. If the billing platform serves the top twenty accounts, the migration window is a customer retention risk. If the warehouse management system runs a single distribution center, downtime has a hard cost per hour. Translate technical sequencing into revenue-at-risk and the conversation changes from an engineering debate to a commercial one.
The License Ledger Nobody Reads
Software agreements are the least-read and most consequential documents in most deals. A few provisions deserve a line in the model:
- True-up and minimum commitment clauses, which convert unused capacity into a liability at renewal.
- Audit rights with retroactive true-up, which can turn an under-licensed estate into an unbudgeted cash call.
- Non-transferable entitlements, common in enterprise agreements, which mean the acquired company's licenses cannot simply be absorbed.
- Bundled terms with unequal expiry, where retiring one component forfeits pricing on another.
A license ledger — every material agreement, its expiry, its assignment terms, its minimum commitment, and its exit cost — takes weeks to build and routinely reprices a deal.
A Pre-Signing Sequence
Run the work in the order that preserves optionality. First, inventory by function rather than by product, so duplication becomes visible. Second, build the license ledger and flag every change-of-control provision. Third, estimate migration cost and timeline for the three highest-value consolidations only — a full plan is unnecessary and a full plan is what people hide behind. Fourth, express the result as an adjustment to price or as a specific indemnity or escrow term.
That last step is the point. Diligence findings that stay in a document become post-close arguments. Findings converted into price, escrow, or closing conditions become someone else's obligation. The difference between those outcomes is decided long before the integration team is assembled — in whether the portfolio was ever treated as part of the thesis.
An application estate is not a technical artifact that happens to sit inside a company. It is the operational shape of the business, with contracts attached. Price it that way.