Two Estates, One Company: An Activation-First Due Diligence Playbook for M&A
A merger is announced as a single event and absorbed as a thousand small reconciliations. Among the least visible, and among the most expensive, is the software estate. Two companies arrive at closing with two identity providers, two contract portfolios, two sets of entitlements, and two very different answers to the question of how much of that software anyone is actually using.
Financial due diligence will examine the contracts. Legal will examine the transfer clauses. Neither discipline, on its own, will determine whether the combined entity is paying for seats that half its employees never activated. That determination sits in the seam between software asset management and integration planning, and it is where a meaningful share of synergy value is either captured or lost.
The characteristic M&A software problem
The arithmetic is simple and unkind. If the acquirer runs a productivity suite with 4,000 assigned seats and the target runs a competing suite with 2,500, the combined organization does not need 6,500 seats. It needs whatever the merged population actually uses — and it will likely be closer to 4,200.
Most of that duplication is invisible at signing because the target's contracts look reasonable in isolation. Each line item has a business justification attached, each renewal has an owner, and each invoice has been approved. What financial diligence rarely surfaces is the activation layer: how many of those licensed seats were provisioned, how many were registered, and how many were used in the past ninety days. A target can be fully compliant and still be carrying a large dormant estate.
Step one: build the activation ledger before the contract ledger
The instinct is to start with contracts, because contracts have dollar values attached. Resist it. Contracts tell you what the target is entitled to buy. Activation data tells you what the target is entitled to stop buying.
Request, at minimum, the following for every application above a materiality threshold:
- Contracted quantity, entitlement model, and true-up mechanics
- Provisioned account count in the tenant or license portal
- Registered or activated account count, with the activation date of each
- Thirty- and ninety-day active user counts
- The reporting source for each figure and the date it was pulled
- Any product with fewer than three named internal owners
The last item is diagnostic rather than quantitative. Software with no owner tends to be software with no activation discipline, and it is the category most likely to have drifted.
Step two: ask questions that produce evidence, not assurances
Targets answer diligence questionnaires carefully. Questions about satisfaction invite narrative; questions about counts require exports. Frame the request accordingly.
Ask how activation is measured today and who produces the number. Ask whether the measurement is based on the vendor's portal, SSO logs, or an internal discovery tool — and accept only one of those as an answer. Ask which applications were purchased in the last twenty-four months and what percentage of those entitlements have been activated. A low figure there is not a moral failing; it is a forecast of the consolidation opportunity.
Ask, too, about the failure cases. Which deployments stalled, and why? A target that can describe three stalled rollouts with specificity is giving the acquirer the integration map for free.
Step three: identify entitlements that will not travel
The commercial risk in M&A software is not only duplication. It is transferability. Several classes of entitlement routinely fail to survive a change of control:
Entity-bound cloud agreements. Enterprise agreements are frequently executed against a specific legal entity with defined affiliates. A new parent company may not be an affiliate under the original definition.
Non-transferable or change-of-control clauses. Some licenses explicitly require vendor consent or terminate on assignment. These clauses are usually short and frequently located in the general terms rather than the order form.
Reseller and marketplace purchases. Keys and entitlements purchased through a reseller may be tied to that reseller's agreement, complicating renewal and sometimes support.
Perpetual licenses with lapsed maintenance. The license survives; the right to update, patch, and receive support may not. An unmaintained perpetual license can become a security liability rather than an asset.
Volume licensing with organizational eligibility rules. Certain US public sector and education programs impose eligibility conditions that a commercial acquirer may not satisfy post-closing.
Each of these should be logged with the contract citation, the risk rating, and the remediation path — renegotiation, re-purchase at a negotiated rate, or planned migration.
Step four: price the duplicate estate into the deal model
Synergy models typically assume headcount rationalization and facilities consolidation. Software is often treated as a variable cost that will fall with headcount. It will not, at least not automatically. Licenses are reduced at renewal, not at termination, and only when someone negotiates the reduction.
A defensible model identifies the overlapping applications, estimates the duplicative spend, applies a realistic capture rate by renewal date, and assigns each opportunity to a named integration owner. Where an overlap will not resolve until a contract term ends, that timing belongs in the integration budget rather than in the unallocated optimism of year two.
Step five: sequence the first hundred days around activation
Integration teams tend to sequence by system criticality. A software-specific sequence is more useful: consolidate identity and provisioning first, then collapse duplicate entitlements at the earliest permissible renewal, then migrate the capabilities that employees in the target company actually used.
That last clause carries the risk. When a tool is retired, the users who relied on its distinctive features need somewhere to go. If the surviving platform offers an equivalent capability that requires activation — a module, an add-on, a configuration step — the migration is not complete when the account is created. It is complete when the feature is enabled and the user has used it once.
The judgment call
The question in the title has no universal answer. Some licenses survive a merger, some do not, and some survive contractually while being economically pointless. What an acquirer controls is whether the answer is discovered during diligence or during the first renewal cycle after closing. The first is a negotiation input. The second is an invoice.