Your company acquires a competitor of about a third its size. Both run their own ERP, CRM, identity provider, data warehouse and payments integration. The board wants "one company" within 18 months and synergy numbers were already published. How do you decide what to consolidate and what to leave alone?
Show the full answer Hide the answer
The deciding property
Not cost and not duplication. The deciding property per capability is whether a single shared instance is required for the business outcome the synergy number depends on. Two CRMs are tolerable for years if the sales teams serve different segments. Two general ledgers are not tolerable past the first consolidated statutory report, because the outcome is one set of numbers.
Choose per capability against the outcome, not system by system against a cost line. Two ERP estates in the mould of large SAP and Oracle installations are not two line items; each carries a configuration that encodes an operating model, which is why the ledger decision is a business decision wearing a systems costume. Any plan that starts with a list of applications and asks which to retire will consolidate the easy things and leave the hard, binding ones for month fourteen.
The sequence that follows
- Identity first, always. One directory, one single sign-on, one joiner-mover-leaver process. Nothing else can be consolidated while two populations cannot be granted access to one system, and every integration you build in the meantime will be rebuilt afterwards. This is also the piece that lets people work across the boundary in month two, which buys political room for the rest.
- Then the capabilities with a hard external deadline: statutory reporting, consolidated financials, regulatory returns, tax. These have dates set by someone outside the company, and they are the only ones where the eighteen-month target is not negotiable.
- Then the capabilities where duplication costs more than it saves — usually licences with a per-user or per-core metric, and anything whose specialist support is 2 people who cannot cover 2 systems.
- Explicitly defer the rest, in writing, with the condition that would change the decision. "Two CRMs until the sales organisations merge" is a decision. "CRM: TBD" is how it ends up on the critical path.
What each option costs
| Option | Gains | Pays |
|---|---|---|
| Consolidate onto the larger instance | One system, one team, licence savings | The acquired business absorbs all the change while also being integrated; the larger system's process becomes mandatory and may not fit |
| Consolidate onto the better instance | Best capability wins | The larger population migrates, which is the expensive direction and the one nobody budgeted |
| Coexist with an integration layer | Fast, low disruption, reversible | Permanent integration cost and a canonical-model problem; coexistence that is not time-boxed becomes the architecture |
| Retire one capability entirely | The only option that actually removes cost | Requires a business decision about which operating model wins, which is why it is usually avoided |
Only the last row removes run cost. The synergy number almost always assumes retirement while the programme delivers coexistence, and that gap is the standard failure of post-merger architecture.
Why copying another company's approach is a mistake
Published integration playbooks come from acquirers with a repeatable model: a standard target stack, a practised migration team, and acquisitions small enough to absorb whole. At a third of your size, the acquired estate is too big to absorb and too small to win, which is the hardest ratio. Serial acquirers also have something you do not: a decision already made about which operating model is canonical. Make that decision explicitly in month one, or every capability argument will relitigate it.
When this is the wrong answer
Coexistence beats consolidation when the two businesses genuinely serve different markets with different regulators, consolidation can destroy value: one instance means one configuration, one release cadence and one set of compromises. Say "two instances, permanently, and here is the integration we will fund" where that is true, and take the synergy number out of the plan rather than leaving it in as an unfunded assumption. A deferred decision with a stated condition is architecture. An undated intention to converge is a tax paid annually.