An acquisition that disappoints is rarely an obviously bad deal.
The target was sound. The thesis was credible. The model worked.
Then nobody decided how the combined business would actually be run.
Twelve months later the expected savings have not appeared, growth has slowed and the people who made the business worth buying have started leaving.
Across the twelve acquisitions our founder has integrated, the recurring problem has usually not been the purchase price. It has been ambiguity about how the acquired company should operate after closing.
Why do acquisitions fail to deliver?
Because the integration type is left implicit. Every workstream then chooses one for itself.
Absorbing a company, leaving it autonomous and building a platform are three different jobs with three different plans.
When diligence and reality do not match, a fourth job is imposed. Stabilize the business before pursuing any of them.
Finance assumes one answer. Sales assumes another. Technology may assume a third.
The first year is often spent pulling against each other.
The failure is rarely visible at the time.
Everyone is busy, the reporting is being built and the disagreement looks like ordinary post-close friction.
It surfaces when the numbers in the model do not arrive.
Clayton Christensen and his co-authors made a related argument in The New M&A Playbook: acquisitions intended to produce economies of scale have to be managed differently from acquisitions intended to create a new platform for growth.
That is the strategic distinction. The four integrations translate it into the operating decisions somebody has to make after the deal closes.
What are the four integrations?
Three of them are destinations a buyer chooses. The fourth is a state you get put into.
1. Consolidation, absorb it
One brand, one system, one team.
Best when the thesis rests on cost savings, shared capabilities or simplification.
The value comes from removing duplication.
The risk is eliminating customer relationships and specialist knowledge before they have been transferred.
2. Leave autonomous, protect it
Integrate reporting, money and governance.
Leave the operating model alone.
Autonomy is not neglect. Performance expectations, governance, capital allocation and decision rights still have to be explicit.
Best when the business was bought because it works and the people are the asset.
The value is continuity.
The risk is drift. Autonomy erodes by accident, one well-meant request at a time, until it has been half-absorbed by nobody’s decision.
3. Roll-up, build the machine
This one will not be the last.
So the job is not integrating a company. It is building the model that integrates the next five.
Best when more deals are already in the pipeline.
The first deal should establish the reporting model, systems standards, decision rights, leadership structure and the rules for what stays local.
Standardize the right things once and each later deal should be faster, less disruptive and less dependent on improvisation.
The risk is over-engineering a platform before the first acquisition has proven out.
4. Fix the mismatch, stabilize it
Diligence missed something and now you own it.
The business, the management team or the operating reality does not match what the deal assumed.
This is not a destination. It is a recovery state.
Establish control, find out what is actually true, stop the deterioration.
Only then decide which of the other three you were really running.
Three integration types are chosen. The fourth is imposed.
Knowing which situation you are in determines the rest of the work.
Can different parts of the business use different models?
Yes, and on most deals of any size they should.
A buyer might preserve the brand and sales team, consolidate finance and payroll and migrate technology slowly.
What matters is that one type is primary and every exception is deliberate.
A named exception is a decision. An unnamed one is two departments quietly disagreeing about whether this is an absorption.
Why does integration planning start at diligence, not at close?
Because that is when operational findings can still change the transaction.
After the purchase agreement is signed, the same discoveries become problems to solve with less standing and at greater cost.
During diligence, operational findings can still change:
- A dependency on the seller becomes a transition-services obligation with a defined scope.
- A fragile management team becomes a retention package rather than a surprise in month three.
- Weak data quality changes the representations, the timeline or the price.
- An unclear handover becomes a written obligation on the seller rather than an understanding.
- A thin working-capital position changes what you fund at close, not what you discover after.
That is why operational due diligence and integration are the same engagement viewed from either side of the signing date.
They are not two services that happen to be sold by the same firm.
What does a failing integration look like in the first two weeks?
It looks like nothing much, which is the problem.
The first signs are rarely financial. They are small operating contradictions.
- Two people give a customer different answers about what happens next.
- A decision that used to take a day waits for a meeting nobody has scheduled.
- The acquired team hears about a change from a customer rather than from you.
- Somebody starts quietly keeping a shadow spreadsheet because the reporting does not agree.
- The best person in the acquired business takes a recruiter call and does not mention it.
None of these appears as a red status on a monthly report.
Months later they appear as customer attrition, turnover, margin erosion and growth that stopped.
Who should run the integration?
One person has to be unmistakably accountable for the integration.
On a larger deal that is usually a full-time integration lead.
On a smaller one it can be a senior operator with protected capacity.
What it cannot be is an informal responsibility shared among executives who are still expected to run the existing business.
An integration is a temporary organization with its own workstream leads, decisions and clock.
It runs alongside two businesses that still have to serve customers on Monday.
It needs one page everybody has read, a name against each workstream and a status rhythm that makes slippage visible while it is still cheap.
Most integration advisors deliver a plan. KPI helps run it.
We establish the workstreams, decision rights, owners and operating cadence, then work alongside your team to keep decisions moving and risks visible.
The approach is built on twelve acquisitions our founder has integrated, plus the operating work KPI has done inside owner-led businesses since 2022.
The decision that has to be made before closing
Before signing, the leadership team should be able to answer five questions the same way:
- Are we consolidating this business, leaving it autonomous, building a roll-up platform on it or stabilizing a mismatch?
- Which functions are deliberate exceptions to the primary type?
- Who has the authority to run the integration, and what capacity has been protected for them?
- Which diligence findings have to become closing conditions, retention actions or transition obligations?
- What has to be true by day 30, day 90 and the end of year one?
Different answers around the table mean the integration has already started with a problem.
That is a cheaper thing to discover now than in month five.
Have a target under review?
The free Deal Risk Check identifies the operating assumptions most likely to create problems after closing, while there is still time to change the deal or the integration plan.
Or see how we work across the whole buy side, from evaluating a target through to running the integration.
Common questions
Can different functions use different integration types?
Yes, and on most deals they should.
Preserve the brand, consolidate payroll, migrate systems slowly.
The requirement is that one primary type is named and the exceptions to it are deliberate, rather than each function quietly choosing for itself.
We have already closed. Can a failed integration be reset?
Yes, and this is when most people call us.
It usually costs more than starting at diligence, because the work is unpicking decisions made by default.
The first move is the same either way. Work out which integration you are actually running, then tell everyone so they stop pulling in different directions.
Who should lead an integration in a smaller company?
Somebody senior with genuinely protected capacity, not a full-time hire the deal cannot justify.
What matters is that one person owns it and has authority.
The work cannot be added on top of a job they are already doing in full.
How long should an integration take?
The decisions and accountability should be settled before close.
Customer, people and reporting stabilization usually comes first.
Systems and operating-model changes can then run for many months.
The aim is not to declare it finished quickly.
It is to keep value realization and unresolved risk visible from the start.
