A merger is the most expensive decision most leaders will ever sign off on, and one of the most likely to disappoint. The maths of a deal can look airtight while the logic underneath it is a story told to justify a price someone already wanted to pay. This guide is about telling the difference, the real reasons companies buy companies, the process a deal moves through, and the small set of questions that separate a sound rationale from an expensive one.

The quick version

  • Rationale comes first, price comes last. A deal is only as good as the reason for it, and "we need to grow" is not a reason.
  • Most acquisitions don't pay. Study after study puts the failure rate at 70–90%; the value, when there is any, usually goes to the seller, not the buyer.
  • The two failure machines are overpayment and over-promised synergies, especially revenue synergies, which rarely show up as forecast.
  • The process exists to test the story. Strategy, screen, valuation, diligence, negotiation, close, integration, each stage is a chance to kill a bad deal cheaply.

The idea in depth

Strip away the jargon and an acquisition is one company deciding it is cheaper, faster or safer to buy a capability than to build it. That capability might be customers, a product, a technology, a team, a geography, or simply scale. A merger is the same logic dressed as a partnership. Either way, the buyer pays a premium over what the target is worth on its own, and the entire bet is that the two businesses are worth more together than apart. That gap, the extra value created by combining, is the synergy. It is also where deals go to die.

Why companies buy: the rationale is the deal

Clayton Christensen and colleagues, writing in Harvard Business Review (2011), make a useful cut: companies acquire for one of two fundamentally different reasons. One is to improve the business they already have, buy a competitor, strip out duplicated cost, gain pricing power, fill a hole in the product line. These deals are easier to value because the gains are concrete and the buyer understands the business. The other is to reinvent the business, buy into a new market or model the buyer doesn't yet understand. These can be transformative, but they are far harder to price, because you're paying today for a future you can't yet see clearly.

Christensen's team also report the number every leader should keep in view: "study after study puts the failure rate of mergers and acquisitions somewhere between 70% and 90%." That isn't a reason never to do a deal. It's a reason to know exactly which kind you're doing and why. Here's the test worth applying before anyone opens a spreadsheet: write the rationale in one sentence a sceptical board member could repeat back, "we are buying X to get Y, which we cannot build ourselves for less by [date]." Can't write that sentence? Then you don't have a deal. You have an urge.

A bad acquisition is rarely a maths error. It's a story that was never stress-tested.

Why most deals don't pay the buyer

The failure rate has a specific shape. Robert Bruner's survey Does M&A Pay?, a synthesis of more than 100 scientific studies from 1971 to 2001, found that M&A creates value in aggregate, but the split is lopsided: target shareholders capture most of the gain through the premium, while acquiring-firm shareholders on average earn roughly zero adjusted return. The combined entity is often worth more; the buyer just hands most of that surplus to the seller on day one by overpaying.

Why overpay? Richard Roll's "hubris hypothesis" (Journal of Business, 1986) gives the cleanest explanation. In a competitive auction, the winner is frequently the bidder who is most wrong in the optimistic direction, the "winner's curse." Add the personal momentum of a CEO who has decided this deal will define them, and the bid drifts above any defensible value. The discipline that should stop this, walking away, gets harder the more time, ego and advisory fees are already sunk.

The counter is dull and it works: set a walk-away price in writing before negotiations heat up, and tie it to a value the target has on its own plus only the synergies you'd bet your own money on. One caveat worth stating plainly, these are averages across thousands of deals and decades. They tell you the base rate, not your odds. Disciplined acquirers, particularly serial buyers of related businesses they understand, do beat the average consistently. The statistics are a warning, not a verdict.

Synergies: where the rationale meets reality

Synergy is the word that launches a thousand bad deals, because it's where optimism hides. The pattern is well documented by McKinsey: acquirers are reasonably good at estimating cost synergies (the duplicated overhead you can remove) and reliably bad at estimating revenue synergies (the extra sales you'll supposedly make together). McKinsey's analysis found roughly 70% of deals miss their revenue-synergy targets, and that buyers routinely ignore dis-synergies, the customers and staff a merger drives away. Cost synergies are subtractions you control; revenue synergies depend on customers and competitors who didn't agree to your plan.

Price the deal, then, on cost synergies you can largely deliver alone, and treat revenue synergies as upside you don't pay for. If the deal only works with the revenue synergies, it probably doesn't work. That's the rule, and it has an exception: some of the best strategic deals, entering a new market, combining complementary products, are revenue stories by nature. The answer isn't to ban them; it's to fund them like venture bets, with explicit owners and milestones, rather than book them as certainties in the valuation. (Our companion piece on synergy assessment & realisation goes deeper here.)

flowchart TD
    A(["Why are we buying?"]) --> B(["Improve the
business we have"])
    A --> C(["Reinvent into a
new business / market"])
    B --> D(["Mostly cost synergies
· easier to value · pay for these"])
    C --> E(["Mostly revenue synergies
· hard to value · treat as upside"])
    D --> F(["Set a walk-away price
before you fall in love"])
    E --> F
Two rationales, two very different risk profiles, and one shared discipline. Leaders Loop

A worked example

Illustrative figures, a composite, not a real company. Northwind, a £400m logistics software firm, wants to enter parcel-locker hardware. Building it would take three years; Loxby, a £60m specialist, already has the product and 40% market share. Loxby is worth about £80m on a standalone discounted-cash-flow basis. Northwind's deal team models £15m a year of cost synergies (shared sales, removed duplicate admin) and £25m a year of revenue synergies (cross-selling lockers to Northwind's 2,000 software customers). On that combined number, the model justifies paying £160m.

Apply the discipline above. The revenue synergies assume Northwind's software buyers want hardware from the same vendor, untested, and exactly the kind of "we'll sell more together" story McKinsey finds usually disappoints. Strip it out. Now the deal is worth standalone value plus the £15m of credible cost synergies, call it a defensible ceiling near £110–120m, not £160m. The walk-away price is set there, in writing, before the auction. If a rival bidder pushes past it, Northwind loses the asset but avoids the £40m+ overpayment that Roll's winner's curse predicts. The revenue synergies aren't abandoned, they become a funded post-deal programme with a named owner and quarterly milestones, upside the company didn't bet the price on.

The lesson generalises: the rationale ("buy, don't build, a product we'd take three years to ship") is sound. The discipline is refusing to pay today for tomorrow's optimism.

flowchart LR
    A(["1 · Strategy
build vs buy?"]) --> B(["2 · Screen & target
shortlist, approach"])
    B --> C(["3 · Valuation
standalone + synergies"])
    C --> D(["4 · Due diligence
verify the story"])
    D --> E(["5 · Negotiate
price, terms, walk-away"])
    E --> F(["6 · Sign & close
approvals, completion"])
    F --> G(["7 · Integrate
where value is won or lost"])
The standard deal lifecycle. Each stage is a cheap chance to kill a bad deal before the next one gets expensive. Leaders Loop

Frequently asked questions

What's the difference between a merger and an acquisition?

Legally and practically, less than the words suggest. An acquisition is one company buying control of another; a "merger" is usually an acquisition framed as a union of equals for political and cultural reasons. In almost every combination one party is effectively in charge. Treat "merger of equals" with mild suspicion, someone is the acquirer, and pretending otherwise tends to slow the integration where most of the value is actually won or lost.

If 70–90% of deals fail, why do companies keep doing them?

Partly because the base rate hides real winners, disciplined, repeat acquirers of businesses they understand beat the average reliably. Partly incentives: growth-by-acquisition is faster and more visible than slow organic growth, and the people who champion deals are rarely the ones still around to own the disappointment. The failure rate is an argument for discipline, not abstinence.

What actually kills most deals after signing?

Two things, in order: paying too much (so there's no room for error), and botched integration, clashing systems, departing key people, and cultures that never knit together. The synergies live in the integration; a great rationale executed badly still loses money. The deal isn't done when it's signed; that's roughly when the hard part starts.

How do I tell a good rationale from a bad one in a board pack?

Look for three things. One: a rationale stated in a single, specific sentence ("buy X to get Y by date Z"), not a list of vague benefits. Two: a valuation that works on cost synergies alone, with revenue synergies marked as upside. Three: a pre-agreed walk-away price. If any of the three is missing, the deal is being sold to you rather than tested for you.

Is it ever better to partner than to buy?

Often. If you need a capability but not control, a partnership or joint venture gets you most of the benefit without the premium, the integration risk, or the irreversibility. Acquisition is the right tool when you need ownership, of the customers, the IP, or the team, and when integration genuinely creates the value. See joint ventures, alliances & strategic partnerships for the alternatives.

Related in the Toolkit

Where to go next