Every deal model has a line that does a lot of quiet work: synergies. It is the number that turns an expensive acquisition into a sensible one, the number that justifies the premium over the target's standalone price. And it is the number most likely to be wrong. Treating that line as a confident forecast, rather than a list of specific actions someone has to take, by a date, at a cost, is how acquirers talk themselves into overpaying.

The quick version

  • A synergy is value the combined company creates that neither side would have produced alone, not the value either was already going to deliver. The acquisition premium is a debt you take on; synergies are how you repay it.
  • Assessment means estimating each synergy bottom-up (what action, who owns it, when it lands, what it costs to capture) and discounting it, not pasting a percentage of combined revenue into the model.
  • Cost synergies are far more reliable than revenue synergies. McKinsey found the biggest estimation errors are on the revenue side, where roughly 70% of deals miss their target.
  • Realisation is a tracked operating programme, not a hope. Name owners, set baselines, run it on a clock, the first 12–18 months after close decide most of the outcome.

The idea in depth: a synergy is a delta, not a destination

The cleanest definition comes from Mark Sirower, whose 1997 book The Synergy Trap remains the standard reference. Synergy, he argues, is the increase in performance beyond what the two firms were already expected to achieve independently. That word "beyond" is the whole game. If a target was going to grow 6% next year anyway, none of that 6% is synergy. Synergy is only the extra you unlock by combining, the call-centre you can close because two became one, the customers you can cross-sell because of the deal and not in spite of it.

Why be so strict? Because the premium demands it. When you pay above market price, Sirower frames it as taking on a debt: the premium is a sum the combined business must now earn back through performance it could not have produced apart. Vague, generous synergy numbers make any premium look affordable. Specific ones tell you whether the price actually clears. Here is the test I would apply before defending any number: write the sentence underneath it, "we will save $X by doing this specific thing, owned by this person, live by this date." A synergy you cannot write that sentence for is not a synergy. It is a wish.

"Synergy is the increase in performance beyond what the two firms were already expected to achieve independently.", Mark Sirower, The Synergy Trap

Cost synergies are real; revenue synergies are mostly hope

Not all synergies deserve the same confidence, and pricing them as if they did is the most common modelling error. The evidence is lopsided. In McKinsey's analysis of its deal database, summarised in "Where mergers go wrong," the largest estimation errors appear on the revenue side, with almost 70% of deals failing to achieve the revenue synergies they expected. Cost synergies fare better, but are still overstated by 20% or more in about a third of cases.

The asymmetry has a simple cause. A cost synergy is mostly inside your control: you can decide to consolidate two data centres or remove a duplicate finance team. A revenue synergy depends on a customer who never signed the deal and owes you nothing. Cross-sell assumes the other side's salesforce will sell your product, that the customer wants the bundle, and that nobody walks during the disruption. Each of those is a maybe, multiplied. That is why revenue synergies typically take years to surface, not quarters, and why disciplined acquirers discount them hard, or keep them out of the price entirely. In practice that means splitting your synergy register into two columns, "cost / control-side" and "revenue / customer-side," then applying a heavier haircut and a longer timeline to the second. The rule I would hold to: never let revenue synergies be the load-bearing part of your premium.

The honest limitation: these are averages across many deals, mostly large public transactions, and they describe a tendency, not your deal. A well-run integration in a business with genuine overlap can beat them; a sloppy one can do far worse. The numbers are a reason to be sceptical and to demand bottom-up evidence, not a formula you can plug into your own model.

flowchart TD
    A(["Gross synergy idea
e.g. 'cross-sell to their base'"]) --> B(["Bottom-up estimate:
action, owner, date"])
    B --> C(["Cost to capture
severance, systems, advisers"])
    C --> D(["Risk haircut
+ probability of capture"])
    D --> E(["Discount to present value"])
    E --> F(["Net synergy that can
justify the premium"])
From slogan to a number you can defend: a synergy survives only if it can pass every gate. Leaders Loop

Realisation is an operating programme on a clock

A synergy assessed perfectly still has to be captured, and capture is where most of the value leaks away. Two ideas do the heavy lifting here. The first is governance with a baseline: every synergy gets a named owner, a measurable baseline (you cannot prove a saving against a number you never recorded), and a tracker that reconciles claimed savings back to the actual profit-and-loss, not to a slide. The second is the clock. McKinsey's integration research is blunt about timing: the first 12–18 months after close shape the result, and momentum compounds. In its work, 79% of acquirers that outperformed their peers 18 months after close were still outperforming at three years, while only a small minority of early underperformers ever turned it around ("Four keys to merger integration success").

There is a legal wrinkle worth knowing, because it shapes how early you can plan. Until a deal closes, the two firms are still competitors, and antitrust rules limit what commercially sensitive data they can share. The standard fix is a clean team (or clean room): a small, ring-fenced group, often outside advisers plus a few named insiders, who can see sensitive pricing and cost data and hand back only approved, aggregated conclusions, leaving a documented audit trail. It lets integration planning start before close without breaking the rules, and EY and McKinsey both report it pulls real synergy capture forward by months. So the practical sequence is straightforward: stand up a clean team the moment a deal is signed, walk in on day one with a baselined synergy register and named owners, and review it on the same cadence you review revenue. If it is not on the operating calendar, it will not happen.

flowchart LR
    A(["Sign"]) --> B(["Clean team:
pre-close planning"])
    B --> C(["Close /
Day one"])
    C --> D(["0-18 months:
capture window"])
    D --> E(["Track vs P&L
baseline, owners, cadence"])
    E --> F(["Synergies banked,
or quietly lost"])
The capture window opens at signing, not close, and most of it is spent within eighteen months. Leaders Loop

A worked example

The figures below are illustrative, chosen to show the method rather than to describe a real deal. Suppose a mid-sized logistics group acquires a smaller regional rival for £240m, a £40m premium over its standalone value. The deal team books £15m a year of run-rate synergies to justify that premium.

On the cost side: closing one of two overlapping depots and removing duplicate back-office roles is worth, bottom-up, about £9m a year, owners named, mostly inside management's control. But it costs roughly £6m one-off to capture (severance, lease exit, systems migration), and it lands over 18 months, not on day one. On the revenue side, the model assumed £6m a year from cross-selling each firm's customers the other's services. Pushed for the bottom-up case, the team can only evidence about £2m within three years, the rest depends on customers who may not want the bundle and reps who have to be retrained and incentivised.

So the defensible, risk-adjusted figure is closer to £9m of solid cost synergy plus £2m of plausible revenue synergy, net of a £6m capture cost, not the headline £15m. That is the number that should have set the premium. The discipline did not kill the deal; it priced it. And it handed the integration team a register of specific, owned actions to run against from day one, rather than a single optimistic total to be quietly missed.

Frequently asked questions

What's the difference between synergy assessment and just valuing the target?

Valuation tells you what the business is worth on its own; synergy assessment tells you what extra the combination creates, and therefore how much premium you can pay without destroying value. They feed each other, standalone value plus net synergies sets your walk-away price, but they are different questions, and conflating them is how acquirers justify overpaying.

Why are revenue synergies treated so sceptically?

Because they depend on customers, not decisions. You can order a cost out; you cannot order a customer to buy the cross-sell. McKinsey's data shows roughly 70% of deals miss revenue synergies, and they typically take years to appear. Treat them as upside, not as the foundation of the price.

What does "cost to capture" mean and why does it matter?

It is the one-off spend needed to unlock a recurring saving, severance, system integration, site closure, adviser fees. A £10m annual synergy that costs £25m to capture is a very different proposition from one that costs £2m. Modelling the gross saving without its capture cost flatters the deal and is one of the most common ways synergy numbers mislead.

When should we start planning to realise synergies?

At signing, not at close. Stand up a clean team so you can plan against real data without breaching antitrust rules, and aim to walk in on day one with named owners and baselines. The capture window is largely the first 12–18 months, late starts rarely catch up.

Isn't "synergy" just a buzzword for layoffs?

Cost synergies often do mean removing duplicated roles, and pretending otherwise erodes trust. But synergy also covers procurement scale, shared infrastructure, faster product reach and genuine cross-sell. The honest version names which kind you mean, costs it, and tracks it, rather than using the word to wave the premium through.

Related in the Toolkit

Where to go next