The Second Opinion Every M&A Model Needs

There is a room we have been inside many times. The lights are good. The model is on a wall-mounted screen. The bankers have flown in. The CFO is leaning forward. The CEO has been told, by three trusted people, that this is the deal that finally moves the company into the next category. The synergies cell is glowing green. The IRR clears the hurdle by a comfortable margin. Somebody — usually a junior on the deal team — has been asked to "pressure-test the model," and has done so by tightening a few formulas and re-running the sensitivity tab.

The room is about to make a nine- or ten-figure decision. And every single document on the table has been authored by someone with a financial or political interest in the deal closing.

This is the structural problem with M&A. Not bad analysts, not unsophisticated boards — structure. The bankers' model is built to support the transaction; they earn on close. The internal analysis is built to support the CEO, who has already said, in front of the board, that the deal is strategic. The target's diligence pack is built to support the target. The lawyers are paid to close. The advisors are paid to close. Nobody at the table is paid to honestly stress-test the case. The room is, in expectation, biased toward "go," and the bias is invisible because everyone in the room is acting in good faith inside the role they were hired to play.

What the literature already tells us

This is not a contrarian thesis. The academic and consulting record on M&A is unusually consistent. McKinsey's long-running post-merger work, repeated across cycles, puts the share of deals that fail to create value for the acquirer somewhere in the neighbourhood of seventy percent. Damodaran's work on synergy valuation makes the point even more pointedly: revenue synergies, in particular, are the assumption category most consistently inflated in deal models, and most consistently disappointing in delivery. The "bidder's curse" — the well-documented tendency of winners in competitive processes to overpay — is in the literature for any executive who wants to read it.

But the literature does not move deals. The literature is read once, in business school, and then forgotten in the heat of a process. What moves deals is the model on the screen, and the conviction of the people standing next to it.

The honest question in an M&A review is not "is the model right?" It is "who in this room was paid to find out it was wrong, and what did they say?"

The second opinion is the answer to that question. Done properly, it is not a re-audit of the bankers' arithmetic. It is a structured, independent, paid-to-disagree analysis of the case — not the spreadsheet — with three specific components.

Component one: priors on each major assumption

Most deal models are built deterministically. A revenue synergy number appears in a cell. It is sourced, somewhere in the appendix, to a "management estimate" or a "benchmarking exercise." The number is then carried, unchanged, through the rest of the model. The model treats it as fact.

A useful second opinion replaces each major assumption with a prior — a stated belief, with a stated confidence range, and a stated basis. Not "the model says revenue synergies are $180M by year three." Instead: "we believe revenue synergies are around $110M with eighty percent confidence between $60M and $170M, and here is the comp-set behind that — the seven most analogous deals of the last decade, the realized-versus-projected synergy ratio in each, and the reasons we think this one sits in the lower half of that distribution rather than the upper half."

This is more work than the deal model does. It is also the only version of the work that actually informs the decision the board is being asked to make. The board's question is not "what does the model show?" The board's question is "what should we believe, given everything that is knowable now?" Those are different questions, and the gap between them is where the overpayment lives.

In one engagement we ran, a strategic acquirer was preparing to close on a target where the deal case turned almost entirely on cross-sell synergies into the acquirer's existing customer base. The bankers' model had cross-sell at $90M of run-rate revenue by year two. Our prior, based on the acquirer's own historical cross-sell rates from three previous (smaller) acquisitions, was closer to $25M with eighty-percent confidence between $10M and $55M. The two numbers were not reconcilable. The decision the board then made was different — they closed, but at a meaningfully lower price, with the renegotiation justified in writing by the prior. The seller accepted, because the alternative was a broken process.

Component two: scenarios with explicit probabilities

Most deal decks contain a "base case, upside case, downside case" structure. We do not, in our practice, treat these as scenarios. They are the same forecast at three different intensities, usually generated by moving the same handful of input cells up and down by the same percentage. They are not independent worlds. They are mood lighting.

A real scenario is a world — a specified state of competitors, customers, regulators, talent, and macro conditions, internally consistent enough that someone reading it could tell, two years from now, whether they were living in it. A useful second opinion contains three to five such worlds, with explicit probabilities attached. Probabilities chosen by the analyst, defended in writing, calibrated against the acquirer's track record of past forecasts.

The probabilities matter more than the worlds. A deal that has a thirty-percent chance of producing a great outcome and a thirty-percent chance of producing a write-down is a structurally different bet from a deal that has a sixty-percent chance of producing a fine outcome and a ten-percent chance of producing a write-down. The expected value can be similar. The risk profile is not. The board, in our experience, is almost never shown the second framing — because the second framing requires the analyst to commit, in writing, to a probability of failure. That is uncomfortable. It is also the entire point.

Component three: a downside the model is allowed to surface

This is the most important component, and the one that requires the most institutional courage to install.

In every deal we have ever reviewed, there is at least one downside scenario that the deal team identified privately and did not put in the formal materials. We do not say this as criticism. The team is reading the room correctly: a slide in the board pack titled "How This Deal Could Lose $400M" will be read by the CEO as disloyalty, by the bankers as deal-killing, and by the deal lead as a career-limiting move. The institutional pressure on the materials is, by the time they reach the board, overwhelmingly toward smoothing.

A second opinion has to be commissioned in a way that is allowed to surface that downside. This means commissioned by someone other than the deal sponsor — typically the board's audit or risk committee, or in some structures the CFO acting independently of the CEO. It means engaged with a written brief that explicitly asks for the worst credible case, named and quantified, with the reasoning shown. And it means delivered into a meeting structured so that the downside cannot be characterized as "negative" or "deal-killing" — it can only be characterized as part of the case. The deal might still go forward. But it goes forward with the full picture in front of the people approving it.

A corp-dev team we worked with, after their first deal where this process was followed, told us the most useful artifact of the engagement was not the lower price they negotiated — though that paid for our fee several times over — it was the document. Eighteen months in, when one of the downside scenarios partially materialized, the board did not panic. They had read the scenario. They had a pre-committed plan. The integration team did not lose six months to a re-litigation of whether the deal had been a mistake. The deal had been understood, at the moment of commitment, to contain that risk. The work was already done.

What the second opinion is not

It is not a re-do of the bankers' model. The bankers' model is, in our experience, usually arithmetically fine. The numbers add up. The formulas reference the right cells. Re-auditing the spreadsheet catches almost nothing of consequence and produces a false sense of rigour. The second opinion lives one level up: at the level of what the model assumes, not what the model computes.

It is not a vote against the deal. Most deals we second-opinion close. Some close at a lower price. Some close with a restructured earn-out. Some close on the original terms with a sharper integration plan, because the act of producing the second opinion has surfaced the integration risks early enough to plan against them. A few do not close — and in those cases, the value of the work is the deal that did not happen, which is the hardest value to credit because it is invisible on the P&L.

It is not, finally, free. A serious second opinion takes weeks of work, costs real money, and inserts friction into a process that is usually being run on a banker's timeline. The friction is the point. Decisions of this magnitude should not be made on a banker's timeline. The cost of the second opinion is a small fraction of the cost of being wrong on the close, and the asymmetry — small certain cost now, large uncertain cost avoided later — is exactly the asymmetry every other part of the deal team has structural reasons not to see.

If your next deal review feels like the room described at the top of this essay — a model on a screen, a synergies cell glowing green, and nobody in the room paid to find out it is wrong — that is the room where a structured second opinion earns its fee. We are happy to discuss what one would look like, in your process, for the specific deal you have in front of you.