Red-Team Your Strategy Without Making It Political

A few years ago we sat in on the post-mortem of a deal that didn't quite work. It had been a large acquisition, well over a billion dollars, in an adjacent category. The integration was nine months in. The synergy case was visibly slipping. The strategy team was reviewing what could be learned.

About forty minutes in, someone asked, almost apologetically, what had happened to the red team report. There was a long pause. Then a slightly embarrassed exchange across the table, and the deputy CFO recalled that yes, there had been a red team, they had produced a memo about six weeks before close, the memo had identified — among other things — exactly the customer-retention failure mode that was now showing up in the integration numbers. The memo had been read. The memo had been thanked. The memo had been filed. The deal had closed on schedule.

We see this almost everywhere we look for it. Red-teaming, as a discipline, has roughly the right idea. Take a major commitment, assign a separate group to find the holes, run the case against itself before the irreversible signature is on the page. The logic is sound. The execution, in most companies, is theatrical. The red team is set up, the red team does its job, the red team's findings are politely received, and the decision proceeds as if the red team had not existed. By the third major commitment, the red team has been quietly defanged or disbanded.

This is not because the executives involved are venal. It is because the structure of red-teaming, as commonly implemented, makes the failure mode almost inevitable. The fix is structural. Four moves do most of the work.

Failure mode: the red team becomes "the people who tried to kill the deal"

The default setup is the problem. A major deal is being worked. A red team is assigned. The red team's mandate is, in effect, to argue the opposite case. They review the deal materials, generate counter-arguments, identify risks, and present their findings to the decision-making body.

Almost immediately, the dynamic settles into a courtroom shape. The deal team is the prosecution. The red team is the defense. The board is the jury. The red team's role is to win, which means defeating the deal. The deal team's role is to win, which means defeating the red team. The board's role, increasingly, is to adjudicate two competing performances rather than to receive analysis.

Two things follow from this dynamic, and both are bad. The first is that the red team becomes identifiable, by name, as "the people who tried to kill the deal." If the deal proceeds and works, the red team was wrong. If the deal proceeds and fails, the red team was unhelpful — they identified the risk but did not stop the bad decision; clearly the process is broken. The red team cannot win. The asymmetry produces, over time, a population of executives who do not volunteer for red team duty, because the role is career-negative regardless of outcome.

The second is that the deal team, knowing the red team is structurally adversarial, begins to defend rather than to absorb. Information that might help the red team is, at the margin, withheld. The deal memo gets slightly tighter, slightly less honest about its own internal disagreements, because the deal team is preparing not for analysis but for cross-examination. The red team, in turn, sharpens its tone to compensate. The dialectic ratchets toward performance and away from rigor.

This is not a culture problem. You cannot exhort your way out of it. It is a structural consequence of pitting two teams against each other in front of a decision-making body. The fix is to change the structure.

Fix one: rotate the role

The first structural change is to make red-team duty rotational, not a fixed function. The same executives should serve on red teams for some commitments and on deal teams for others, on a published rotation. Over a two- to three-year cycle, every senior leader should have served as both prosecutor and defendant on commitments of comparable size.

The effect is immediate and almost mechanical. An executive who knows they will be on the deal team for next quarter's acquisition cannot easily punish the red team that is reviewing this quarter's. The norms of red-teaming become self-enforcing because every member of the leadership group has either just been on a red team or is about to be on one. The "people who tried to kill the deal" label becomes uninhabitable, because next month those same people are running the deal.

Rotation also produces a quieter benefit: it improves the quality of the red-team work itself. An executive who has just spent a quarter as a deal-team principal knows where the soft spots in a deal memo tend to live — which assumptions are usually waved at, which integration claims are usually under-specified, which customer retention numbers usually don't survive contact with the field. A standing red-team function staffed by outsiders does not have this knowledge. A rotating red-team staffed by people who were recently building deals does.

Fix two: scope it tightly to specific assumptions

The second structural change is to scope the red team's work narrowly. The default mandate — "find the holes in this deal" — is too broad to be productive. It invites the red team to argue the case as a whole, which puts them in opposition to the deal team as a whole, which produces the courtroom dynamic.

The better scope is to identify, in advance, the two or three load-bearing assumptions on which the case actually depends, and to assign the red team to stress those specifically. Not "review the deal." Review the customer-retention assumption. Review the synergy timing assumption. Review the regulatory approval assumption. The red team's deliverable is not "should we do the deal" — that is the decision-making body's job — but "is this specific assumption defensible, and what would we need to see to be more confident in it."

The narrow scope changes the politics. The red team is no longer arguing against the deal. They are arguing about a parameter. The deal team is not defending its judgment. They are defending an assumption. The conversation moves from "yes / no" to "how confident, on what basis." That is a substantially less adversarial conversation, and a substantially more useful one.

It also forces the deal team, ahead of the red team's work, to identify which assumptions actually matter. This is the question they should have answered anyway. Most deal memos do not identify their load-bearing assumptions explicitly — the assumption that the case hangs on is buried somewhere in slide forty-seven, surrounded by twelve other assumptions of varying importance. Making the red-team scope explicit forces the prioritization upstream, which improves the original memo whether or not the red team finds anything.

Fix three: position it as a service to the deal team, not opposition

The third structural change is to reframe the red team's relationship to the deal team entirely. Not as an opposing function. As a service function. The red team's client is the deal team. The deal team is the consumer of the red team's analysis. The output is delivered to the deal team first, before any decision-making body sees it, with enough time for the deal team to absorb, respond, revise, and incorporate.

This sounds like a small change. It is not. It reverses the political flow. Under the adversarial model, the red team's job is to embarrass the deal team in front of the board. Under the service model, the red team's job is to make the deal team's memo to the board better — sharper, more honest about its assumptions, more rigorously stress-tested. The deal team's incentive is to absorb the red team's findings into their final memo, because the final memo is what the board will see, and a memo that has already absorbed the strongest counter-arguments is a stronger memo.

The board, when it eventually sees the deliverable, sees one document, jointly worked, with the deal team's case and the red team's most serious concerns both surfaced inside it. There is no courtroom. There is no winning side. There is one memo, with its weaknesses identified by the people best positioned to identify them, presented honestly. The board's job is, finally, to decide, rather than to adjudicate between two performances.

A red team that reports to the board is in opposition. A red team that reports to the deal team is in service. The first is theater. The second is insurance. The architecture determines which one you get.

This fix is the one we find hardest to install in practice, because it requires the deal team to accept the red team as a quality-improvement function rather than a threat. That acceptance, in turn, requires the leadership group to have norms that reward the deal team for absorbing uncomfortable findings rather than for batting them away. Those norms, in turn, require the rotation. The three fixes interlock.

Fix four: reward the red team for finding problems regardless of outcome

The fourth structural change is the one that ties the system together. The red team must be rewarded for finding real problems, regardless of whether the deal ultimately proceeds.

In most companies, this is exactly inverted. The red team is implicitly graded on whether the deal goes through. If the deal goes through and works, the red team was wrong. If the deal goes through and fails, the red team was unhelpful. If the deal doesn't go through, the red team killed it. Every outcome is bad for the red team. So the rational red team minimizes the strength of its findings, hedges its language, and gets credit for being collegial.

The fix is to reward the red team on the quality of the findings relative to what was knowable at the time. A red team that surfaced the customer-retention failure mode six weeks before close should be credited regardless of whether the integration team subsequently solved the problem, ignored it, or got lucky. The output is the analysis. The outcome is downstream of a hundred other things. Confusing the two — grading the red team on whether the deal worked — produces exactly the timid, ratifying red team that everyone privately complains about.

This is the same principle that has to underlie the annual decision audit and the broader question of how a leadership team scores its own work. Process and outcome are different things. The system has to be designed to reward them separately, or the process work will, eventually, stop being done.

What good red-teaming actually looks like

Done with the four fixes installed, red-teaming stops being theater and becomes — for the cost it incurs — almost certainly the cheapest insurance available to a leadership team contemplating a large commitment.

The shape is recognizable. A rotational pool of senior executives, drawn from outside the deal team but inside the firm. A narrow scope, focused on the two or three load-bearing assumptions identified by the deal team itself. A reporting line back to the deal team first, with the integrated memo going to the board afterward. A grading system that rewards the red team on the quality of its analysis, not on the verdict of the decision-making body or the outcome of the deal.

The output is not a competing case. The output is a strengthened original case, with its weakest assumptions clearly surfaced, its kill criteria sharpened, and its post-commitment review structure already half-written. The board receives a better document. The deal team retains its ownership. The red team retains its credibility. Everyone is still talking to each other on Monday. The premortem and the red team are complementary: the premortem identifies the failure modes; the red team stresses the assumptions on which the case depends. Done together, they cover most of the ground.

We have seen leadership teams adopt this architecture and stick with it. We have seen others adopt only the label, run an adversarial red team for two cycles, and quietly drop the practice when it produces too much friction. The difference between the two outcomes is not commitment or culture. It is whether the four structural fixes were installed at the start, or whether the team relied on goodwill to substitute for structure. Goodwill, in our experience, does not substitute. Architecture does.

If your last large commitment had a red team and the red team's findings ended up in the file rather than in the memo, that is the room we are usually called into. We are happy to walk through what changed and what didn't. The fix is structural and, once installed, durable. The cost of not installing it is paid downstream, in the deals that closed when they shouldn't have and the integrations that surprised everyone except the people whose memo never made it to the front page.