Decision Rights Matrix kit.

The Conversation You Are Avoiding

There is a specific conversation that operators in the $5M-$50M range have been avoiding for, in most cases, somewhere between eight months and four years.

The conversation has a recognizable shape. It happens — or fails to happen — between the operator and their leadership team, usually one direct report at a time. The substance is: who actually decides what in this operation. The shape is that the operator believes one answer (the org chart's), the leader believes another answer (the lived reality's), and the conversation that would surface the gap between the two has been deferred because both parties have been managing around the ambiguity rather than naming it.

The Decision Rights Matrix is the artifact that forces the conversation. Not by being a clever document. By being a structured forcing function that requires the operator and the leadership team to name, in writing, on a single page, who decides each consequential category of decision. The act of producing the matrix is the install; the matrix itself is just the residue.

This is why the kit's installation effort is "two working sessions" instead of "one document." The document is two hours of writing. The conversation underneath the document is two working sessions, often four, sometimes a full quarter of distributed effort. The cost is the conversation. The value is the resolution.

This essay is about the conversation. The kit guide covers the mechanics. This piece covers why the mechanics matter and what happens to operations that skip them.

The Pattern That Always Shows Up

Across the operations I have watched run the matrix install, three patterns surface with such reliability that they are functionally constants.

Pattern one. The CEO believes the leadership team has been operating with substantially more authority than the leadership team actually feels they have. The CEO points to the org chart, the role descriptions, the formal delegations. The leadership team points to the last several consequential decisions that all routed through the CEO despite being formally delegated. Both parties have evidence. The evidence is reconciled by the audit step — the Twenty-Decision Audit kit that the matrix install often runs alongside. The audit traces actual decision flow against the org chart's stated flow, and the gap that surfaces is almost always closer to the leadership team's read than to the CEO's.

Pattern two. The CEO has been overriding the matrix-equivalent of every consequential decision for some duration before the install is recognized as necessary. The overrides have been informal, often well-intentioned, sometimes invisible to the CEO themselves. The pattern is recognizable from the team's behavior: leaders who routinely escalate decisions that the org chart says they should own, escalations that produce the CEO's input "just to be safe," approvals that arrive faster from the CEO than from the formal owner because the team has learned that the CEO is the real decider. The matrix install is structurally a request that the CEO stop overriding for 90 days. Most CEOs experience this request as harder than they expected.

Pattern three. The leadership team includes one or two leaders who have been quietly carrying decisions outside their formal scope, and these decisions have generally been better than what the formal owner would have produced. The matrix install threatens to take these decisions away from the leaders who have been making them well. The threat is real but should not block the install — the right resolution is usually to formally re-scope the role to match the decisions the leader has been making competently, not to take competent decisions away in service of the org chart. The install is the moment when this re-scoping becomes possible because the alternative (continued unofficial carrying) has been named as a problem.

The three patterns are not exceptions; they are the install. Operators who think the matrix install is producing the matrix are missing what it is producing. The matrix is the byproduct. The structural rebalancing of who decides what is the actual deliverable.

The Decision That Defines Whether the Install Takes

The single decision that most predicts whether a matrix install will take and produce structural change is the CEO's commitment to defer to the matrix in the first 30 days after it ships.

The commitment sounds simple in the working session. It is structurally hard to execute. Within the first two weeks, the CEO will receive at least three decisions that the matrix routes to someone other than them, where the CEO's judgment disagrees with what they predict the named owner will decide. Each of these is an override opportunity. Each override the CEO takes in the first month teaches the leadership team that the matrix is decorative.

The discipline that produces real installs is the discipline of pushing the decision back to the named owner, even when the CEO would have decided differently. The named owner makes the call. The call may be worse than what the CEO would have made; usually it is similar; sometimes it is better because the named owner has context the CEO doesn't. The CEO learns to predict the named owner's calls over time, which is the real install.

I have watched CEOs fail this test by overriding once, in the second week, on what felt like an obviously consequential decision. The override was visible to the team. The matrix became decorative within 30 days. The install was lost.

I have watched CEOs pass this test by overriding once, in the second week, then publicly acknowledging the override in the next working session as a failure of their own discipline. The acknowledgment reset the install. The team saw that the CEO took the matrix seriously enough to confess a violation. The matrix held from week three forward.

The discipline is not perfection. The discipline is the visible commitment to the matrix when violations occur. CEOs who can confess overrides publicly produce installs that survive their own occasional lapses. CEOs who cannot produce installs that die quietly with each unacknowledged override.

The Categories That Get Hardest

When I work through matrix construction with leadership teams, the categories that produce the longest discussions are consistent across operations.

Hiring decisions above the senior IC level produce the longest discussion in nearly every install. The org chart usually puts these in the function head's scope. The lived reality usually puts them in the CEO's scope, because senior hires affect culture, leadership team dynamics, and the operating system itself in ways that justify CEO involvement. The matrix needs to resolve this explicitly: either the function head decides with CEO consult (which preserves cultural visibility while honoring the formal delegation), or the CEO decides with function head consult (which preserves the structural reality while clarifying the actual decision-maker). Either resolution is valid; the matrix needs to make the choice rather than leaving the ambiguity in place.

Pricing exceptions produce the second-longest discussion. Sales leaders want the discretion to close deals; finance leaders want the discipline of consistent pricing; the CEO wants to be involved when discounts are large enough to affect margin materially. The matrix needs a threshold structure that distributes authority across the levels — sales lead decides under X%, COO under Y%, CEO above Z%. The threshold values matter less than the discipline of having them explicit; operations without explicit pricing thresholds make exception decisions by polling whoever is in the room, which produces inconsistent pricing and erodes margin over time.

Vendor decisions in the $25K-$250K range produce the third-longest discussion. Below $25K is usually clearly delegated; above $250K is usually clearly CEO. The middle is where the ambiguity lives. The matrix needs to draw a line and hold it, with documented thresholds and named owners. The discussion usually surfaces that the leadership team has been making these decisions through a polling process for some time, with no clear ownership and no clear escalation. The install replaces the polling with the structure.

Customer concession decisions produce the fourth-longest. The escalation pattern in most operations is that concessions get bumped up the chain at each level because nobody wants to own the margin impact. The matrix needs to fix this with explicit authority — the customer-facing leader has discretion up to a threshold, beyond which the COO or CFO has authority. Without the matrix, the concession decision routinely climbs to the CEO, who is structurally the wrong person to be making customer-concession decisions at the volume most operations require.

Each of these four categories is hard for a specific structural reason — the org chart says one thing, the operating reality says another, and the leadership team has been managing the ambiguity informally. The install is the moment when the ambiguity gets resolved formally. The resolution is the value.

The Bottleneck Remediation, Specifically

The most expensive Decision Rights pattern in mid-market operations is the CEO bottleneck. The matrix install is the structural intervention; the operator-narrative is what makes the intervention land culturally.

The bottleneck pattern is asymmetric in a specific way. From inside the bottleneck, the CEO experiences a productive week — every decision they're making feels important and consequential. From outside the bottleneck, the team experiences a slow operation — every decision is waiting for the CEO's attention, which arrives in batches, which produces 7-14 day cycle times on what should be 2-3 day decisions.

The CEO cannot see the cost of being the bottleneck because the CEO is on the inside of the bottleneck. This is the structural property that makes the bottleneck self-perpetuating. The operator who is the bottleneck cannot diagnose themselves as the bottleneck without external diagnostic — the matrix install is one of the cleanest sources of that diagnostic.

The remediation has three steps that have to be done in order:

Identify the five rows where the CEO is most safely delegated. Usually these are decision classes the CEO has been carrying out of historical accident rather than current necessity. The leadership team can usually name these in the first working session if the CEO asks. The CEO has to be willing to hear that they have been carrying things they shouldn't have been.

Re-name the Accountable on those rows. The new owners should be named individuals, not roles. The new owners get walked through the cases where their judgment would likely differ from the CEO's, so both parties have shared expectations about the range of acceptable outcomes.

Hold the CEO accountable for not overriding. This is the part most operators skip. The CEO commits, in writing, to defer to the new owners for 90 days, even when the new owner's call differs from what the CEO would have made. The leadership team is given explicit permission to flag overrides in the working session if they occur. The first month is the test; once the discipline holds for 30 days, the team has learned that the matrix is real and the bottleneck starts to dissolve.

The bottleneck doesn't dissolve in 30 days. It dissolves in 90-180 days, because the operating team has to develop the muscle of deciding independently after years of routing through the CEO. The 90 days is the install; the 180 is the consolidation.

The Cost of Skipping the Install

Operations that skip the matrix install — or run it superficially and revert to the prior pattern within a quarter — produce a recognizable set of costs over the following 12 months.

Decision velocity flat or declining. The half-life of decisions made in the working session stays above 7-10 days. Decisions that should have been one-cycle commitments become two- or three-cycle commitments. The cumulative drag on operating velocity is enormous over a full year — measured in revenue plan misses, missed customer concessions, hiring delays, and the compound cost of every other slow decision.

Leadership team voluntary attrition. The best leaders want to make decisions. Operations that route decisions away from named owners produce leaders who feel structurally underutilized. Voluntary attrition in the leadership team, in my experience, runs 2-4x higher in operations without functioning matrices than in operations with them. The cost of replacing senior leaders is substantial; the cost of replacing them more frequently than the operating environment warrants is structural.

Operating risk concentration. The CEO becomes the single point of failure for the entire operation. The bus test is brutal here — operations without matrices typically have 50-70% of meaningful decisions running through the CEO. If the CEO is unavailable for two weeks, those decisions stall. The risk concentration produces operating fragility that doesn't show on the balance sheet but shows in stress events.

Diligence and capital cost. PE buyers, lenders, and lead investors all run versions of decision-rights diligence on operations they're evaluating. Operations without matrices read as immature, founder-dependent, structurally fragile. The diligence read affects the price. The discount can be material — I have watched specific deals re-priced because the decision-rights diligence surfaced concentration patterns the deal team had not modeled. The matrix is operating infrastructure that also functions as capital infrastructure.

When the Install Is Wrong

There are operations where the Decision Rights Matrix is not the right install. Naming them matters for operator self-diagnosis.

Operations under 15 employees. The matrix overhead exceeds the matrix benefit at this size. The operating team is small enough that decision rights can be coordinated through direct conversation. The matrix install at this stage produces document overhead without proportionate operating benefit; the right move is informal clarity that converts into the matrix at the right scale (usually $5M revenue or 20+ employees).

Operations in active acute crisis. The matrix install requires steady operating conditions to land. Operations in week one of a major customer loss, leadership transition, or capital crisis should defer the install until the acute phase passes. The 30-day discipline of CEO deferral is incompatible with the rapid-decision tempo of acute crisis; the install should happen 60-90 days after the crisis resolves, when the operating cadence has returned to normal.

Operations whose CEO is structurally unwilling to defer. This is the hardest case to diagnose, and it is the case where the install is most likely to fail. CEOs who cannot commit to defer to the matrix for 30 days should not run the install. The install will produce a matrix that becomes decorative within the first month, which costs operating credibility (the leadership team experiences the install as theater) and structural muscle (the team learns that operating discipline does not hold under pressure). Better to defer the install until the CEO is ready than to ship one that doesn't take.

What to Do This Week

If you've recognized your operation in the patterns described above, the install path is concrete.

Run the Twenty-Decision Audit first. Pull the last 20 significant decisions. For each one, capture what the decision was, who actually made it, who the org chart says should have made it, and how long it took. The audit takes a week and produces the evidence base that makes the matrix install land. Without the audit, the matrix install is an abstract conversation; with the audit, the matrix install is a specific response to specific evidence.

Schedule the two working sessions one week apart. Both should include the full leadership team, no observers, no other agenda items. The first session populates the matrix; the second stress-tests it against three real decisions from the prior quarter to confirm the matrix would have produced acceptable outcomes.

Write the CEO deferral commitment in the matrix itself. Not as a separate document. As a line item on the matrix, with the CEO's signature or initials next to it. The commitment becomes harder to violate when it is visible alongside the matrix it governs.

Hold the first 30 days religiously. The team is watching. The first override the CEO makes — or doesn't make — sets the tone for whether the matrix is real or decorative.

The kit guide at /playbooks/decision-rights-matrix covers the structural detail. This essay is what makes the case for the install. The Ops Health Check is the diagnostic that surfaces whether Decision Rights is your top-ranked risk; if it is, this is the next install.

The conversation you are avoiding is the install. The matrix is the residue.

Go have the conversation.