Twenty-decision audit.

What Twenty Decisions Tell You That No Org Chart Will

There is a specific operating diagnostic that produces more pointed insight per hour invested than any other tool I have run with mid-market leadership teams. It costs two weeks of structured collection. It produces a single page of evidence. It cannot be argued with, because the evidence is the team's own decisions.

The Twenty-Decision Audit pulls the last twenty significant decisions an operation has made and traces, for each one, four pieces of evidence: what the decision was, who actually made it, who the org chart says should have made it, and how long the decision took from initiation to commitment.

The output is uncomfortable on purpose. Most leadership teams discover that their org chart and their decision flow do not match. The mismatch is not a surprise to the team — they have been living inside it. The surprise is the structural specificity of where the gap lives. Bottleneck mode? Fog mode? Bypass pattern? The audit answers the question that the team has been managing around for months.

This essay is about what the audit surfaces, what to do with the surfacing, and why most operations need this diagnostic before the Decision Rights Matrix install will land. The kit guide covers the structural mechanics; this is the operator narrative.

The Three Patterns the Audit Always Finds

When the audit runs across the operations I have worked with, the same three patterns surface with consistent reliability.

Bottleneck. Every decision in a specific category routes through one person, regardless of what the org chart says. Usually the CEO; sometimes a strong COO or function head. The org chart says authority is distributed; the audit shows authority is concentrated. The bottleneck person doesn't realize how bottlenecked things are, because from inside the bottleneck, the work feels productive — you're making important decisions all day. The audit makes the pattern visible from outside the bottleneck.

Fog. Decisions in a specific category are made by whoever shows up to the meeting first. There's no clear owner. There's no clear escalation path. The actual decider varies decision to decision in a way that doesn't track any consistent logic. The CEO believes the operation is delegated; the team experiences it as politically driven with no map. Fog patterns are usually the result of formal delegation without structural follow-through — the role exists, but the authority hasn't been operationally clarified.

Bypass. A specific role consistently makes decisions outside its named authority — often a strong functional leader who has built operational power that exceeds the role's formal scope. The bypass leader is usually producing good decisions; the cost is in the rest of the team that has learned the org chart is decorative and stops trusting any of its named authorities. The bypass pattern is structurally different from the bottleneck pattern because the work is happening at the right level but the wrong person is named.

Each of these patterns has a specific remediation in the Decision Rights Matrix kit. The audit's job is to surface which pattern is dominant in which decision categories. The matrix's job is to fix it.

Why the Audit Is the Prerequisite, Not the Endpoint

Operators sometimes assume the audit is itself the install — that surfacing the pattern is what fixes it. The audit is structurally a diagnostic, not a remediation. The Decision Rights Matrix kit is what closes the gap.

The audit's value is in making the matrix install land. The matrix is a one-page document; the install is the conversation underneath the document. The conversation is much easier when the team has shared evidence from the audit than when it has only abstract assertions about authority.

I have watched matrix installs without the audit produce decorative matrices that get bypassed within a quarter. I have watched matrix installs with the audit produce real structural change because the leadership team cannot argue with twenty real decisions. The audit changes the conversation from "what should the matrix say" to "what does the matrix need to say so that this decision routes the way it should have."

The cost of the audit is two weeks of structured collection. The benefit is matrix installs that take instead of installs that produce posters.

The Four Pieces of Evidence Per Decision

For each of the twenty decisions in the audit, the structured collection captures four pieces of evidence:

What the decision was. A one-sentence description. Specific enough to be unambiguous — "approved the $80K marketing spend for Q3 campaign A" rather than "approved marketing spend." The specificity prevents later disputes about whether the audit captured the decision correctly.

Who actually made it. The person who said the final yes or no, not the person who was copied on the email or attended the meeting. Decisions made by consensus require a follow-up: who would have overridden if consensus had failed? That person is the actual decider. This column is the hardest one to populate honestly, because operators often want to credit a more distributed decision pattern than actually exists.

Who the org chart says should have made it. The person whose role description names this decision class. If multiple roles could plausibly own the decision, name them all. The audit is allowed to surface organizational ambiguity — that is part of what it's for.

How long the decision took. From the moment the decision was first surfaced to the moment commitment was made. Hours, days, or weeks. The latency data is what surfaces bottleneck patterns; chronic high latency in specific categories indicates that the decisions are routing through someone who's structurally over-allocated.

The four-column matrix is the artifact. Twenty rows. Sixty data points of evidence. The patterns emerge from the patterns.

What "Significant" Means

The audit requires the leadership team to select "significant" decisions, which inevitably produces a definitional discussion. The right definition is operational rather than dollar-amount-based.

A decision is significant if (a) it materially affects the operating performance of the business for the current quarter or beyond, or (b) it sets a precedent for how future decisions of the same class will be made, or (c) it required leadership-team input to land. Routine decisions made within a function (a marketing manager approving a $2K vendor invoice) are not significant for this audit; consequential decisions involving cross-functional coordination, customer-facing implications, or material resource allocation are.

A useful heuristic: if you had to brief a new VP on the operation, would this decision be one of the examples you'd cite to illustrate how things get done? If yes, it's significant. The discipline is to be inclusive about what counts as significant during week one of collection; the patterns emerge from the breadth rather than from any specific subset.

A typical mid-market audit produces a mix of capital decisions, hiring decisions at various levels, pricing exceptions, vendor selections, customer concessions, organizational changes, and strategic pivots. If a category is unrepresented in the last twenty significant decisions, the absence is itself diagnostic — the operation may be under-deciding in that category, or the team may be making decisions in that category without recognizing them as significant.

The Two-Week Install Cadence

The kit guide lays out a two-week structured collection. Week one is collection. Week two is synthesis and remediation planning. The cadence matters for two reasons.

Week one's collection has to be inclusive rather than selective. The team identifies the twenty decisions, populates the four columns, and resists the temptation to pre-interpret the data. Pre-interpretation produces a sanitized audit that misses the patterns; full collection produces an audit the patterns can emerge from.

Week two's synthesis is when the patterns get named. The leadership team meets to walk through the populated audit. Bottleneck, fog, bypass — each pattern gets named where it shows up. The audit produces a heat map showing where each pattern dominates which decision category. The heat map is the artifact that drives the matrix install.

The two weeks are not consecutive eight-hour days. They are structured collection happening in the background of the normal operating week, with two-to-three hours of leadership-team time invested in synthesis. The total operator time investment is roughly a day across two weeks. The diagnostic produced is worth the investment by a wide margin.

The Conversation That Has to Happen With the CEO

The audit will surface, in almost every case, that the CEO is the bottleneck in some categories. This produces a specific conversation that the kit install requires.

The conversation goes like this. The audit shows that X decisions in the last twenty routed through the CEO despite being formally delegated. The CEO acknowledges the pattern. The leadership team acknowledges the pattern. The question is: what changes.

The honest answer is usually one of two things. The CEO recognizes that the bottleneck pattern is hurting operating velocity and commits to deferring to the matrix-named owners for the next ninety days. Or the CEO decides that some of the formally-delegated decisions actually belong to the CEO operationally, and the matrix needs to be updated to reflect that reality.

Both answers are valid. The discipline is the explicit decision. Operations where the CEO can neither defer nor formally re-claim the decisions produce matrices that don't take — the leadership team learns that the formal authority is decorative, and the matrix erodes within a quarter.

The audit forces this conversation by producing evidence the CEO cannot argue with. Without the audit, the conversation gets had in the abstract and produces ambiguous resolutions. With the audit, the conversation gets had in the specific and produces resolutions the matrix can codify.

What to Do This Week

If you have run the Ops Health Check and Decision Rights is in your top three risks, the audit is the next install.

Schedule the two-week structured collection. Assign the audit owner from the leadership team — typically the COO, chief of staff, or operations lead. The audit owner is responsible for the four-column data integrity, not for interpreting the patterns.

Pull the last twenty significant decisions. Use the operational definition from above. Resist the temptation to curate.

Populate the four columns per decision. "Who actually made it" is the hardest column; populate it honestly even when the answer is uncomfortable.

Run the synthesis session at the end of week two. Leadership team in the room. Walk the heat map. Name the patterns. Identify the top three intervention candidates for the matrix install.

The matrix install follows the audit by 30-90 days, depending on operating cadence and CEO bandwidth. The audit's value compounds even if the matrix install is deferred — the team operates with sharper awareness of where decisions are routing once the heat map is visible.

The kit guide at /playbooks/twenty-decision-audit covers the structural detail. This essay is the operator narrative for why the diagnostic is worth the two weeks. If Decision Rights ranks high on your Ops Check, the audit is the install that makes the matrix install land.

Twenty decisions. Four columns. Two weeks. The evidence base for the install the operation has been deferring.