How to Build an Accountability Chart That Holds Up Under Growth

Start One Level Above the Org Chart

The job in this exercise is not to list tasks. It is to isolate the three to seven things a department exists to deliver such that, if those things are true, the department is winning. A job description is a list of verbs. A core accountability is an outcome, with one owner, and a number.

Work one level above the org chart and one level below strategy. Above the org chart, because accountabilities follow function, not headcount. People change, functions don’t. Below strategy, because the company’s objectives have to decompose cleanly into departmental results, or you have a gap that no one owns.

Three failures show up over and over, and the system is built to prevent each one.

• Job descriptions masquerading as accountabilities. Activity gets tracked instead of outcomes. The fix is the expectation column: a number or a binary state with a date, never an adjective.
• Authority defaulting upward. Everything rolls to the department head, so nobody below them truly owns a decision. The fix is explicit decision rights.
• Accidents at the intersections. Most accidents happen at or near intersections, and the same is true in business. Departments hit every internal metric and the company still crashes where their work crosses. The fix is a handoff register that makes clear who has the right of way.

The Three Layers: From Accountability to Operational Excellence

These are not three separate documents. They are three altitudes of the same thing. Each layer answers a different question, and detail deepens only as you descend. The upper layers stay clean and readable precisely because the detail lives below them.

Layer	The question it answers	How much of it
1. Core Accountabilities	Who is responsible, who has authority, who do I go to when it breaks. General, stable, changes rarely. “Qualified pipeline volume” sits here: it names an owner and a number, and says nothing about how the work happens.	3 to 5 per department
2. Process Map	What is the work, in what sequence, and where does it live. The high-level buckets of activity that together produce the accountability above. Still high level, so the whole flow fits on one page.	Several per accountability
3. SOPs	Exactly how is this one step performed, by whom, click by click. Each bucket of work in the process map links to one SOP. This is where the detail lives, and it lives here and only here.	One per bucket of work

The relationship is strictly hierarchical and one-to-many downward. One accountability owns several processes. One process contains several buckets of work. Each bucket resolves to one SOP.

Authority flows down the same lines. Every accountability has an Outcome Owner who owns the processes beneath it. Every process has an owner, usually a Decision Owner from layer one, who owns the SOPs beneath it. When something breaks, you start at layer one to find who is accountable, drop to layer two to locate which process failed, and open the layer three SOP to see exactly where the step went wrong.

You do not build all three at once. Establish accountabilities first, then decompose only where the work demands it. Most of the value in this system is available at layer one, in a single working session with your leadership team.

The Diagnostic Question Set

Run these against every department, in order. They move a team from describing what it does to declaring what must be true for it to have won.

A. Defining the accountability

1. What does this department exist to produce? State it as an outcome the rest of the company consumes, not an activity the team performs. Marketing does not run campaigns, it delivers qualified pipeline.
2. If this department disappeared tomorrow, what specifically breaks? The answers are your candidate accountabilities. If nothing breaks, the accountability is not real.
3. Why does this matter to the level above? Tie each accountability to a key result or to company strategy. If you cannot, it is probably not core. This is your pruning tool: start with a larger set, then pare back to four or five.

B. Assigning decision rights

1. Who can make this call without checking with anyone? That person is the Decision Owner. If the answer is “ultimately the head,” you have not delegated, you have only assigned a task.
2. Who answers when the expectation misses? That person is the Outcome Owner. On most rows this is the same person as the Decision Owner.
3. Where does this person’s authority stop? Name the scope. The stop-line is what connects to the handoff with the next department.

C. Setting the metric

1. How will we know it succeeded? The expectation must be a number or a binary state with a date, not an adjective. “Improve onboarding” fails. “90% of new hires productive within 30 days, by manager sign-off” passes.
2. Where does the number come from? Name the system, report, or dashboard. No source, no accountability.
3. Can the owner actually move it? Favor leading or outcome metrics the owner controls over lagging vanity figures they merely report.

D. Decomposing into processes

1. What are the handful of high-level buckets of work that, performed together, produce this accountability? Aim for three to seven. More than that and you have dropped into SOP detail too early.
2. Is this a sequence or a set of parallel responsibilities? One feeds the next, or each stands alone. Map it the way it actually runs.
3. For each bucket: who does it, what system does it live in, what triggers it, and what does it hand to the next step? These become your process map columns.
4. Which buckets are high-risk, high-frequency, or frequently done wrong? Those are the ones that earn an SOP first. You do not SOP everything on day one.

E. Mapping the handoff

1. Walking the end-to-end process, where exactly does one department’s accountability end and the next begin? Draw the line on the specific step.
2. Who owns the work in the moment of transfer? There is always a gap of ownership at a handoff unless you name who holds it.
3. What is the definition of done that both sides sign? Misaligned definitions of done are the single largest source of interdepartmental failure.
4. How is the receiver notified, and what happens if they do not act? No trigger and no service level means the handoff runs on memory and goodwill, which fail under load.

Decision Rights: Stopping Authority From Defaulting Upward

The problem this solves is the silent default. Everything rolls up to the department head, so in practice nobody below them owns a decision. They execute and escalate. That kills speed and accountability at the same time. A single “owner” field collapses three different things: who does the work, who decides, and who is answerable for the outcome. The chart separates the last two.

Decision Owner. The single named person who makes the call on this accountability and does not need the department head’s sign-off to act. The head’s name should appear here rarely, only for genuine department-level calls such as budget reallocation, hiring, or strategic trade-offs. If the head’s name fills most of this column, nothing has actually been delegated.

Outcome Owner. The single named person who is answerable if the expectation misses. In nearly every case this is the same person as the Decision Owner, and that is exactly what you want. The person who makes the call should wear the result. Expect these two columns to match on most rows.

When they differ, treat it as a flag, not a feature. A split is tolerable only when the Outcome Owner controls the rules the Decision Owner operates inside. A deal desk analyst holds the decision right to approve a specific discount in the moment, deciding each case against the approved bands. The VP of Sales is the Outcome Owner, answerable if discounting erodes margin across the quarter. That split is healthy because the VP owns the bands the analyst decides within. Apply the test to every split you find: does the Outcome Owner control the standards and targets the Decision Owner works under? If yes, it is a sound delegation. If not, you have separated authority from consequence, and you have a problem to fix before it produces a miss.

The department leader sits above the grid, accountable for the whole, but is not the Decision Owner for specific functions. State it explicitly on every department page:

The department leader is accountable for every result below. They are the Decision Owner only where named. Everywhere else, decision authority is delegated to the named individual, who acts without leader sign-off within the scope of that accountability.

Give them the decision frame, not just the decision right

Naming a Decision Owner answers who decides. It does not answer how they should decide. Hand someone authority without the reasoning that is supposed to govern it and you will get inconsistent, misaligned calls, and you will find yourself asking, “how could you possibly have made THAT decision?” The answer is almost always that nobody told them the rules of the road. Authority without a decision frame is not delegation. It is abdication.

Every Decision Owner inherits a frame made of three things already present in the system. Make them explicit and decisions get consistent, because everyone is deciding from the same inputs.

• Standards set the boundaries. These live in the department header. They define what the Decision Owner may not do regardless of the result: the quality bar, the compliance line, the brand rules, the discount bands.
• Targets define what they are optimizing for. These are the expectations in the grid, the metric and the number. They tell the Decision Owner which direction is better, so the call is made toward a known outcome rather than by instinct.
• The decision framework sets priority when the two conflict. Hitting the pipeline number may tempt someone past the brand standard. State the order of precedence in advance. Without it, each person resolves the conflict their own way and the outcomes scatter.

A Decision Owner should be able to answer three questions before acting. What standards must I stay inside? What target am I optimizing for? When those two conflict, which wins? If any answer is missing, the decision right is not yet safe to delegate.

Worked Examples: Three Departments

Three contrasting departments, filled in to show the standard. The names are illustrative. Note that revenue is never assigned as a single department’s accountability. It is decomposed so each function owns one clean, sourced number.

Marketing

Leader: VP Marketing, Dana Reyes. Key result: deliver qualified pipeline sufficient to support the quarterly revenue target. Standards: brand and messaging standards upheld on every asset, no unapproved claims, data-privacy compliance on all capture.

Core accountability	Decision Owner	Outcome Owner	Expectation (metric + source)
Qualified pipeline volume	Demand Gen Lead, P. Okafor	Demand Gen Lead, P. Okafor	$4.0M MQL-sourced pipeline per quarter. Source: CRM pipeline report.
Cost per qualified lead	Demand Gen Lead, P. Okafor	Demand Gen Lead, P. Okafor	$180 or less blended cost per qualified lead. Source: ad platforms plus CRM.
Content production throughput	Content Manager, J. Liu	Content Manager, J. Liu	8 publish-ready assets per month. Source: editorial board.
Brand integrity	VP Marketing, D. Reyes	VP Marketing, D. Reyes	Zero unapproved-claim incidents. Source: quarterly brand audit.

Sales

Leader: VP Sales, Marcus Bell. Key result: convert qualified pipeline into closed-won revenue at or above the quarterly target. Standards: CRM hygiene maintained daily, approved discounting bands only, no commitments outside signed scope.

Core accountability	Decision Owner	Outcome Owner	Expectation (metric + source)
Closed-won revenue	AE Team Lead, S. Tran	VP Sales, M. Bell	$3.2M closed-won per quarter. Source: CRM closed-won report.
Sales cycle time	AE Team Lead, S. Tran	AE Team Lead, S. Tran	45 days or less median qualified-to-close. Source: CRM stage history.
Discount discipline	Deal Desk, R. Patel	VP Sales, M. Bell	12% or less average discount. Source: deal desk log.
Forecast accuracy	VP Sales, M. Bell	VP Sales, M. Bell	Within 10% of committed. Source: weekly forecast versus actual.

Note the one deliberate split. On discount discipline, the deal desk holds the decision right case by case while the VP of Sales owns the quarterly outcome. That is sound because the VP sets the bands the deal desk decides within. Every other row keeps the two owners identical, which is the norm.

Customer Success

Leader: VP Customer Success, Aisha Nadir. Key result: retain and expand the installed base to protect and grow recurring revenue. Standards: every account has a current success plan, escalations logged within one business day, no silent churn.

Core accountability	Decision Owner	Outcome Owner	Expectation (metric + source)
Gross revenue retention	CSM Lead, T. Gomez	VP CS, A. Nadir	92% or better gross revenue retention. Source: billing plus CRM renewal report.
Net expansion	CSM Lead, T. Gomez	CSM Lead, T. Gomez	110% or better net revenue retention. Source: billing system.
Onboarding time to value	Onboarding Manager, K. Shah	Onboarding Manager, K. Shah	90% of new accounts live within 30 days. Source: onboarding tracker.
Escalation resolution	Support Lead, N. Ortiz	VP CS, A. Nadir	95% resolved within the service level. Source: ticketing system.

The Handoff Register: Who Has the Right of Way

Most accidents happen at or near intersections. The same is true in business. Departments can each hit every internal metric and the company still crashes where their work crosses. So you make it absolutely clear who has the right of way, the same way roads have stoplights, stop signs, and rules everyone knows.

The handoff register is that set of rules. It lives at the company level, and every handoff maps to a Decision Owner on both sides. The From owner holds the right of way up to the definition of done, then yields it to the To owner, whose authority begins there. The boundary is the stoplight. The definition of done is what turns it green. A handoff is simply the point where a process in one department’s map connects to a process in another’s, so this is a cross-departmental view of the same layer two flow, not a separate idea.

• Process or handoff. What is being passed across the boundary.
• From and To. Where the work ends and where it is received, each with the single named owner on that side.
• Definition of done at the boundary. The standard both sides sign. The From side is not complete until this is true, and the To side cannot refuse once it is. This single column prevents most handoff accidents.
• Trigger or mechanism. What physically moves the work and how the receiver knows. “Someone remembers to email” is a flagged risk.
• Failure mode and service level. What happens at the seam when it breaks, with a time bound. Turns the boundary from a hope into a managed control.

Handoff	From	To	Definition of done at the boundary	Trigger	Failure mode and service level
Qualified lead	Marketing, P. Okafor	Sales, S. Tran	Lead meets the jointly signed MQL definition: ICP fit, budget confirmed, intent action logged.	CRM stage change to MQL auto-assigns an AE and fires an alert.	Unworked past 24 hours auto-escalates to the AE team lead.
Signed deal	Sales, S. Tran	Customer Success, K. Shah	Closed-won with signed scope, paid invoice or PO, and completed handoff form.	CRM closed-won triggers the onboarding ticket.	No CSM assigned within 2 business days flags the VP of CS.
Onboarding complete	Customer Success, K. Shah	Customer Success, T. Gomez	Account live, success plan created, first value milestone hit.	Onboarding tracker status “live” notifies the CSM.	Stalled past 30 days flags the VP of CS.
Product defect	Customer Success, N. Ortiz	Engineering lead	Reproducible bug with severity, steps, and customer impact attached.	Ticket tagged for engineering routes to the triage queue.	Severity one unacknowledged past 1 hour pages the on-call.
Win/loss intel	Sales, S. Tran	Marketing, D. Reyes	Closed deal tagged with reason code and competitor, notes complete.	Monthly win/loss export to marketing.	Missing reason codes block the month-end close.

Layer Two: The Process Map

Each core accountability decomposes into a handful of high-level buckets of work. This is the middle altitude: detailed enough to see the flow, general enough to fit on a page. Each bucket names who does it, the system it lives in, what triggers it, what it outputs, how long it takes, and the SOP that explains exactly how.

• Description of work. The bucket of activity, a noun phrase, not a full procedure.
• Owner. The role, named. Usually a Decision Owner from the layer one grid.
• System. Where the work physically lives: the tool, platform, or environment.
• Trigger or input. What starts this step and what it needs to begin. This is what makes the flow a flow rather than a list.
• Output and definition of done. What this step hands to the next, the same discipline as the handoff register applied inside the department.
• Cycle time or frequency. How long it takes or how often it runs. This is where you later find bottlenecks.
• SOP link. The pointer to the layer three document, so the map is navigable.

Some accountabilities decompose into a sequential flow, where one step feeds the next like an assembly line. Others decompose into parallel responsibility buckets that each stand alone. Here is one of each, both from the marketing grid above.

Sequential: qualified pipeline volume

#	Description of work	Owner	System	Trigger or input	Output and definition of done	Frequency	SOP
1	Build target account list	Demand Gen Analyst	Enrichment tool plus CRM	Quarterly ICP refresh	Validated account list loaded to CRM	Weekly	MKT-PIPE-01
2	Validate and enrich contacts	Demand Gen Analyst	Verification tool plus CRM	New list from step 1	Clean, deliverable contact records	Weekly	MKT-PIPE-02
3	Launch outreach sequences	Demand Gen Lead	Outreach platform	Enriched list ready	Live multi-touch sequences sending	Per campaign	MKT-PIPE-04
4	Score and qualify leads	Demand Gen Lead	CRM scoring	Reply or engagement signal	Lead marked MQL or recycled	Daily	MKT-PIPE-03
5	Hand off MQL to sales	Demand Gen Lead	CRM	Lead reaches the MQL threshold	MQL accepted by an AE, per the handoff register	Real time	MKT-PIPE-05

Parallel: brand integrity

#	Description of work	Owner	System	Trigger or input	Output and definition of done	Frequency	SOP
A	Maintain messaging guidelines	Brand Manager	Brand wiki	Any positioning change	Current guideline doc published	Ongoing	MKT-BRAND-01
B	Review assets pre-publish	Brand Manager	Asset library	Asset submitted for review	Approved, or returned with notes	Per asset	MKT-BRAND-02
C	Run the quarterly brand audit	VP Marketing	Audit checklist	Quarter close	Audit report plus incident log	Quarterly	MKT-BRAND-03
D	Manage trademark and claims compliance	VP Marketing with legal	Legal tracker	New claim or campaign	Cleared claims register updated	As needed	MKT-BRAND-04

Layer Three: The SOP Standard

SOPs do not live inside the accountability document, because they would bury it. Each bucket of work in a process map links to its own SOP file. What the document does is set the standard every SOP must meet.

• SOP ID. A unique identifier that ties the SOP to its process and accountability, for example MKT-PIPE-03.
• Owner. The single role accountable for keeping this SOP current, not necessarily the person who performs it.
• Trigger. The event or condition that starts the procedure.
• Steps. Numbered, click by click, specific enough that a new hire can follow them unaided.
• Systems and tools. Every platform the procedure touches.
• Definition of done. The exact output that signals completion, the same one shown in the process map.
• Last reviewed. A date. An unreviewed SOP is a liability, and stale steps are worse than none.

The SOP link in each process map row points to that SOP’s file, and the SOP’s ID, owner, trigger, and definition of done mirror the process map row exactly, so the two layers always agree. Change one, change the other.

Here is the SOP behind “score and qualify leads” in the sequential example above. Notice how its header fields mirror that row.

SOP ID	MKT-PIPE-03 (process: qualified pipeline volume, step: score and qualify leads)
Owner	Demand Gen Lead, P. Okafor, accountable for keeping this SOP current
Trigger	A reply or tracked engagement signal arrives from an active outreach sequence.
Steps	1. Open the daily engagement queue in the CRM scoring module. 2. For each lead, confirm ICP fit against the saved scoring rubric, firmographics and role. 3. Check the engagement signal: reply sentiment, content opened, or meeting requested. 4. Confirm the budget or intent indicator is present; if it is missing, recycle to nurture with the reason code “no intent.” 5. If the score clears the MQL threshold, set the stage to MQL, which fires the sales handoff. 6. Log the decision; the CRM timestamps the action for the cycle-time metric.
Systems and tools	CRM lead scoring module, outreach platform engagement data, team chat alerts.
Definition of done	Lead is either marked MQL and routed to an AE, or recycled to nurture with a reason code. Mirrors the process map output for this bucket.
Last reviewed	A date, refreshed every quarter.

Two Checks, and the Rollout Order

Once every department is mapped, run two passes across the whole set.

• Coverage. Does every strategic objective have at least one departmental accountability laddering to it? Gaps are objectives nobody owns.
• Overlap. Does any single outcome appear in two departments’ grids with two different owners? Resolve it now, on paper, rather than in a quarterly miss later.

Then roll it out in this order.

1. Map the company’s strategic objectives first, so every key result has a parent.
2. Establish layer one across the whole company. Run the question set with each leader, fill the grid, and force one name per Decision Owner and Outcome Owner cell. Start broad, then pare back to four or five.
3. Run the coverage and overlap passes, and resolve every gap and collision before going deeper.
4. Decompose to layer two process by process, building each process map with the seven columns.
5. Walk each end-to-end process across departments and populate the handoff register.
6. Write layer three SOPs only where the process map reveals fragility: high-risk, high-frequency, or frequently-done-wrong buckets first. Do not boil the ocean.
7. Review quarterly. Accountabilities drift as the business changes, processes break when volume grows, and SOPs go stale.

What you get at the end is not a chart on a wall. It is a company where every outcome has one name against it, every decision has an owner who does not need permission, and every seam between two teams has a rule instead of a hope.

Where This Connects

The named term and the short definition live at the Core Accountability System. The language your team uses inside the grid, commitment by commitment, is the Accountability Framework. Once every number has an owner and a source, your leadership meeting can manage by exception instead of walking through everything that is already fine, and the meeting cadence becomes the place the grid gets reviewed. If most of the decision rights currently point at you, that is the founder bottleneck, and this is the way out of it.