An accounting firm can look busy and still have capacity. It can also look fully staffed and have none.
The difference is often hidden in work that has started but cannot finish: bookkeeping files waiting for documents, reconciliations waiting for an answer, cleanup jobs sitting in review, and completed work returning for correction. Those jobs occupy attention, create follow-up, and consume reviewer time even when nobody is actively working on them.
This is why a staffing plan alone cannot tell a firm how much new work it can safely accept. Available hours matter, but usable capacity depends on whether work can move through the whole delivery system, including intake, preparation, review, correction, and final approval.
The practical question is not only, “How many hours do we have?” It is, “Where is work waiting, how long has it been there, and what must happen before it can move?”
The headcount answer is real, but incomplete
Staffing pressure is not imaginary. In a 2024 AICPA survey of 667 respondents, finding qualified staff was the top issue for every accounting-firm category except sole practitioners. Larger firms also ranked effective staff utilization and management among their leading concerns. The profession has a genuine talent problem, and no workflow dashboard can make specialized judgment appear where it does not exist.
At the same time, current industry research keeps pointing to a second problem. The Thomson Reuters Institute reported that operational efficiency remained accounting firms' top priority for 2025, while firms also faced the effort of learning new technology and fitting it into existing systems. An AICPA small-firm article published in 2026 made a similar recommendation: map the real engagement, find the logjams, and solve the problem the firm actually has before buying technology.
That distinction matters. A firm may need another experienced reviewer. It may instead have enough review time but poor intake discipline, so reviewers keep receiving incomplete files. It may have adequate staff at every level but too many jobs open at once. It may have a recurring process that was never standardized, which means each new team member adds another version of the workflow.
All four situations can feel like “we are at capacity.” They do not have the same solution.
Capacity is a property of the whole workflow
Imagine a bookkeeping team with several client files due at month-end. The preparation schedule says there are enough hours. Yet some files are missing statements, others are waiting for answers about unusual transactions, and a batch of completed reconciliations is sitting with one reviewer.
If the firm looks only at assigned hours, the team appears to have room. If it looks only at completed tasks, the problem appears near the deadline. The operational truth is visible earlier, in the age and location of the unfinished work.
This idea has a formal foundation in operations research. Little's Law describes a relationship among average work in process, average throughput, and average time in the system. Under stable conditions, more work in process at the same throughput implies a longer average time in the system. An accounting engagement is not a factory unit, and complex work varies, so the equation should not be treated as a precise promise. The useful lesson is simpler: opening more jobs does not create more finishing capacity.
For an accounting firm, capacity therefore has at least three layers:
- Planned capacity: the available time and skill by role, service line, and period.
- Flow capacity: the firm's ability to move work from intake through preparation and review without excessive waiting.
- Acceptance capacity: the firm's ability to check evidence, resolve exceptions, correct errors, and approve work at the required quality level.
A good dashboard connects all three. A staffing schedule without flow data is optimistic. A workflow board without role capacity can ignore an actual skill shortage. A speed dashboard without review and rework data can reward work that is not ready to approve.
Start by defining one unit of work
Capacity measures become misleading when unlike jobs are mixed together. A clean monthly bookkeeping file, a historical cleanup, and a complex tax engagement do not consume the same resources or follow the same path.
Choose one recurring workflow first. For example, define the unit as one client-month of bookkeeping from complete intake to firm approval. Then define the start and finish points plainly:
- Start: all required source records have arrived and the file is accepted into production.
- Finish: the accounting firm's reviewer has approved the work or recorded the final disposition.
The start definition is especially important. If incomplete files enter production, waiting for documents becomes mixed with preparation time. That may be acceptable if the firm wants to measure the entire client experience, but the dashboard should then show the blocked period separately. Otherwise, teams can appear slow when the work was not actually ready to begin.
Once the unit and boundaries are stable, the firm can compare like with like and learn from changes over time. The objective is not to create a universal benchmark. It is to create an honest baseline for that firm's workflow.
The seven signals that reveal usable capacity
The dashboard does not need dozens of charts. It needs a small set of measures that explain movement, waiting, and quality.
1. Work in process and work age
Work in process is the number of jobs that have started but are not yet accepted as complete. Work age is how long each one has been in the system or in its current stage.
The count shows load. The age shows risk. Ten recent files may be healthy; ten files that have sat in review for two weeks tell a different story. Show the median age and the oldest items, then segment them by workflow stage so one queue cannot hide inside the total.
2. Blocked-work rate and blocked age
A job is blocked when the next required action cannot happen. The reason might be a missing statement, an unresolved client question, unclear scope, system access, an internal dependency, or a review decision.
Track the share of open work that is blocked, how long it has been blocked, and the reason category. Do not label every delay as a client problem. Internal review, poor instructions, access failures, and scope changes need their own categories.
This distinction turns a vague complaint into a decision. If missing documents dominate, improve intake and escalation. If internal questions dominate, clarify policies and ownership. If access issues recur, fix provisioning. If scope changes are common, strengthen the intake and change-order process.
3. Reviewer queue volume and age
Preparation capacity is not delivery capacity if finished work cannot clear review. Show how many items are waiting for each review role and how old those items are.
A rising review queue can mean the firm genuinely lacks reviewer capacity. It can also mean work reaches review in uneven batches, evidence is incomplete, preparers escalate too many routine decisions, or reviewers receive files without a consistent acceptance checklist. The queue identifies where to investigate; it does not prove the cause by itself.
4. Cycle time by stage
Cycle time is the elapsed time from the defined start to the defined finish. Break it into stages such as intake, preparation, waiting, review, correction, and approval.
The stage view matters more than one overall average. If total cycle time rises because blocked time increased, adding preparers will not solve it. If preparation time rises while blocked and review time stay stable, the firm may have a production-capacity or standardization problem.
Use medians alongside averages so a few extreme jobs do not distort the normal pattern. Keep the oldest items visible because they often reveal a control or ownership failure that summary statistics smooth over.
5. Throughput and on-time completion
Throughput is the number of comparable work units accepted as complete during a period. On-time completion shows how many met the agreed internal or client deadline.
Throughput should be read beside work in process and quality. A higher completion count is not evidence of more usable capacity if rework rises, reviewers lower the acceptance standard, or unfinished inventory keeps growing.
6. Rework rate and rework reason
Rework is work returned after a review because something must be corrected, completed, or supported. Record the reason rather than only the count.
Useful categories might include missing evidence, unsupported classification, incomplete reconciliation, instruction mismatch, data-entry error, scope issue, or reviewer preference. A high rework rate can consume the exact senior capacity the firm hoped to protect. It can also point to weak instructions, training gaps, inconsistent acceptance criteria, or source-quality problems.
7. Exception mix and decision owner
An exception is an item that cannot proceed under the normal rule and requires additional evidence or judgment. Track exception categories, how often they occur, who owns the next decision, and how long they remain unresolved.
This measure keeps automation and outsourcing discussions honest. A workflow with high repeatable volume and a small, well-defined exception set may be a good candidate for process improvement. A workflow where most items require specialized judgment may need a different staffing and service decision.
What the dashboard should let you decide
The purpose of measurement is not to admire the dashboard. It is to distinguish among different bottlenecks while there is still time to act.
If work is waiting before it starts, the firm may have an intake, document collection, scoping, or scheduling problem. If work moves through preparation but accumulates in review, reviewer capacity or review design is the likely constraint. If files repeatedly return for correction, the firm should inspect instructions, evidence requirements, training, and acceptance criteria. If every stage is stable but planned demand exceeds available hours by role, the headcount or sourcing decision becomes clearer.
That sequence prevents a familiar mistake: buying software, hiring staff, or outsourcing work before naming the bottleneck. Any of those choices can be sensible. None is a substitute for diagnosis.
It also gives the firm a safer way to evaluate improvement. Instead of promising that a tool or provider will “save time,” establish the baseline first. Then compare cycle time, blocked time, reviewer queue age, rework, evidence completeness, and accepted throughput for the same type of work. Keep scope and quality rules constant enough that the comparison means something.
Run the first version for four weeks
Choose one recurring workflow and capture only the fields needed to calculate the seven signals:
- work item and service type;
- date accepted into production;
- current stage and stage-entry date;
- blocked status, reason, and owner;
- review-entry and review-exit dates;
- rework status and reason;
- final acceptance date.
Review the data weekly with the people who prepare and review the work. Ask which items are aging, what is blocking them, whether the reason categories are accurate, and what decision would release the most work safely. Do not turn the exercise into individual surveillance. The unit of analysis is the workflow and its queues, not a leaderboard of people handling different work.
Four weeks will not produce a universal capacity formula, especially for seasonal or low-volume work. It can still reveal whether the firm's immediate constraint is incomplete intake, too much concurrent work, review concentration, recurring rework, or a real shortage of skilled time.
The result is a better conversation. Instead of saying, “We are too busy,” the firm can say, “Our reviewer queue is aging,” “A quarter of open files are blocked for missing records,” or “Rework is consuming the capacity we planned to use for new clients.” Each statement leads to a different next action.
Capacity becomes manageable when waiting becomes visible
An accounting firm should absolutely plan hours, skills, deadlines, and hiring. But those plans become more reliable when they are connected to the work already moving through the firm.
The most useful capacity dashboard is not the one with the most utilization data. It is the one that shows what is moving, what is waiting, why it stopped, who must decide, and whether completed work is accepted without avoidable rework.
Before adding the next client, tool, hire, or outside provider, measure one workflow from complete intake to firm approval. The answer may still be “we need more people.” It may also reveal capacity that is currently trapped in the queue.
Sources
- AICPA and CIMA: Beyond Email, Using AI and Automation to Enhance Capacity
- AICPA and CIMA: Staffing, IRS Service Problems and Leadership Development Are Top Issues for CPA Firms
- Thomson Reuters Institute: 2025 State of Tax Professionals
- Financial Cents: 2025 State of Accounting Workflow and Automation Report
- INFORMS: Little's Law as Viewed on Its 50th Anniversary
This article provides general operational education. It is not accounting, tax, legal, employment, or technology advice. Each firm should set workflow definitions, quality requirements, security controls, and approval authority for its own services and professional obligations.