Clarifying Responsibilities with the RACI Matrix: Guide and Example

Why do tasks stall even though everyone is involved?

A customer asks for a custom product. Sales needs a feasibility check from engineering, engineering needs exact dimensions from sales, and purchasing does not know whether it should already reserve the material. After a week, the customer calls, and all three honestly say they were waiting for the others.

Such gaps do not arise from carelessness. They arise because nobody has said out loud who stands behind the result and who only contributes. This leads to the guiding question of this article: When does a RACI matrix help a small company clarify responsibilities, and how do you use it without creating bureaucracy?

To answer it, we first explain the four roles and the thinking model of role clarity behind the matrix. Then we look at findings on unclear roles and at two perspectives on the tool. A guide, an example, and the checks by which a good matrix can be recognized follow.

Where everyone is responsible, the gap is already built in.

What do R, A, C, and I mean?

The RACI matrix is a table whose rows contain the activities or decisions of a workflow and whose columns contain the people or roles involved. Each cell shows which role the person has in that activity. The method belongs to the so-called responsibility assignment matrices that have been used in project management for decades (Kerzner and Zeitoun 2026).

The difference between R and A is the heart of the method. Kerzner and Zeitoun (2026) sum it up: you cannot have accountability without responsibility, but you can very well have responsibility without accountability. Many companies have enough people who do something. What is missing is the one person who stands behind the result and says yes or no at the end. In small companies, R and A are often the same person. That is allowed and often even right.

C and I are also often confused. Whoever is a C must be heard before the decision, and that takes time. Whoever is an I learns the result afterward. Every C that could actually be an I slows the workflow down.

Role clarity: what research knows about unclear responsibilities

Behind the matrix is a thinking model from organizational psychology, role theory. It describes a role as the bundle of expectations others direct at a person in their position. Two disturbances are distinguished. Role ambiguity means that the person does not know exactly what is expected of them and what they are responsible for. Role conflict means that expectations contradict each other, for example when sales demands quick commitments and engineering demands complete checks.

Tubre and Collins (2000) summarized the studies on both disturbances and job performance in a meta-analysis. Role ambiguity was associated with lower performance, with an average relationship of minus 0.21. For role conflict, by contrast, the relationship was negligible, at minus 0.07. Gilboa, Shirom, Fried, and Cooper (2008) confirmed the pattern with 169 samples and more than 35,000 employees: among seven stressors studied, role ambiguity and hindering work conditions were most strongly associated with lower performance.

For a small company, this finding is revealing. Contradictory expectations can apparently be tolerated as long as it is clear who owns what. Uncertainty about one’s own responsibility, on the other hand, does measurable harm. A survey of 511 employees in ten care facilities also found that formalized procedures and processes perceived as fair go along with less role ambiguity (Rai 2016). All three studies show associations and no proven cause. Nor do they come from small companies. But the direction is clear: clarity about responsibilities is not a formality but a working condition.

Clarity tool or bureaucracy? Two perspectives on RACI

The first perspective sees the RACI matrix as a simple tool for clarifying what research shows to be important. A study of software projects in India found that responsibility matrices were hardly used or used only half-heartedly, and saw this as one reason for delays and cost overruns (Khan and Quraishi 2014). Kerzner and Zeitoun (2026) describe how in earlier project organizations everyone felt responsibility but nobody took accountability, resulting in mutual blame.

The second perspective is skeptical, and with good reason. A matrix with forty rows and twelve columns in which every second cell has a C clarifies nothing. It is created once, filed in the project folder, and never opened again. And it can hide a leadership problem: whoever is listed as A but has no authority to decide is accountable only on paper. Kerzner and Zeitoun (2026) therefore stress that accountability always requires decision-making authority as well.

I consider both perspectives correct, but for different use cases. For workflows that stay within one department, one name per step is usually enough, as the one-page method provides. The RACI matrix is worthwhile where a task runs across department boundaries and regularly stalls. According to the article on interfaces between departments, these are exactly the places where most information gets lost.

How do you create a RACI matrix in one hour?

1. Choose a workflow (5 minutes)

   Take the task that most often gets stuck between departments. Not “order processing” but “custom request to quote.”

2. List activities and decisions (15 minutes)

   Five to ten rows, each with a verb: check, calculate, approve, order. Decisions get their own row.

3. Enter the people involved as columns (5 minutes)

   Roles instead of names where several people share the same role. Only those who are really involved.

4. Set the A’s first (10 minutes)

   For each row, first the one person who owns the result. Only then R, C, and I.

5. Go through it with everyone involved (20 minutes)

   Everyone confirms their entries. Objections at this point are welcome, because this is where the real ambiguities show.

6. File and use it (5 minutes)

   The matrix sits next to the description of the workflow, not in a separate folder.

The fourth step is the most important. Whoever starts with R fills the matrix with everyone who does something and argues about the A at the end. Whoever starts with A has to answer the difficult question first, and the remaining entries follow more easily.

Example: RACI for a custom request

The following example shows the matrix for the workflow “custom request to quote” in a manufacturing company with 30 employees. The entries are an example and must be adapted to your own company.

What stands out is how rarely the owner appears. He is A only where a price limit is exceeded. All other decisions lie with the department that has the knowledge. What such approval limits can look like for purchases is shown in the article on purchasing and approvals.

Before the cleanup, the matrix in this example looked different. The owner went through the twelve rows of the first version with the checks from the next section and counted the following defects.

How do you recognize a good RACI matrix?

A matrix can be checked in a few minutes. Read each row first, then each column.

• Every row has exactly one A and at least one R.
• Every row has at most two C’s. More are a sign of missing trust or a step that is too large.
• Every person listed as A also has the authority to decide.
• No column is full of A’s. Whoever is A everywhere becomes a bottleneck.
• No column is empty. Whoever has no role anywhere does not belong in the matrix.
• The matrix has no more than ten rows. Longer workflows are split.

The column check often reveals the real problem of a small company. If the owner’s column has an A in every second row, the matrix is honest, but the workflow is not healthy. Then no table helps; what helps is handing over responsibility. How to do that is described in the article on delegating tasks with a clear handover. For projects in which every milestone has exactly one name anyway, the matrix complements the one-page project plan only where several departments work together.

Conclusion: when a RACI matrix helps a small company

The opening question was
a RACI matrix helps a small company and how to use it without bureaucracy. The answer: it helps where tasks run across department boundaries and regularly stall. It works when every row has exactly one A, the A’s are actually allowed to decide, and the number of C’s stays small. For workflows in one person’s hands, it is superfluous.

As shown, in large meta-analyses unclear responsibility is more consistently associated with lower performance than contradictory expectations. That argues for clarifying responsibilities explicitly rather than leaving them to chance. At the same time, a matrix that includes every conceivable participant becomes a source of ambiguity itself. In the end, a simple observation applies: a RACI matrix is only as good as the one question it answers in every row, namely who says yes or no at the end.

What you can put into practice right away

1. Ask in your next meeting: which task most often stalls between two departments here?
2. List five to ten activities and decisions for this task.
3. First set the one A for each row, then R, C, and I.
4. Go through the matrix with everyone involved in 20 minutes and check it against the six rules.
5. File it next to the workflow description and check after four weeks whether the task runs faster.

If a gap does not close even with the matrix, it is usually not the table’s fault but a decision nobody has wanted to make yet. That, too, is a useful result.