A dotted line is often an agreement to disagree later.
It says 2 relationships matter. It rarely says what happens when they point in different directions.
One leader owns the person's craft, career, and functional standards. Another owns the project, product, region, or customer result. Both have legitimate work. Both can set priorities. The employee is expected to "manage stakeholders," which is corporate language for resolving a design problem through diplomacy.
This is the clarity tax of a matrix organization. The structure buys access to expertise across formal boundaries, but authority becomes harder to read. Every dotted line creates questions that a solid line once answered badly but at least answered visibly.
Who sets the priority? Who evaluates performance? Who can commit the person's time? Who decides when quality, schedule, and customer need conflict? Who is responsible when the compromise satisfies nobody?
A matrix can be the right structure for complex, cross-functional work. It is not a defect simply because it has 2 dimensions. But it is expensive. If leaders do not pay the clarity tax explicitly, the people in the matrix pay it through negotiation, delay, duplicated meetings, and political guesswork.
Why organizations choose a matrix
A functional hierarchy groups people by expertise. Engineers report through engineering. Marketers report through marketing. Finance maintains its standards, sales maintains its customer focus, and each function develops people who share a discipline.
That arrangement becomes awkward when the work itself cuts across those boundaries.
A product needs design, engineering, marketing, finance, legal, and sales. A global company needs both regional judgment and worldwide standards. A consulting engagement needs specialists who still belong to a professional home. A hospital, aerospace program, or major transformation cannot simply choose 1 dimension and ignore the others.
The matrix is an attempt to keep both.
An early Project Management Institute account of matrix structure describes its growth in aerospace during the 1950s and 1960s, where large programs had to coordinate scarce specialists across functions and projects. The account is historical practitioner guidance, not a controlled study. It nevertheless captures the design problem well: complex work often needs more lateral coordination than a functional chain can provide.
The matrix can offer real advantages:
- scarce expertise can serve several products or projects;
- professional standards remain connected across teams;
- customer, geography, function, and product can each have a formal voice;
- people can move between projects without losing a stable craft home;
- decisions can include more of the knowledge they affect.
These benefits are not free. A matrix does not remove hierarchy. It overlays authorities that would otherwise sit in separate structures.
Flexibility enters through the same door as ambiguity.
The evidence is thinner than the folklore
Matrix management has accumulated a familiar catalog of promises and warnings. It breaks silos. It improves information flow. It creates conflict. It slows decisions. It develops people. It exhausts them.
Some of those claims are plausible. The confidence with which they are repeated exceeds the evidence behind them.
Robert Ford and W. Alan Randolph reviewed research on matrix and project structures published since 1976. Their 1992 review described common advantages, disadvantages, and conditions for effectiveness, but it also emphasized the need for much more empirical work to resolve the field's questions and paradoxes.
More than 3 decades later, recent organizational scholarship still notes that many claims about multiple-authority structures remain theoretical or anecdotal. That does not mean matrix structures do nothing. It means there is no responsible basis for saying that a matrix always improves collaboration, always slows work, or comes with a standard performance penalty.
Even PMI's early account, despite listing conflicts and complexity, explicitly cautioned against assuming that multiple authority inherently produces ineffectiveness.
The honest conclusion is conditional. A matrix may fit work that genuinely has 2 important dimensions. It also creates recurring coordination and authority problems that the design must solve. The balance depends on task interdependence, resource scarcity, leadership behavior, performance systems, and whether anyone has written down how decisions work.
That last condition sounds basic. It is remarkable how often the dotted line is treated as the explanation rather than the thing requiring one.
A line is not a decision rule
An org chart can show that a person reports solid-line to a functional manager and dotted-line to a product lead. It cannot encode the operating agreement between them.
At minimum, a matrix needs explicit answers to 5 questions.
1. Who sets priorities?
If both leaders can assign work, someone must decide when the total exceeds capacity. "They should align" describes a hope. Name the person or forum with the final call, the information required, and the time in which a conflict must be resolved.
2. Who owns performance?
One manager may conduct the formal review while another observes most of the work. Decide how evidence will be gathered, whose judgment carries what weight, and how the employee can challenge contradictory feedback.
3. Who controls time and resources?
A project lead who owns an outcome but cannot secure people's time has accountability without authority. A functional manager who can reassign capacity without owning the resulting delay has authority without the consequence. Define commitments in units more concrete than "support."
4. Who decides the trade-off?
Functions often protect standards. Products and projects often protect delivery. Both can be right. Decide who makes the final call when quality, scope, cost, speed, and customer need cannot all win.
5. Who develops the person?
Career growth, coaching, pay, staffing, and craft development may sit with different leaders. The employee should not have to reverse-engineer which conversation matters.
These questions are the work between the boxes. Without answers, the matrix does not share authority so much as leave it in dispute.
The employee becomes the integration layer
When leaders do not resolve matrix tensions, the person at the intersection has to.
They translate priorities between bosses, protect each relationship, explain the same capacity constraint twice, and guess which request will matter during performance review. The work may look like collaboration from above. From inside, it often feels like being the only shared spreadsheet between 2 departments.
The burden is not limited to formal dual reporting. Many employees serve on several teams at once, each with its own leader, goals, meetings, and deadlines. Mark Mortensen and colleagues have described the prevalence and risks of multiple team membership: access to knowledge and resources can improve, while attention becomes fragmented and coordination grows more difficult.
The core problem is aggregate load. Each project sees the portion of the person it needs. Nobody sees the whole allocation.
This is why a manager with 6 reports may be carrying more coordination than a manager with 12. Formal span of control counts reporting relationships. Matrix load includes project memberships, functional obligations, customer commitments, and the switching costs between them.
A healthy matrix makes total demand visible. It maintains 1 shared view of commitments, reserves capacity rather than merely requesting it, and gives the employee a fast escalation path when priorities conflict.
An unhealthy matrix celebrates collaboration while allocating the same 20% of a person 6 times.
Dotted does not mean optional
Organizations often use line style as a substitute for defining the relationship.
A solid line is assumed to mean formal authority. A dotted line might mean influence, service, project accountability, geographic coordination, subject-matter guidance, or "this person should probably be kept informed." The visual convention is doing far more semantic work than it can support.
People then interpret dotted as weaker. Sometimes that is correct. Sometimes the dotted-line leader owns the outcome that determines whether the entire quarter succeeds.
Use a dotted line only after naming what travels through it. For example:
- The functional manager owns hiring, pay, craft standards, and long-term development.
- The product lead owns the weekly priority order and acceptance of delivery.
- Capacity is agreed quarterly and cannot be changed unilaterally inside the period.
- Performance input is shared, while the functional manager owns the final review.
- Conflicts unresolved within 24 hours go to a named leader.
That agreement can fit on a page. It may save months of polite confusion.
The chart should link to or sit beside the operating rules. A line without a legend is decoration.
Conflict is not proof the matrix failed
Some conflict exists because the matrix is doing its job.
The functional leader should care about technical quality, capability, and standards across projects. The project leader should care about delivery, customer needs, and the integrated result. If those interests never differ, the second dimension may not be necessary.
The objective is not to remove conflict. It is to make conflict useful and resolvable.
Useful conflict has:
- shared facts about demand, capacity, risk, and consequence;
- a defined decision owner;
- a deadline for resolution;
- permission to raise the conflict without being labeled difficult;
- a record of the decision so it is not renegotiated by each participant;
- performance measures that do not punish the person for following the agreed priority.
Unresolved conflict has 2 leaders separately giving reasonable instructions whose sum is impossible. The employee works longer, reduces quality quietly, or chooses based on perceived political risk. The organization later calls the result an execution issue.
The system created a standoff and delegated it downward.
When the clarity tax is worth paying
A matrix is most defensible when the work truly needs sustained authority from more than 1 dimension.
It may be worth the cost when expertise is scarce and must be shared, projects are complex and interdependent, local conditions genuinely require authority alongside global standards, or a product outcome cannot be delivered through functional handoffs alone.
It is less defensible when the second line exists because leaders will not choose an owner, because every stakeholder wants veto power, or because a temporary collaboration has been formalized without defining its end.
Before adding a dotted line, ask:
- What valuable decision or resource does this relationship enable?
- Why cannot an ordinary cross-functional agreement provide it?
- What new conflict does the relationship create?
- Who resolves that conflict?
- When will we review whether the line is still necessary?
If the first 2 answers are vague and the last 3 are blank, the matrix is not buying flexibility. It is postponing a decision.
Draw the structure people actually have
The answer to matrix ambiguity is not to hide the matrix.
If people take direction from a product lead, show the relationship. If a regional leader controls meaningful decisions, state them. If an employee is effectively split across 4 teams, make the allocation visible. A simplified chart that conceals real authority is easier to read and harder to use.
At the same time, do not ask the org chart to explain everything. Pair it with a short decision-rights record, shared capacity view, and conflict protocol. Review all 4 when the work changes.
This also helps reveal the informal organization. If everyone bypasses the formal matrix and goes to 1 trusted person for a decision, the network has found either a useful bridge or an unacknowledged authority. Learn which before redrawing another line.
A matrix organization is not doomed. It is simply unforgiving of ambiguity disguised as flexibility.
Pay the clarity tax at design time. Otherwise it will arrive later as meetings, escalation, and work that waits while 2 leaders agree that somebody else should decide.
Sometimes that meeting will be necessary. Quite a few will not. The next article asks why so much unclear ownership ends up on a calendar.