QM Management Depth (14) | Quality Organization Structure Design: Choosing Between Centralized, Decentralized, and Hybrid Models

A certain electronics manufacturing company, with an annual revenue of about 1.2 billion yuan, operates three production bases. The group quality department has 28 people responsible for the quality management system (QMS), laboratory, and customer interface; each of the three bases has its own quality department, totaling more than 60 people, responsible for incoming inspection, process inspection, and outgoing inspection.

The conflict erupted during a major customer complaint: the batch returned by the customer was judged to have terminal plating peeling. Base A claimed that the semi-finished products supplied by Base B were defective, while Base B blamed the wave soldering parameter drift at Base A. Both sides had their own inspection data and criteria, and neither would acknowledge the other. The group quality department wanted to unify the judgment, but the two base quality managers had inconsistent reporting and evaluation criteria—one was promoted by the base general manager, and the other was dispatched by the group. After three weeks, the customer recall letter arrived, and the internal debate over "whether it was incoming material or process" continued.

1. The Essence of the Problem: This is About Responsibility and Authority Allocation, Not Just Drawing an Organizational Chart

Most quality managers understand the organizational structure issue as "who reports to whom," and thus focus their efforts on revising the organizational chart. However, what truly determines the operational effectiveness is the combination of three variables:

1. Reporting lines and evaluation authority—Who decides the quality manager's appointment, evaluation, and bonus? 2. Functional boundaries—What matters are decided by the group, and what matters are decided by the bases? 3. Resource allocation—Who controls the budget and staffing?

The debate between centralization and decentralization is essentially about where these three aspects are placed. Changing only the first aspect (reporting lines) without addressing the other two (boundaries and resources) will not alter the organizational chart.

There are three common misjudgments to be wary of in this issue.

Misjudgment One: Treating the organizational structure issue as a personnel issue. Inconsistent judgments and non-uniform standards are often due to the lack of a clear "final authority" in the organizational structure, not because of an individual's lack of capability. Replacing personnel can only lead to a new round of localized bargaining.

Misjudgment Two: Assuming "centralization equals uniform standards." Centralization can indeed unify standards, but at the cost of response speed and on-site responsibility. Once quality personnel at the bases no longer report to the base managers, they are likely to lose their "on-site judgment under delivery pressure" and become an approval step at the group headquarters.

Misjudgment Three: Making significant changes to the organizational structure during turbulent periods. Major customer complaints, new product introductions, and audits are the least suitable times to make changes to the organizational structure—the risks of these two events will compound. The prudent approach is to address the current issues first and make organizational adjustments during a relatively stable window.

2. Comparison of Three Organizational Structures

Centralized Model is suitable under the following conditions: multiple bases but each with a small scale, high homogeneity in products and processes, customer requirements for a unified external interface, or a weak group system capability that needs to be rapidly leveled. The advantages are uniform standards, resource reuse (laboratory, audit resources, expert manpower), and strong independence—inspection and judgment do not have to yield to production. The disadvantages are slow response, disconnection between base quality and production, and the base general manager not being accountable for quality outcomes.

Decentralized Model is suitable under the following conditions: significant differences in products and processes, a sufficiently large scale for a single base, delivery sensitivity, and independent financial accounting at the base level. The advantages are proximity to the site, quick response, and clear responsibility. The disadvantages are standard drift, weak independence (quality managers report to base general managers, making judgments more likely to yield under delivery pressure), and duplicate investment.

Hybrid Model is usually the practical answer, and the key is to correctly allocate the three levels:

• Must be Centralized: Functions that "must have a single answer"—QMS and documents, metrology and laboratory, customer interface and complaint judgment criteria, supplier access and key supplier audits, quality data platform, and quality personnel qualification standards.
• Must be Decentralized: Functions that "must be close to the site"—process quality control, on-site inspection, initial judgment and isolation of nonconforming products, and on-site improvement promotion.
• Must be Matrix: Functions that "require a dual perspective"—final judgment of major nonconformities, on-site investigation of customer complaints, and quality reviews during the new product introduction phase.

3. Implementation Actions: Five Practical Steps

Step One: Write the "Three Lists" Before Discussing the Organizational Chart. List out the centralized, decentralized, and matrix functions, specifying "who sets the standards, who executes, and who evaluates" for each item. The criterion is: each row must have a single primary responsible person, and "joint responsibility" is not allowed. Do not issue any organizational adjustment documents until this step is completed.

Step Two: Define Reporting Lines as Executable Authorities. The dual reporting lines for quality managers (solid line to the base general manager, dotted line to the group quality director) must clearly define what the dotted line manages: qualification review authority, annual evaluation weight (recommended to be no less than 30%), and the authority to stop production and veto major nonconformities. The criterion is that these contents are explicitly written into the "Quality Organization and Authorization Regulations" and signed off by the CEO.

Step Three: Ensure Result Responsibility Does Not Fall Solely on the Quality Department. Include at least one result-oriented quality metric (customer complaint frequency, external failure cost rate, major quality incidents) in the base/department general manager's evaluation, and provide an acceptable weight. The criterion is to review the general manager-level evaluation forms and identify the row that can be shown to a third party.

Step Four: Institutionalize Interfaces. The interface between quality and production should include daily quality meetings, stop-line rules, and nonconforming product handling timelines; the interface between quality and R&D is the quality sign-off points in the stage gate reviews of new products; the interface between quality and procurement is the supplier access and change review. The criterion is that each interface has a fixed meeting frequency, decision-making authority table, and escalation path, rather than relying on personal communication.

Step Five: Proceed in Steps to Protect the Transition Period. The sequence is: first unify standards and data criteria, then unify the reporting and judgment authority of key functions, and finally adjust staffing and evaluations. Allow one to two quarters for each step, and during the transition period, implement a "dual-signature" rule—judgments are co-signed by the base and the group to avoid creating a vacuum on-site due to a one-size-fits-all approach.

4. Case Development: From Forced Centralization to Layered Hybrid

Returning to this electronics company. After being "pushed back," the quality director did not continue to forcefully push for centralization but instead revised the plan to a layered hybrid model with a three-step approach.

Step One: Centralize "judgment authority" and "customer interface." Establish a group quality judgment team consisting of three people: one group quality engineer and two process quality supervisors from different bases. Major nonconformities are judged within 24 hours and responsibility is assigned once; all customer complaints are handled through a single external channel, and bases no longer respond to customers individually.

Step Two: Centralize standards, decentralize execution. The QMS documents, inspection standards, metrology laboratory, and supplier access list are unified by the group; bases retain the execution and initial judgment authority for process quality and on-site inspection. The laboratories of the three bases are merged into a single metrology center.

Step Three: Implement reporting lines and evaluations. Base quality department managers report solidly to the base general manager and dottedly to the group quality director; qualification reviews are conducted by the group, and the group's annual evaluation weight is 40%, with the authority to stop production for major nonconformities residing with the group. Two metrics are added to the base general manager's evaluation: customer complaint frequency and external failure cost rate, with a combined weight of 15%.

The results became evident by the seventh month: the average judgment time for major nonconformities decreased from 18 days to 3 days; customer complaints in the first year decreased by about 40%, and the recurrence of similar issues dropped from 7 to 2; the metrology center reduced its workforce by 4 people, and the calibration cycle shortened from 15 days to 6 days.

The costs were also significant. First, the reform actually took nearly three quarters, far exceeding the original plan of one quarter—during the rush to merge the laboratories, the calibration schedule at one base was interrupted, resulting in two batches of products with dimensional overruns being sent to the client, leading to a customer letter. This was entirely avoidable. Second, the resistance from the base general managers did not disappear; it merely shifted from "opposing centralization" to "focusing on evaluation weights." The quality director had to renegotiate the weights at the beginning of each year. His later summary was: organizational structure is not a one-time design but an agreement that needs to be renewed annually.

5. Self-Inspection Checklist

• Have the centralized, decentralized, and matrix lists been written into documents, and does each row have a single primary responsible person?
• Are the solid and dotted reporting lines for quality managers clearly defined, and are the dotted line authorities (qualification, evaluation weight, stop-line authority) specific and executable?
• Does the base/department general manager's evaluation include at least one result-oriented quality metric?
• Are the interfaces between quality, production, R&D, and procurement clearly defined with fixed meeting frequencies, decision-making authority tables, and escalation paths?
• Is there a phased plan and transition period rules for organizational adjustments, rather than a one-time issuance of documents?

Organizational structure changes the allocation of responsibilities and authorities, not just the boxes on an organizational chart.

Knowledge code: 13.1.1

Version: v20260924

Author: QTank QTank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.