Worker Interview Protocols & Confidentiality SOPs

Social compliance audits rest on three kinds of evidence: documents, site observation, and worker interviews. Documents show what the factory claims. Site observation shows what the factory looks like. Only worker interviews show what workers actually experience: whether overtime is truly voluntary, whether wages arrive in full, whether safety equipment is actually used, and whether grievance channels are trusted. That is why SMETA and every serious buyer audit treat interviews as a core method, not an optional extra.

Interviews also serve as the check on the other two evidence types. When payroll sheets claim full overtime payment and workers report unpaid hours, the contradiction itself becomes a finding, and auditors investigate until the records and the testimony can be reconciled. When the factory floor looks clean but workers report blocked exits at night, the site tour is repeated under different conditions. This cross-checking is called triangulation: no finding rests on one source alone, and no single source can clear a risk on its own.

The auditor selects the sample, never management. The factory provides a complete, accurate worker list covering department, shift, gender, employment type, and start date, and the auditor chooses from it. A credible sample is a cross-section: workers from every major department, from day and night shifts, permanent and contract workers, women and younger workers, recently hired workers and long-serving workers. Each dimension matters because conditions differ across them: night-shift overtime patterns do not appear on day shift, contract workers face different recruitment risks, and new hires reveal recruitment practices that veterans cannot.

Two selection rules are absolute. The list must be complete: omitting night-shift, contract, or recently hired workers is sample manipulation, and auditors detect it by comparing the list against gate logs, production records, and headcount on the floor. And the selection must not be influenced: suggesting friendlier workers, substituting names, or objecting to particular selections all damage audit integrity. A factory that hands over an honest list and accepts the sample without comment has already passed the first interview test.

The interview room must be genuinely private. It should sit away from the production floor, where conversation inside cannot be heard from the corridor, and where no manager can see who enters or leaves. It needs chairs, water, and no cameras. A room next to the production manager office, a room with a glass wall facing the line, or a room where supervisors wait outside all fail this test, and auditors will reject them and delay the audit until a suitable room is found.

Logistics follow the same principle of non-interference. Selected workers are released from their lines without loss of pay and without supervisor escort. They are not told what to say, and they are not debriefed afterwards. Where workers need interpreters, the factory provides them, briefed on accurate translation only, never on steering answers. Workers with hearing, language, or literacy needs are accommodated, because excluding them from the sample would itself bias the audit.

A practical audit-day schedule for a factory of 800 workers across sewing, cutting, finishing, and packing runs like this: the complete worker list is handed over at the opening meeting; the private room is confirmed before the site tour; interviews are scheduled across both audit days so no single department dominates; workers are released in small groups so lines keep running; and after the audit, the factory files no record of who was interviewed or what was said.

Confidentiality has four rules. First, interview content and identities are not disclosed to management or other workers without the worker consent. Auditor reports describe themes and patterns, not named individuals, unless a worker consents or there is an immediate safety risk. Second, participation is voluntary: a worker who does not want to be interviewed is not pressed, and the auditor selects another worker from the sample frame. Third, interview notes are held securely by the auditor, not shared with the factory. Fourth, nothing adverse happens to a worker because they were interviewed: no roster change, no shift change, no transfer, no dismissal. Any adverse change to an interviewed worker after an audit is presumed to be retaliation until the factory proves otherwise.

Factories sometimes ask, innocently, what workers said. The answer must always be no engagement: work only from the auditor findings and the corrective action plan. Even a well-meant follow-up question feels like intimidation to a worker who cannot distinguish curiosity from a threat to their job.

Three behaviours turn a manageable audit into a critical failure. Coaching means instructing workers what to say, or rehearsing answers with them. Interference means observing the interview room, suggesting interviewees, or withholding parts of the worker list. Retaliation means punishing a worker for speaking to auditors or raising a grievance, including dismissal, transfer, shift change, or harassment.

Consider a real pattern, presented here as an illustrative example: a sweater factory in Gazipur prepared carefully for its SMETA audit, with documents reconciled, exits cleared, and an interview room ready. On audit morning, however, a line supervisor told his workers to say overtime was always voluntary and always paid on time, and stood near the interview room door watching who entered. Two workers reported the supervisor comments in their interviews. The auditor recorded an integrity finding that outweighed the otherwise clean labour record. The factory removed the supervisor from all audit contact, issued a written management instruction against coaching and retaliation, and retrained all supervisors before the follow-up audit. The integrity finding closed only after evidence of retraining and the new instruction was verified. The lesson is stark: one supervisor ten minutes of interference cost more than any single labour finding would have.

Prevention is straightforward and cheap. A written management instruction, issued before every audit and included in supervisor induction, stating that no one may discuss interview content with workers, observe the interview room, or change an interviewed worker duties. A pre-audit briefing for workers with exactly three messages: interviews are confidential, you will not face retaliation for speaking truthfully, and you should answer honestly even if the truth is uncomfortable. Nothing more. No fourth message about what to say.

Auditor interviews typically cover wages and how they are paid, working hours and overtime voluntariness, health and safety conditions, freedom of association, grievance channels, treatment by supervisors, and living and dormitory conditions where applicable. Workers are also asked whether they know their rights, whether they hold their own identity documents, and whether they paid any fee to get the job. These topics map directly onto the labour standards, safety, and ethics pillars, which is why interview findings so often confirm or contradict the document review.

Factories should not rehearse answers to these topics. They should instead fix the underlying conditions: pay overtime correctly, keep exits clear, run a grievance channel workers actually trust, and return any held documents. A factory where conditions are genuinely good has nothing to rehearse, and workers who trust confidentiality will say so.

After the audit, the factory works only from the closing meeting findings and the corrective action plan. Each finding gets a root cause, a corrective action, a responsible person, a deadline, and evidence of completion. Interview-derived findings are handled exactly like document findings: no attempt to identify who said what, no investigation of interviewees, only investigation of the underlying condition. A finding that workers fear retaliation, for example, is closed by evidence of a functioning grievance channel and a communicated non-retaliation policy, not by questioning workers about their answers.

Worker interviews work only when workers trust them, and trust comes from three guarantees kept visibly: the auditor chooses who speaks, nobody from management hears what is said, and nothing bad happens afterwards because of it. Factories that protect those guarantees get audit evidence they can actually use to improve, and auditors who trust the process write findings the factory can accept and close. That closes Module 1 of the Social Compliance course: SMETA preparation, labour standards, working hours and payroll, and interview integrity. Module 2 takes the same discipline into WRAP and international social certifications.