In-depth interviews with experts and decision-makers: the method for B2B and policy

1. What the in-depth interview is

The in-depth interview (IDI) is an individual conversation, usually of 45-90 minutes, between a trained researcher and a person chosen for what they know or what they have experienced. It is semi-structured: there is an interview guide with the main themes and questions, but their order and depth adapt to the answers, and the researcher follows the thread that emerges, not the one planned. The aim is to understand a person's reasoning, experience, criteria and language, in a level of detail no questionnaire can capture.

The difference from the focus group is the unit of analysis: the individual, not the group. There is no group dynamic, no conformity to the majority, no dominant participants; there is, instead, time for a single person, confidentiality and the possibility of scheduling the conversation when and where it suits them. For a detailed comparison of the two methods, see the guide on the focus group and the Focus groups and IDI page.

2. When the in-depth interview is the right method

• Experts and decision-makers. Directors, doctors, engineers, public officials, opinion leaders. They are few, busy, hard to gather, and what they know does not fit into predefined answer options.
• B2B research. Industrial customers, distributors, institutional buyers. The purchase decision is complex, involves several people and criteria, and is understood only through conversation.
• Sensitive topics. Health, personal finances, discrimination, violence, socially disapproved behaviour. Individual confidentiality makes possible answers a group would inhibit.
• Public policy and programme evaluations. Key informant interviews with implementers, partners, local authorities and beneficiaries are the standard qualitative component of any evaluation.
• Exploring a new topic. When it is not yet known which questions a survey should ask, 15 interviews with well-chosen people produce the map of the topic.
• Decision processes and journeys. How someone chose a supplier, a bank, a treatment; step-by-step reconstruction requires individual time.

The in-depth interview is not the right method when the interaction between participants is itself the information sought, for example the reaction to a consumer product concept, when budget and time call for efficiency per person, or when a figure is wanted.

3. How many interviews

As in any qualitative method, the number is set by saturation: interviews continue until new ones stop bringing new themes. Experience shows that for a homogeneous category of people, for example procurement directors in medium-sized companies in one sector, saturation usually appears after 12-20 interviews; for heterogeneous categories or for several categories to be compared, 25-40. A programme evaluation study may have 8-10 interviews with implementers, 10-15 with beneficiaries and 5-8 with institutional partners, each group with its own saturation.

The number is planned with a reserve: a minimum per category is set, and interviews are added if the last ones still bring new themes. A proposal offering five interviews for a category offers a preliminary exploration, not a study, and must be presented as such.

4. Recruiting experts and decision-makers

Recruitment is the longest and hardest stage of an in-depth interview study, and the one proposals underestimate most often. A director does not answer a generic e-mail; they answer a personal request that explains who is conducting the study, for whom, for what purpose, how long it takes, how confidentiality is ensured and what they receive in return, usually a summary of the results, sometimes a reward proportional to their time. The acceptance rate at first request is, for decision-makers, 20-40%, and scheduling takes from a few days to a few weeks.

Recruitment sources: lists of companies and organisations from public registers, professional associations, chain referrals from the first interviewees, known as snowball sampling, and, for evaluations, the programme's lists of partners and beneficiaries. In B2B studies, the client can provide contact lists, but interviews with the client's own customers require a decision: the people will know who the study is for, and the agency must decide, together with the client, whether the client's identity is disclosed or not, because the answers differ.

5. The interview guide

The guide contains 8-12 main themes, each with an open entry question and with probing questions the researcher uses if the answer does not cover them spontaneously. The classic structure: the person's context and role; concrete experience with the topic studied, reconstructed chronologically; criteria and reasoning; problems and barriers; comparison with alternatives; the outlook for the future; and, at the end, what was not asked and should have been. Questions are neutral, do not suggest the answer and ask for concrete examples: "describe the last time you changed supplier" produces more than "what matters when you choose a supplier".

For experts, the guide is adapted to each person's specialisation; for decision-makers, it is usually shorter, 45 minutes, and starts with the most important questions, in case the interview is cut short. The guide is tested in the first two or three interviews and adjusted; the test interviews are kept in the analysis if the guide has not changed substantially.

6. Conduct: face to face, telephone, video

Interviews are conducted at the person's office, in a neutral place, by video or by telephone. Face to face is the standard for sensitive topics and long interviews; video works well with busy decision-makers and allows materials to be shared; the telephone is acceptable for short interviews and for people hard to reach otherwise. The interview is audio-recorded, with the person's explicit consent, and transcribed in full. Immediately afterwards, the researcher notes the context and impressions that do not appear in the transcript.

In the Republic of Moldova, the interview is conducted in the language the person prefers, Romanian or Russian, with a native researcher; for international projects, in English or with interpretation, in which case the duration doubles and nuances are partly lost. The multilingual team service covers these situations.

7. Analysis: thematic coding and patterns

Analysis starts from the transcripts and proceeds in three steps. Coding: every relevant fragment receives one or more thematic codes, from a list that starts with the guide and is completed with the themes that emerged spontaneously. Grouping: codes are organised into themes and sub-themes, and for each it is noted in how many interviews it appears, in which categories of people, with what intensity and with what nuances. Interpretation: themes are linked to the research questions, common patterns and divergences are identified, and conclusions are formulated with their degree of support: "all interviewees", "most of those in sector X", "a single interviewee, but with an argument worth checking".

The report contains the synthesis by theme, the differences between categories, anonymised quotes chosen to illustrate the pattern, the typologies of behaviour or decision that emerge, and the implications. It contains no percentages. "Six out of fifteen" may be written as a description of the material, not as an estimate of the population.

8. Confidentiality and ethics

Experts and decision-makers are identifiable: in a small sector, "the procurement director of a large retailer" may be a single person. Confidentiality is ensured by anonymising quotes, aggregating characteristics and obtaining the person's explicit agreement on how they will be described in the report. Recordings are stored securely and destroyed at the agreed date. Interviewees receive, if promised, the summary of the results. These rules are not formalities: they determine whether the next study in the same sector will still find people willing to talk.

9. What to ask the agency

• The number of interviews per category, with the saturation criterion and the planned reserve.
• The selection criteria and recruitment sources, with the estimated duration of recruitment.
• The text of the participation request and what interviewees are offered.
• The interview guide, for approval before the first interview.
• Who conducts the interviews, with what experience in the field and in which language.
• Recording, full transcription, the anonymisation procedure.
• The coding and analysis method, and the structure of the report.
• A realistic timetable: recruiting decision-makers takes weeks, not days.

The in-depth interview is the method through which research reaches the people who know the most and have the least time. It produces no figures, but it produces something figures cannot: an understanding of how decisions are taken in a sector, an institution or a life. For organisations working with decision-makers, institutional partners or B2B markets, it is often the only method that answers the question.

See also: Focus groups and IDI Qualitative research How focus groups work Social and political surveys Multilingual team Methodology FAQ

The Perspicuus research team · 10 May 2026