The Art of the Discovery Interview

“What?” raced through Janet’s head as she read the email. “Now that’s a surprise.”

The message was from Jack Johnson, vice president of development. It said she would receive a meeting request from Rajan Alak, an outside consultant, to interview her about the problems with the new system. The message went on to say the company had made a significant capital investment in the development of Synergy and problems with the system were preventing the company from enjoying the expected ROI. Jack asked Janet to give Rajan her full cooperation.

“He wants me to give an outside consultant—a total stranger—my full cooperation?”

The problems with Synergy didn’t surprise Janet. She had invested almost all of her time during the past year in developing the system. She believed the business planners had been too aggressive with the system integration plans. She thought the company’s chance of achieving the projected ROI was zero. Her suspicion was that the projection was based on politics rather than reality.

A portion of Jack’s message was a surprise: In addition to fixing the problem with Synergy, he wanted to fix problems in the development process that had caused the issues with the system. Solving problems with the development process was an initiative Janet had wanted to see since she started her job seven years ago.

She waited to exhale and asked herself, “How much should I tell this outside consultant? Will my statements be used against me? Or my manager?” And as she exhaled, no answers came.

The next morning Janet received the meeting request from Rajan to interview her the following Wednesday. The request included a copy of Jack’s message and told her she would receive more information about the interview in a forthcoming email.

A part of her kept wondering, “How much can I say?”

Rajan’s email titled “The Interview Process” arrived two days later. Janet read the message carefully. It said she had complete control over the information she shared. She could choose to have information marked as originating from her, originating from an anonymous source, or recorded as off the record. Rajan said that after the interview she would receive a transcript of the “on the record” parts of the interview for her review and approval. Rajan emphasized he would not share any of her comments with anyone else until she approved them.

“Hmm . . . ” she thought. “Maybe it’s OK to share what I know.”

Fundamentals

The quality of an interview depends on how safe the interviewee feels. People guard their knowledge when their answers may endanger themselves or a valued colleague. The safer the interviewee feels about answering questions, the higher the quality of information available to the interviewer.

Creating a safe environment is only the start. In addition to safety, the quality of a set of interviews–whose purpose is to discover problems and solutions–depends on managing the sponsor, interviewing the right people, and interacting skillfully with the interviewees.

Let’s explore effective actions available to the interviewer before, during, and after the face-to-face interview.

Before the Interview

Sponsor Agreement

Gaining clarity about the objectives of the sponsor saves everyone time. Help your sponsor write down what is important to him and, just as importantly, what will gain the cooperation of the interviewees. Have the sponsor sign off on a written set of discovery objectives and a list of people to be interviewed.

For instance, if a vice president says her objective is to fix the problems with a system, her message will be compelling to some of her people. Adding that she also desires to solve the development problems that caused the system’s problems may energize additional people.

Rarely does anyone create objectives that are compelling to everyone. Objectives that are compelling to some people may de-energize others. So, focus on creating compelling objectives for the people whose opinions matter most to your sponsor.

Overcome Restrictions

You will want to talk to the organization’s customers. Some organizations carefully restrict who communicates with their customers. Despite these barriers, assume management wants you to speak with them. Work with your sponsor to identify which customers you will interview. If your sponsor objects to your talking directly to the customer, negotiate. Explain that without customer feedback, the most that can be discovered is less than half the available information.

I prefer to interview the customers first. I want to hear their unfiltered perspectives about outcomes that were expected but never satisfied. Next, I like to interview key middle managers to gain additional perspective. The customer and middle management interviews reveal a panorama of the most visible problems and provide an opportunity to find out more about whose opinion is the most influential.

Sponsor Communication

Regardless of whether you are an inside or outside interviewer, the person you are interviewing needs to hear from his management why he is being interviewed, who will perform the interview, and what actions management expects from him. In the introductory story, a vice president, Jack Johnson, provided that information to Janet.

Prepare an email for the sponsor to send to all the people being interviewed. Take control; if the context isn’t set properly, it will be a barrier to your success. If the sponsor is uncomfortable with your message, ask him to discuss it with you and work with him to revise it so it works for the sponsor and you.

Ask the sponsor to send the message to each interviewee individually. My experience is that messages addressed to a single recipient gain more attention than messages addressed to a group.

Also ask the sponsor to mention the interviews in staff meetings and to emphasize the importance to the interviewees. Scheduling interviews is difficult in busy organizations. When upper management deems the interviews to be of high priority, middle management will more readily support the scheduling of its people’s time. Otherwise, the interviews will be a low-priority event that may never happen.

Interviewer Communication

After the message is sent from the sponsor, it’s up to you to schedule the interview. I suggest you attach the sponsor’s original message to your scheduling request so recipients can review it. The inclusion of the original message prevents confusion by people who may not have read the message from their management.

Follow the meeting request with an email from you to each interviewee explaining how the process will work. This message lets the interviewee know he controls the use of the information he shares. This email will surprise the recipient. People in large organizations frequently receive messages about protecting the company’s rights but rarely receive messages giving them rights.

My preference for the first interview requires no preparation by the interviewee. If you are interviewing the right people, they already know everything they need to know. Inform the interviewee that he doesn’t need to do anything prior to the interview.

I strongly suggest you telephone the interviewee the day before the interview to confirm the time and location. Priorities change regularly in organizations and the interviewee may need to cancel the interview. Knowing about cancellations early will enable you to reschedule your day. If your call is transferred to voice mail, let the interviewee know the time and location of the interview, leave your cell phone number, and let him know that you’ll assume everything is as scheduled, unless you hear from him.

Question Sequencing

The sequence of your questions contributes significantly to a successful interview. A key aspect of most interviews is gathering information about problems. I like to look at questions as either branches or stems. Branch questions move to a new subject area. Stem questions (indented below) gather more detail about a branch. Let’s look at a high-level plan for sequencing questions during a sixty-minute interview:

Q: Who is your customer?

How does your customer relate to Synergy?
Who else is your customer?
Would you recommend that I interview any of the people you mentioned?

Q: What problems did Synergy solve?

Tell me more.
Anything else?
Someone in a previous interview mentioned that Synergy retired a number of older applications. What’s your take on that?

Q: What problems did Synergy create?

Tell me more–what evidence do you have?
Who else should I talk to about that problem?
Who might see this differently?
Anything else?
Francois suggested that I ask you about complaints about poor performance. What can you tell me about that?
Why did this problem occur?
Could something have been done to prevent it?
What suggestions do you have for fixing the problems?

Q: What problems happened during development?

Tell me more.
How did that affect you?
What else?
What recommendations do you have for fixing the problems?

Q: What other questions should I be asking you?

How would you answer your questions?
Anything else?

Q: Do you have any questions for me?

Q: May I contact you if I have additional questions?

These questions can be asked to anyone in the organization. As you gain information from each interview, adapt your questions to fit the person you are interviewing.

Metaquestions

In addition to questions on the topic of interest, effective interviewers equip themselves with metaquestions to gather feedback about the interview process itself. Metaquestions are questions about questions. For instance, if you see a puzzled look on the interviewee’s face, you might respond, “I see a look on your face that suggests to me that you might be puzzled by my question.”

I find answers to metaquestions open new possibilities about what to do next. For instance, you may discover that the person you are interviewing has a different role than you thought and the role isn’t relevant to the discovery. Rather than continue the interview and waste his time and yours,you now have the option of ending the interview. The following is a list of metaquestions I have found valuable in any interview situation:

Do you have any questions for me?
Do my questions seem relevant?
Do my questions puzzle you?
Are you the right person to answer these questions?
Is there anything else I should be asking you?

Don’t Worship the Plan

Plan the interview, but don’t worship your plan. Effective interviewers adapt to the desires of the interviewee. Don’t be the type of interviewer who never deviates from his list of questions. I have experienced that kind of interviewer, and I wondered if he even heard or cared about my responses.

If the interviewee makes it clear that he would enjoy answering more questions, you have connected. And connection is an objective of every interview.

During the Interview

Virginia Satir, a pioneering family therapist, created an interaction model that offers interviewers insight into how to conduct an interview. Satir insightfully broke down each interview interaction into a series of steps. She suggested that careful processing of each step offered new choices for strengthening the connection between the interviewer and the other person.

Satir’s interaction model can be summarized as follows: Perceive -> Interpret -> Evaluate -> Respond.

Let’s use the interaction model to examine a portion of the interview from the perspective of Rajan, the interviewer. For example, Rajan asks Janet, “What problems did Synergy solve?”

Perceive

The first step in the interaction model is to perceive the interviewee’s response. Rajan hears and sees Janet’s response. The words are a single component of Janet’s response; other components–such as tone, pace, breathing, and facial expression–are also part of her response.

For instance, before Janet uttered a single word in response to the question about the value of a solution, Rajan noticed her eyes narrow and her forehead crinkle. Rather than rush to interpret the words, the interaction model suggests there is an opportunity to gather more data before interpreting meaning.

Rajan has the opportunity to say something like this to Janet: “I noticed that your eyes narrowed and your forehead crinkled before you answered my question. I’m not sure how to interpret that reaction. What can you tell me about it?” Regardless of how Janet responds, Rajan has gained additional and perhaps more relevant data about Janet’s response.

Janet blinks, straightens herself, and answers, “It would mean a whole lot to the department. We could process work faster.” Let’s analyze this. Notice that Janet’s words are about the value of the solution to her department rather than to herself. Without further probing, valuable data could be missed.

An effective interviewer explores how something directly affects the interviewee. That’s the subject where the interviewee has total expertise. Rajan, an experienced interviewer, then asks Janet a clarifying question, “What would a highly effective solution to the problem do for you?” Rajan might ask several probing questions to gain more specific data about the value of the solution to Janet.

Interpret

The second step in the interaction model is to interpret the data. Rajan decodes Janet’s meaning from the data he gained through his senses. Successful completion of this step happens when the interviewee agrees that the interviewer’s interpretation is the same as his meaning.

Sometimes interpretation is simple. For instance, Rajan says, “Janet, I understand that solving the problem would save you four to six hours per week. Does that capture the value of the solution for you?” If Janet says yes, Rajan is done with that question. But watch for her wanting to say more. After a long pause from Rajan, she may say, “But the most important thing is that I then could rely on the accuracy of the results.” Now Janet has revealed the real value to her.

Sometimes interpretation is difficult. Transmission errors are normal. Your perception might be wrong. The interviewee might have said something wrong and not realized it. That’s why it’s crucial to gain the interviewee’s agreement about this meaning. After you publish the findings and recommendations, the last thing you want to hear is an interviewee saying, “That’s not right,” “That’s not what I said,” or “That’s not what I meant.”

Let me suggest a method for confirming that you have captured an interviewee’s meaning correctly. Ask the interviewee a series of “Do you mean X? Do you mean Y? Do you mean Z?” questions until you hear three “Yes” answers. For instance, Janet may have provided Rajan with a lot of data about the value of the solution that doesn’t have a single simple interpretation. Rajan asks Janet:

“Do you mean the solution will save you four to six hours per week?”
“Do you mean the solution will enable you to more effectively communicate the status of the client’s requests?”
“Do you mean the solution will enable you to help your colleagues with their work and for them to help you with yours?”

A “Yes” answer confirms your interpretation. “No” answers provide opportunities for finding out what was meant.

Evaluate

The third step in the interaction model is to determine the significance of the meaning. Explore how the meaning connects to value for the interviewee, organization, and customers.

For instance, consider the response “X will save me four to six hours per week.” On the surface that sounds terrific. But how significant is that savings? During the interview with the head of engineering for an airplane manufacturer, I informed him that someone in his organization said that a new system would save each of his engineers four hours per week. He squinted his eyes and said, “So what? That doesn’t guarantee me increased productivity. They may take that time savings and stare at the holes in the ceiling tiles.”

In other words, without connecting the time savings to something else, a benefit that seems obvious at one level may not be obvious at all to a different level or perspective. Dig deeper. Ask follow-up questions such as, how would you use the time savings? Keep probing until you uncover a benefit that is meaningful to the interviewee and, if possible, to his management.

Respond

The final step in the interaction model is for the interviewer to choose the next question. You can choose to continue asking stem questions, ask the first question in a new question branch, ask the first question in an unplanned branch, or ask a metaquestion to help you decide what to do next.

If you are like me, you may have times when you aren’t sure what to ask next. I have found a comment and a metaquestion that has worked well. I tell the interviewee, “I’m not sure what question to ask you next,” and then ask the metaquestion, “What question should I be asking you?”

After the Interview

Observe And Transcribe

I suggest an interviewer use a pocket recorder so you can keep your eyes on the interviewee rather than looking at your notes. Be sure to ask for permission to make a recording, and if you don’t get it, don’t record. Throughout the interview, watch for signs that the interviewee is uncomfortable with the recording. Be willing to switch it off if it’s obstructing the interview process.

Write the transcript of the interview as soon as you can. The transcript only includes the material you may want to use in your discovery report. Share only what is relevant and needs to be confirmed by the interviewee.

I like to tell the interviewee that anything in the transcript is something that I might quote directly. I believe it’s extremely powerful to include in the discovery report quotes from people in the organization as well as customers. If the interviewee grants you permission to quote him, give him the credit for discovering a problem and how to fix it.

Don’t Share Until Approved

Don’t share information from the interview with anyone until the interviewee has given you permission. And let me be clear: I mean no one else. That includes the sponsor and your manager. You have made a commitment to the interviewee with the hope he would feel safe to share things with you. Don’t break your promise.

Adapt Your Plan

From each transcript, follow the suggestions about questions to ask other people. The interviewee gave you a person’s name because he thought that person knew something important or his thoughts had significant influence within the organization. Use this information.

Revise your question plan based on what you have learned during the interview.

Thank the Participants

Thank the participants at the end of the interview. Thank them when you send the transcript. Thank them when all the interviews are done. Thank them all in the preface to your report.

The more appreciation you show the participants, the more they will appreciate you.

Summary

Interviewing is an art. Learning how to do it effectively takes practice.

I’ve made many suggestions. If you can only do three of them, I recommend:

Building a foundation of safety so interviewees will tell you what they know.
Conducting face-to-face interviews so you hear and see what is being communicated.
Planning your questions and using metaquestions to adapt to the needs of the interviewee.

By executing a set of effective interviews, you will gain knowledge about the organization and its problems that no single person in the organization can offer you.

Remember to conduct yourself with integrity every step of the way. It’s fundamental for gaining the trust of people you are interviewing.

Originally published as Drawing Out the Facts: The Art of the Discovery Interview in Better Software Magazine, July, 2007 issue.