I have finally read First, Break All The Rules by Marcus Buckingham, which has been on my to-read list for a while. If I wanted to sum up its essence in one sentence, it would probably be:

Everybody has a unique combination of talents which is extremely hard to change, so, as a manager, you should do everything you can to leverage this uniqueness instead of working against it.

While the book has other great insights about what it means to have a strong workplace, this is the central theme that can be found throughout the book. Talents are part of a person’s personality. More specifically, talents are defined as recurring patterns of thought, feeling, or behaviour. They are hard-wired into a person’s brain, and they are enduring, i.e., almost impossible to change.

Buckingham contrasts talents to knowledge or skills, in that you can teach talents only to a low degree, while skills and knowledge can be acquired or taught quite effectively. Examples of skills, knowledge, and talents include:

• Skills: Arithmetic (for accountants), how to give a safe injection (for nurses), how to commit changes to a git repository
• Knowledge: The order of the parameters of a function, FAA safety regulations (for flight attendants), the rules of double-entry bookkeeping
• Talents: Empathy, strategic thinking, the ability to win others over, the ability to recognize patterns

While skills and knowledge have the advantage of being transferable from one person to another, they are often situation-specific. For example, if you can teach somebody to adhere to the Single Responsibility Principle in the context of the data access layer, but the same person then goes and mixes responsibilities freely when working on the controller, there might be a lack of talent in that area. While the person is able to recognize and apply the principles and patterns of software architecture in a few known use cases, he will feel lost in a lot of other situations. With an unanticipated scenario, skills and knowledge lose much of their power.

Talents, on the other hand, are very hard to transfer from one person to another, but they are transferable from situation to situation. Given the right mental stimulus, they fire spontaneously. If you have a talent for software architecture, you will apply its principles regardless of the domain you are dealing with. If you have a talent for organizing data in an optimal way, it will not matter to you if you are dealing with a relational database or a JSON file: You will look for relationships, patterns, and ways to find a good structure.

Three kinds of talent

Buckingham further distinguishes between three kinds of talents:

• Striving talents give answer to the question: What drives this person? What motivates her? Is she competitive? Does she want to make the world a better place? Does she want to reach mastery in her craft?
• Thinking talents are about how a person thinks and organizes himself. Is he a focused, structured person, or is he creative and a bit chaotic? Does he think strategically and far ahead into the future, or is he only interested in the here and now?
• Relating talents explain how a person interacts with other people. Does she easily mingle with strangers, or does she feel comfortable only with her closest friends? Does she stay calm and rational in conflict situations, or does she avoid them at all cost? Does she easily trust people, or is she generally mistrustful?

Implications on recruiting

Because talents are very hard to transfer from one person to another, it logically follows that you should hire people who already have the right set of talents for the role. Otherwise, no matter how much time and energy you invest into “fixing” an employee, she will never reach excellence in her role - while your relationship with her might seriously suffer (see Enhancing Strengths).

Buckingham gives some very useful generic advice related to recruiting. Below, I will explain the methods I find the most interesting, promising, and practically applicable. Moreover, I have tried to transfer the ideas from the book to the field of software engineering.

1. Ask some open-ended questions and then remain quiet

The idea with open-ended questions is that the candidate has to make some decision about the direction in which he wanders when answering. This decision tells you something about his way of thinking, and about certain preferences. Quote from the book:

“The talent interview should mirror verbally what will face him on the job behaviorally. On the job, he will face thousands of situations every day to which he could respond in any number of ways. How he consistently responds will be his performance.”

Examples for such open-ended questions are:

• How closely do you think people should be supervised?
• What do you enjoy most about software development?
• How important is software quality?
• How do you feel when there is a bug and you haven’t found it for hours already?

It is important to ask the question as neutrally as possible, without somehow hinting at the “right” or desired direction. Ask, then remain silent. If the candidate asks you to clarify, just don’t. Tell him that it is important what he makes of the question, and how he interprets it. The more unaided his answer, and the more spontaneous, the more predictive it is of his future behaviour.

Buckingham adds one more piece of advice: No matter what the candidate answers to such an open-ended question, believe him. If you ask him what he enjoys most about software development, and he never mentions problem solving or creating something, believe him, even if you would like to hear something else.

2. Listen for specifics

This is the “tell-me-about-a-time-when” category of questions. Since past behaviour is a good predictor for future behaviour, we are interested in past behaviours that are correlated with the talents we are aiming for. For example, a talent for strategic thinking might be revealed by: “Tell me about a time when you made a tough choice that played out in the long run.” Buckingham gives us words of warning when using these questions, though:

• Only consider answers that give a specific example, ideally including time, the involved persons, or other details. If the candidate gives you a general, eloquent narrative about the topic, without providing any specifics, it does not count.
• Give credit only to the person’s top-of-mind response. Talents are recurring patterns of behaviour. If the behaviour is recurring, the candidate should have no trouble remembering an occurrence at only one prompt.

If you are aiming for the talent Restorative, which stands for people who have a drive to make things better, you might ask something like: “Tell me about a time when you fixed or improved something without somebody else telling you to do it.” (This question is inspired by Kate Matsudaira’s Epic Interview Questions List). Here are two possible answers:

1. “Well, I think refactoring falls under that category. And refactoring certainly has its value, because it keeps code from rotting, and enables you to keep delivering new features without slowing down over time. Of course, you have to balance refactoring activities with actual feature development, or else you will be perceived as unproductive. But without refactoring, your development will keep getting slower. So I prefer doing small refactorings regularly, instead of doing some big-bang refactoring once a year or so.”
2. “Well, just last month I realized that our documentation on the development environment was really frustrating for newcomers, because it was disorganized and incomplete. I took some time to update it, and make it better accessible.”

While there is nothing wrong with answer 1, is rather generic, and not about something specific the candidate did. Therefore, it is not predictive of future behaviour. Answer 2 offers a concrete example, even within a rather recent time frame. That’s the kind of answer you are looking for. It is a clue that fixing and improving things is a recurring part of the candidate’s life.

Another word of advice: If you get an answer like answer 1, you should refrain from follow-up questions, like asking the candidate to be more specific, and tell exactly what happened. Buckingham even goes so far to say that “This is the cardinal sin of interviewing”, because:

“Regardless of the detail the candidate eventually provided, if she needed two or three probes to describe a specific example, then the chances are that the behavior in question is not a recurring part of her life.”

In other words, if you judge the candidate by how much detail she provided, you will end up evaluating if the person has a good memory or is articulate, which is not what you were aiming for. So, whatever you do, only take top-of-mind answers into account, and only consider specific examples.

3. Satisfactions

Instead of choosing your questions carefully to reveal certain talents in the candidate, you can also do the opposite: Ask extremely broad questions, and see if the candidate reveals her talents for you. One way to do this is to ask about the candidate’s satisfactions, the “psychological oxygen” she breathes.

Great accountants smile when the books balance, great nurses feel satisfied when they know they took good care of all their patients, and great software developers are happy when all tests pass. Asking for satisfaction is rather straightforward. Example questions are:

• What do you find fulfilling?
• Which kinds of situations give you strength?
• What is your greatest personal satisfaction?

The answers to these questions can give you valuable hints for finding a good role fit for this candidate. Doing what you do best every day is extremely beneficial for employee happiness, and these employees tend to be extremely valuable to the company.

4. Separating questions

The last category of questions is an especially interesting one. Buckingham introduces them as “questions where great managers know how the top performers respond”, and he gives an illustrative example.

The question “How do you feel when somebody doubts what you have to say?” is answered differently by top salespeople than by average salespeople. Top salespeople are personally invested in what they are selling. They care deeply, and take rejection personally. If somebody doubts what they say about the product they are selling, to them, it is like they doubt the integrity of their character. This feels horrible to them. Average salespeople, on the other hand, are not as personally invested, so the thought of being doubted does not upset them so much. They will answer the question differently.

The very same question can also be used on teachers, but the associations they have will be different. Outstanding teachers will think of bright, keen children who develop thoughts of their own, and do not just believe everything they are told. To a great teacher, this is the most beautiful thing in the world, so their answer is: “I love it!” Average teachers don’t love being doubted like great teachers do. In fact, they rather dislike it, because they interpret it as a challenge to their competence.

Since Buckingham does not really provide a name for this category of questions, I will call them “separating questions”, because they can separate outstanding candidates from average ones.

Identifying separating questions

The tricky part about separating questions is identifying them. They do not work universally: “How do you feel when somebody doubts what you have to say?” might work with salespeople and teachers, but it will not work with nurses, just to give an example. The greatest nurses will not answer in a consistently different way compared to mediocre nurses. Can this question separate excellent software developers from average ones? I am not sure, but I doubt it. I plan to test a variation: “How do you feel when somebody criticizes your work?” However, I do not really know what to expect.

If you want to find more of these questions, there are at least two things you can do:

• Study your best: Ask people on your team or in your department. Collect the answers, or make notes. Find out if your most productive, most talented colleagues answer consistently differently to a question compared to “average” colleagues.
• Experiment in interviews: Do the same with candidates who come in for interviews. Keep a record of their answers, along with the hiring decision you made, and - in case you hire them - their development on the job.

Both of these methods require patience and diligence, but the result can be very valuable. Imagine you had good separator questions for one striving, one thinking, and one relating talent that you are aiming for: Your interviews would be so much more efficient, and the confidence of your hiring decision so much higher.

There is also danger, though, in over-relying on your questions. People are complex, and maybe there are “grey” answers to any separating question. Answers that are difficult to judge. Therefore, you should not trust a question’s power blindly, or you might end up with false positives or false negatives.

Finding candidates

Moreover, while we were talking about how to verify candidate questions, we took the second step before the first, because: Where do we even get these candidate questions?

Again, this is difficult. The only thing I can think of is using creativity and common sense. You are looking for a good problem solver? Have some problems ready to solve, and watch how the candidate responds. Is she enthusiastic about the tasks, or rather disinterested? Or, are you looking for a quick learner? Ask him how he feels about the current explosion in JavaScript tools and libraries. Does he feel uneasy about it, or does he see it as an opportunity?

Another thing that can help is a personality assessment test that you take with your entire team. Again, use the results to identify the strengths or talents of your best colleagues. Interview them, and talk about their talents with them. How do these talents manifest themselves in everyday life? In certain situations at work? Conversations like this can lead to good candidate questions, which you can then verify.

Conclusion

Currently, I am interspersing some of the questions mentioned above throughout my interviews: Open-ended questions, “tell me about a time when…” questions, questions about satisfactions, and separating questions (or, at least, candidates). While it is still early to draw conclusions, I believe some of them work quite well. I will keep you updated when I have better data. Meanwhile, I would love to read your comments: Do you know good interview questions for some of the question categories?