Hiring is a process to be designed
Let's talk about finding the "best designer for our team."
I've seen a recent swell in writing about "crushing design challenges" and a handful of write ups from people who spend days on interview homework problems. At the same time there is a rise in developers writing about how screwed up these practices have become in their own interviewing processes.
As someone who as been tasked with hiring designers, I understand there are a lot of questions you want to answer about a candidate. As someone who has been hired many times as a designer since 1999, I can tell you that it wasn't until about four years ago that I started getting regularly asked to do speculative work, homework, or on the spot designs as part of the interview process. I've noticed it again in my current job hunt, and while I hope this doesn't disqualify me as a candidate, I do think the interview process is something design leaders should be talking about, analyzing, questioning, and improving. So I'm talking about it now, and I hope you will join me in the conversation.
You may think that giving someone an assignment to do is the best(or only) way to evaluate a design candidate. It might be the way you were evaluated for your job. Learning more about how someone works isn't necessarily a bad goal, it just depends on how you go about it. The following is a set of questions most hiring managers tell me they are always trying to answer in the interview process. Most of the time they go about this with a homework problem or occasionally an on-the-spot "design challenge" even when those things don't actually answer the question they have about a candidate. And the truth is that's most of the time.
It's important to remember that the most important question you are trying to answer with your interviewing process is, "Is this person the best designer for us?" You can't objectively evaluate "the best designer" like you might "the best microwave." All of these other questions are only to try and help you answer the big one: "Is this the best designer for us?" so lets get some better data and try to answer it.
Hiring Manager Question: "How does this candidate approach solving problems?" and/or "Can this candidate work with our team?"
This is what I hear the most from people when they tell me why they use "design homework" as part of their interview process. Here's the question I always ask in response: "If they're doing it all alone - at home - how does it tell you any more about their approach to problem solving than any of the case studies in their portfolio?"
If you are trying to use homework problems to understand someone's approach to solving problems, you should be able to explain why their other work completed out of your view isn't telling you what you want to know. And you should also be honest if that is your fault or theirs.
If you're thinking, "But we give them a problem to solve that has to do with our product, that way we can see how they'll solve our problems!" I'm going to stop you right there. This idea sounds so so good when you say it, but the reality is the same as the idea of fat-free ice cream or whole-grain Goldfish crackers being "healthy," it's just not true.
Most importantly, the candidate has no idea about: your company and the details of your business, the people on your team, the things you've already tried, the things that are politically inappropriate to propose, the current state of the system and why it's designed the way it is, or your users. All of these are essential in solving YOUR problem, so at best what you're doing with this kind of assignment is hoping they guess correctly. Congratulations, you're filtering your candidate pool for good guessers. Good guessing is not a valuable design skill.
Additionally, by using work you actually have to do, you're asking for speculative work. "Just solve this one problem for us for free, and if we like it we'll totally pay you for the next designs." I again point to the NoSpec site to explain all the reasons spec work is bad, for anyone, not just designers.
Better ways to answer questions about approach to problem solving and team fit
The best way to actually answer questions about someone's approach to problem solving is to watch them start to solve a problem. Better yet participate in their solving a problem.
Schedule a brainstorming meeting with your design team while the candidate is onsite and have them attend. On the developer side of things, forward-thinking companies are having engineering candidates pair-program for a few hours with an existing member of the team. There is no reason why something similar couldn't happen as part of the process of hiring designers.
Send the candidate a brief a few days ahead of time giving them topics to prepare for a workshop they will run the day of the interview. This means thoughts and notes, not wireframes, not prototypes, not mockups, thoughts and notes. If you want to see how they work with your people, have them working with your your people.
Schedule time for a developer and a product manager to join them and have the three brainstorm a redesign of something from everyday life. My personal favorite is "What would the public library be like if it was a SaaS startup?" but you can do this with almost anything. The point is to see how they approach the problem and work with a PM and an engineer. Save everyone some time and ask the candidate to give a full 45–60 minute presentation on one of the pieces in their portfolio. The reality about design homework is that it doesn't tell you anything more than a portfolio case study, which your candidate has already put together, even if you intend to have her present her solution. Every good design candidate can present at length about a previous project, even on very short notice. You can also ask questions about how they worked with their former coworkers, which you can't do with your homework problem they worked on in complete isolation.
Hiring Manager Question: "Can the candidate really do work like the work in their portfolio?"
Most times this question is about filtering out the candidate's individual contribution in a project. Most of the time when you ask this question, it means the portfolio pieces aren't clearly presented or they're missing important details.
At the extreme end, this question it's really asking if someone actually did the work in their portfolio. In these rare situations, I am always amazed at how often a hiring manager is afraid someone forged their portfolio, but not afraid that same person would forge an unsupervised homework problem. If you have serious trust issues with a candidate before you've even hired them, that's a sign that you should not hire them. Why would you even consider moving forward with a candidate you think is lying to you?!?
Better ways to answer questions about individual contributions
The most effective way to answer questions about an individual's contribution to a project is to have them present the work in more detail and ask them what they specifically did. If they can't do this or you aren't satisfied with their answers, you have your answer.
Ask them to see more work. Every designer, even a student, has a stack of work that didn't make the cut for their portfolio. Ask them to present some of it. When you do this, be specific about what kind of things you're looking for in the work so they can show you pieces that actually answer your questions.
Hiring Manager Question: "Can the candidate do the kind of design work that we want her to do for us?"
The answer largely depends on what you mean by "the kind of design work we want."
If you mean you don't see something in their portfolio that looks like what you envision your future design to be, you are treating design like fast food and the designer's portfolio's like the Value Menu. You shouldn't be looking at a design portfolio as a menu to be ordered from. The reality is, at best, the designer's control of the final design of any even moderately complex project is maybe 30%. Unless you plan on hiring all the other people that were involved in the project, it's unlikely you're going to be able to get the same exact same design for your company. If this is the question you're asking: Go read chapter 2 of Mike Monteiro's "You're My Favorite Client". Actually you can read part of it online, then go buy Mike's book because it's really good.
Most of the time this is a question about wanting to see a specific kind of design work (e.g. logos, on-boarding flows, identity systems, pattern systems, etc.) that aren't present in the portfolio. The worst thing you can do to answer this question is ask someone to "just do something real quick" because you're not going to get a very good example of their ability in a rush job. Think about the quality of information you want to have to make your decision.
Better ways to answer questions about specific kinds of work
Ask the candidate about the lack of a specific kind of work in their portfolio. It might be that they have some examples that didn't make the cut for their portfolio (remember, EVERY designer has this cache of work that you never see.) Or it could be that they really don't have any experience in that area but they want to gain some. Then you have to decide if you want to hire them to learn based on the other examples they have shown you. It could also be that they hate doing that kind of work and so they are a bad fit, which is valuable information for both of you to have. Asking can save everyone a TON of time.
If it turns out the designer doesn't have any examples of the kind of work you're looking to hire them to do, this is one of the few times a homework problem makes sense, as long as you follow some rules: Make the assignment small (e.g. don't ask someone to build a style guide from scratch), make sure the problem is general enough that there is no fear of their work being stolen, and make sure you give them a realistic amount of time to do the work. A more general design homework problem can be turned into a portfolio piece, if you don't hire them, and they continue refining it and developing that specialty.
There are some questions you're not going to be able to answer with a homework problem or an on-site challenge, so stop trying
Remember, "Is this person a good designer?" is not a question you can answer by testing people with a single assignment. This isn't even the right question to be asking about a candidate. Being a "good designer" is a complicated mix of qualities. Soft skills are some of a designer's most important. The important question to answer is, "Is this person the best designer for us?" The designer who is "a good designer for us" is one that works well with your team, the processes you have set up, and is interested enough in what you're doing to move your whole design practice forward. You can't find all of that out with a homework problem or a white-boarding exercise because it's about more than some hastily assembled design artifacts. Plenty of "not good for us" designers can crank out great wireframes or mocks. You're going to have to spend some time getting to know them as a person to really evaluate if they are "good for us." The same rules that apply about knowing your users apply double to anyone you consider adding to your team.
A last and important thought: Hiring is a skill and it's hard to master. Just like every other skill you learn by doing it and by learning from your mistakes. You will never be able to completely avoid making a bad hire, but you can seriously reduce the chances if you are thoughtful about your hiring process.