Improve design iteration with crazy checks

The bottom line: Iterate more and faster to design better experiences. Here's an easy method I use with my teams to increase iteration.

Enter your email and I'll send a quick note when the next post is up.

You're set up for efficiency. You're using your favorite tool. You've memorized the keyboard shortcuts for those small tweaks you always make. You have a set of templates and reusable widgets, so you can drag and drop common design patterns. You spend the majority of your time focused on the tricky bits: the messaging, information design, the triggers, thinking about your users. You spend a day working and re-working the wireframe. It's beautiful. It solves all of your data problems, meets business needs, and helps your users accomplish their goals.

You share it with your team. It's just wrong.

You did everything right. You've got the process down, and you just wasted six hours.

Why? Because you spent six hours exploring the wrong concept.

It's not a total waste. You've learned. Now you can iterate. You've built. You measured, and you learned. What if you had learned the same thing in one hour instead of six? What if you had five of those six hours back to spend on the right concept?

Iteration is learning. How can you learn faster?

Crazy checks help you learn faster

That great idea you have? Are you brilliant? Or crazy? That's the question you ask with a crazy check. With a crazy check, you show a concept earlier to make sure you're not crazy before you spend more time on it.

Crazy checks combine three core concepts that drive the UX process:

  1. Iteration: how often can you measure and learn, make and check your building and thinking 
  2. Fidelity: how perfect, or "real", does something have to be before you can check and learn from it
  3. Critique: how can you structure your checks to make them more efficient
My crazy team (from left to right): Ryan, Joshua, and Shima

My crazy team (from left to right): Ryan, Joshua, and Shima

I've implemented crazy checks with numerous project teams over the last several years, and the response has always been the same. It feels a little weird at first, but it's awesome, and after a while, there's not other way you want to work.

To conduct a crazy check, you rough out an idea, or several, and show them to someone close to you. They provide feedback as to whether or not you're crazy or should continue exploring that idea. That's it. It's that simple.

Iteration is learning

You learn when you iterate. Your team learns when the team iterates.

Iteration happens when :

  • you think about something and have an idea
  • form the idea and make it into something you can share
  • share it with someone to check the idea by receiving feedback 

Think, make, check, and build, measure, learn aren't just lean UX and lean startup concepts. Those are design concepts that describe the process of iteration, of learning, so you can create better experiences.

More iteration, more learning

It's pretty human to want to avoid rejection. As a designer you have the urge to not show something until it's finished, until it's just right. Unfortunately, it takes longer to finish something than to start something. If you wait to get feedback on your idea until it's finished, then it will take longer for you to get feedback.

It's not that you waste time working on a brilliant idea that's really just crazy. When you take longer to get feedback, you take longer to learn. Until we invent time machines, taking longer to learn means you have less time to improve an idea before it has to ship.

If you're not learning, you're not designing. Design is about validation.

Increase how much you iterate

It takes longer to finish something than to start something. If you wait until it’s finished to get feedback on your idea, then it’ll take longer for you to get feedback.

The easiest way to learn more frequently is to iterate more frequently. The easiest way to iterate more frequently is to reduce the time it takes between iterations. You have to spend less time on each iteration.

This is where the crazy check comes in. As soon as you have a concept together enough that you can show it to someone, you show it to them for feedback. Before you spend six hours on the concept, you spend one hour.

This goes entirely against your need to avoid rejection, to show something half-baked, that's ugly, riddled with typos. That's why we position these reviews as crazy checks. Don't ask if you're purposefully half-baked idea is awesome. It will never be awesome. Go in expecting failure. Open the door for your team mates to be critical and ask, "Am I crazy?"

How crazy? Be crazy like a fox

Just enough crazy with Michael and Jessica

Just enough crazy with Michael and Jessica

You can't create a fully functioning prototype, for an entire website, from scratch, in an hour. That's just crazy. Too much crazy. To do a crazy check you need just the right amount of crazy. 

To share just the right amount of crazy with your team mates, you have to really focus on the one important crazy idea you want feedback on.

What specific problem do you want to solve first? What's the first problem your user needs to figure out? Or, what's the one problem you have with the data? Or, what business goal really hamstrings your design?

That core problem is the first idea you want to crazy check.

One you know your problem, you have a sense of exactly how much work you need to do check your solution. Lower fidelity is good for concepts, and faster to create. You want higher fidelity when you want to work out the details.

Do as much work as you need to do in order to validate your concept. Anything more is waste. Add more and more detail as you refine your idea more and more.

Do as much work as you need to do in order to validate your concept. Anything more is waste. Add more and more detail as you refine your idea more and more.

On a project I'm working on now, we want to have one detail page for a location. The problem is, we have two entirely different groups of users who want entirely different sets of information from our one detail page. The problem we have to fix is how do we prioritize one screen for two incompatible priorities?

For more information on fidelity, check out my  Hacking UX ebook on Slideshare .

For more information on fidelity, check out my Hacking UX ebook on Slideshare.

To check some approaches, we don't have to spend six hours wireframing everything out. We can take 30-minutes, do a handful of sketches, and then show them to the team for some feedback.

Is it better to learn how to refine our approach after 30-minutes? Or is it better to learn how to refine it after six hours? That's easy. Spend 30-minutes, ask if we're crazy, and then refine.

Focus your crazy

More often than I'd like, I show a wireframe, and one of the client stakeholders will ask why the text is wrong or why the images are wrong or why the logo is gray. I want feedback on layout and structure, and I'm getting feedback on visual design.

The same thing can happen when you show things to your team. You have to learn how to focus feedback, so you spend time discussing the problem you're trying to solve, and don't waste any time discussing problems you haven't gotten to yet.

For more on framing feedback, check out my post on  outcome-driven meetings

For more on framing feedback, check out my post on outcome-driven meetings

The trick is to be specific about the feedback you're looking for. Explain the problem you're trying to solve, show your crazy, and then ask whether your approach is crazy, or not. When you frame your feedback, you help your audience help you evaluate the problem and the solution.

Using our example from above, we would tell the team we're trying to figure out to help each user without getting in the other user's way and show sketches of the page layouts.

Get Feedback in layers

As you refine your idea more and more, you can starts asking more and more people how crazy you are. 

Your first feedback is always yourself. You finish something, sit back, tilt your head, and ask yourself if it's right. It's usually wrong, so you lean back in and make some changes or redo it before leaning back and taking another look.

Your second layer of feedback is always someone else. Poke your head over the cubicle wall, grab someone walking by, or take a pic with you phone and email it to someone. Does this look crazy?

On the project I'm working on now, we have seven layers of feedback.

  1. Ourselves
  2. Crazy check with the UX team (conversation or email)
  3. Crazy check with our company team (conversation or email)
  4. Crazy check with our agency partners (email)
  5. Review with our direct stakeholders (email or quick meeting)
  6. Review with the end customer (quick meeting or formal review)
  7. Review with users (formal usability testing)
Your feedback audience kind of grows from just yourself to including everyone.

Your feedback audience kind of grows from just yourself to including everyone.

On any project, you can institute the same kind of layered feedback. Show rougher sketches to those closer to you, and slowly show more refined work to people who are farther away.

Our project plan schedules two formal reviews and two rounds of changes to make sure we can deliver according to the expected timeline. If we stuck to the project plan, we'd only iterate twice. We'd only learn twice. 

By incorporating crazy checks and a layered approach to feedback, we iterate at least seven times, usually more. That means, we learn seven times. We refine seven times. We improve the user experience seven times.

Design is about validation

Design is about thinking, making, and checking; building, measuring, and learning. The goal of any design activity is ultimately to validate the design. You will iterate and validate over and over and over. The more often you iterate, the more often you learn, and the faster you iterate, the faster you learn, the faster you get to a better experience.

So, here's your mission, should you choose to accept it. This week sometime, you are going to design something. And you are going to get a crazy check.

  1. Identify the main problem you need to solve
  2. Limit yourself to 30-minutes to generate a few options. Sketch them on paper. Make them ugly little buggers you're embarrassed to share.
  3. Share them with someone you trust. 
  4. Ask if you're crazy.

Start conducting crazy checks for everything. Layer your feedback, so you expand your audience with each crazy check. At first, it might feel weird showing unfinished ideas and concepts, but after a few crazy checks, I promise, not only will it feel normal, but it will become second nature and a vital part of your process.

If you're feeling really adventurous, phrase your problem in 140 characters or less and post your crazy check to Twitter with the hashtag #uxcrazycheck. You're already primed for the worst answer: yes, you are crazy, and it looks like crap. Every bit of good feedback is just tons of icing on your design cake.