The Limits of Prototyping / by David Sherwin


A few months ago, I saw a presentation by two interaction designers about the benefits of prototyping. During the Q&A, they received the following question from the audience: “So, at what point are you done with prototyping?”

“I don’t know if I want to get philosophical about this,” one of the designers said. The other designer mulled the question for a few more seconds, then responded:

“You can treat everything as a prototype. You’re always trying new things and learning. So I don’t think the prototyping process ever ends.”

This wasn’t the first time I’d heard this point of view in the past year. In Kathryn McElroy’s recently published book Prototyping for Designers: Developing the Best Digital and Physical Productsshe states in the first chapter:

“Everything can be prototyped, and everything is a prototype. There can always be a better, improved version of what you are creating, and it takes time and practice to develop that urge to always be prototyping.”

She draws upon the original definition of prototype (n.) in the OED: “A first, typical or preliminary model of something, especially a machine, from which other forms are developed or copied.” She recognizes that there’s tension in the design community about whether we should be using a more focused or a broader definition of prototyping in our work. Her suggestion is to keep the definition broad:

“If you choose to see everything as a prototype, interactive or static, and use it as an opportunity to test your assumptions through any means possible, you will develop the mindset of incremental improvement and constant feedback that will greatly benefit your product.”

I wholeheartedly agree that testing prototypes is a great way to disprove risky assumptions and hypotheses when working on products. Prototypes are useful for soliciting early customer feedback and iterating on your product design. You may need a prototype to fully understand an idea and communicate that idea to your stakeholders. Creating prototypes can help you create the right design and ensure that the design is right. Prototyping is a fundamental activity in the design thinking process, and it has even appeared in recent self-help books as a way to experiment with major life decisions.

But I disagree with the notion that “everything can be prototyped, and everything is a prototype.” If we start looking at everything in the world as a bunch of models and treat them as such, we are creating problems downstream for our teams and our products.

Think about your product managers. They have to make sure their products consistently deliver customer and business outcomes. A prototype may be acceptable for gathering feedback from customers about a product’s usefulness or desirability, but a product manager wouldn’t allow such a thing to be a formal release. That would be poor product management.

Think about your engineers. They do prototyping in order to understand how to use a technology. They also may help you in creating prototypes of your designs to test with users. The designs may look so good that everyone thinks they’re ready for release. However, your engineers will need to move beyond these prototypes and fulfill the necessary engineering and quality assurance work in order to ship a durable product. Otherwise, that’s shitty engineering.

Think about your customers. What are you promising to them if you’re selling them ‘just a prototype’?

No matter whether you’re improving a feature for a heavily used mobile app or inventing a new kind of snack mix that you can purchase at the airport, you’re going to reach a point where the prototyping process has to end. By their very nature, prototypes lack the formal constraints that a final design implementation requires. Prototyping should always be considered a modeling activity, not formal engineering or construction of products and services. If you spend all of your time prototyping, you may not be thinking through the implementation details that are needed for dozens or thousands or millions of people to make best use of your products.

Then there’s the notion that the prototyping process never ends. Bill Buxton, in the final chapter of his classic book Sketching User Experiencestalks about how designers go too far with prototyping, and end up confusing a prototype with design or research itself. In his words:

“Without the accompanying analysis, reflection, projection, hypothesis formulation, and testing, what one gets is a shotgun-type splattering of isolated demos that are far more suited to fundraising than they are to constructing a body of knowledge, or experience on which we can build anything solid.”

The key words here are constructing a body of knowledge. Prototyping can help you come up with ideas, model them, test them, and evaluate what you learned. But prototyping serves a goal in product development: Making team decisions that help you confidently fulfill customer needs with shipped products.

The above goal can be fulfilled in many ways—and McElroy is speaking to it when she says prototyping contributes to “the mindset of incremental improvement and constant feedback.” I would rather teach designers and product team members to bring an experimentation mindset to their work, and to recognize and challenge their assumptions through a variety of methods, whether it’s rigorous critique, conducting foundational and generative research, or creating and testing prototypes with users.

Prototyping doesn’t drive experimentation or learning. Prototyping serves it. We should not be wishy-washy about this.