Process and Prototypes
This weekend I was hanging out with Asher, my 2.5-year-old son. After hard-boiling and peeling some eggs, I tried asking him to clean up the eggshells. After breaking it down for him, he diligently got 90% of them in the trash can and said he was all done. I asked about the remaining pieces on the ground, to which he covered them with Tupperware lids and said: “can’t see it.”
Shenanigans aside, I was reminded of the Exact Instructions Challenge, as at 2 years old every task becomes a game of specific steps. As software developers, this is our challenge every day. We codify complex processes, procedures, and functions and assign them to workers--the computers tasked with executing those exact steps. But what happens if we, as developers, get those steps wrong?
At Vaporware, we’re huge proponents of the Lean Startup process: minimizing the build and release cycles directly with users so we as developers can verify our procedures and correct our mistakes as quickly as possible by measuring the outcomes directly with usage. This works great with easy to create activities, like filling out forms or presenting data, but there are a few cases this doesn’t work well.
For one, the Lean approach works well in Summative research, when we already have an answer and want to check the validity of it. But Lean does not work well in Generative or Formative research--when we want to understand current problems and form initial ideas. Check out this article from our friend Basia Coulter for a deep-dive into the differences between these research methods.
For those cases, we turn to Prototypes as a solution to quickly validate conceptual problems before we suggest investing more. But creating a prototype is only one part of the step. You have to test that prototype with actual users and observe their behavior. This user testing can be a challenge, especially when you’re not able to observe the exact usage of the prototype. Luckily Joshua Rudd, our Design Lead, has a fast and cost-effective way to do this remote user testing. Read his article to learn how to do it yourself. Through a combination of modern online tools and proper implementation, we’re able to get amazing access to better results at a fraction of the time.
The other case where Lean doesn’t work well is with non-UI experiences. With some more modern tooling, we’re able to prototype Voice Assistants, Chat Bots, and even Machine Learning efficacy before investing extensive time and money in a custom solution.
Do you have an idea that could be validated with the right Prototype?