• UX for AI
  • Posts
  • Faster, UX Pussycat! Kill! Kill!

Faster, UX Pussycat! Kill! Kill!

Sure, AI can design, but can it kill? The new UX superpower is killing bad ideas and projects with low ROI. Welcome to the New Normal.

Sure, AI can design, but can it kill? The new UX superpower is killing bad ideas and projects with low ROI. Welcome to the New Normal. And we are doing it today, Dungeon Crawler Carl style.

Midjourney /imagine: Princess Doughnut from the Dungeon Crawler Carl series is shooting lasers out of her eyes, killing a bunch of silly defective robots (ideas).

You never forget your first kill

I still remember the first project I killed: it was during my days at a giant internet retailer. One of the PMs proposed a particularly bad implementation of the bulk shipping calculator. The vision and ROI were unclear, so we scheduled an off-site to brainstorm various ideas to see how the project might be brought to life. In preparation for the offsite, I sketched a quick storyboard borrowing liberally from the two characters from the “Money for Nothing” video by Dire Straits:

I tried to express, in a story, how the product might work. I have spent a few days thinking and 2-3 hours over the weekend sketching the storyboard. As soon as the workshop attendees saw the storyboard on a single handout slide, they understood the implications, and the project was dead, just like that. I killed it. It was a beef-brained idea: an expensive, large dev and marketing investment, no integration with existing features, no clear onboarding, hard to learn, low user benefit, and no clear ROI. It was a classic “Money for Nothing” early-2000s internet bullshit. Making this into a UX storyboard helped everyone get an immediate, visceral understanding of just how bad of an idea it was.

In the age of smart machines that easily generate multiple ideas in just a few minutes, 

The UX super-power is not just coming up with new ideas. It is often our ability to utterly kill, destroy, and stomp out the bad ideas that is of the biggest value to our team.

How many projects have you killed last year? 

So far this year, I have killed five large projects, and frankly, I wish the number was much higher. 

If you answered “none” to how many projects you killed this year, you might want to rethink how you work. 

Are you:

  • Just “painting by numbers” designing based on a fixed set of requirements provided to you by the PMs? 

  • Designing without business and dev numbers at all, based instead on some heroic personal conception of “creating a good experience,” no matter the implementation cost? 

  • Routinely recommending complete re-writes of entire existing product features with a bunch of corner cases?

  • Have fewer than 2 meetings with customers per week?

Well, I got news for you – that party is over. Dead. Pow. It might not be this month or the next, but let me promise: that type of UX work is done

On the other hand, if you:

  • Consult developers early in the process to ascertain that the idea has a positive ROI

  • Continuously validate with customers at every step (storyboard → paper mocks → high-def mocks → edge cases & specs → UX QA → post-deployment) using RITE (Rapid Iterative Testing and Evaluation) 

  • Understand how much every feature costs to build and maintain

  • Know specifically how much customers will pay for a specific feature 

  • Can recite from memory which specific KPIs the feature you are working on is supposed to improve

  • Can tell a clear story about how that feature will impact customer acquisition, retention, and lifetime customer value

  • Routinely exploring lower-cost alternative designs for every feature, not just going with your knee-jerk/darling design/sacred cow idea using Disciplined Brainstorming/Bookending 

  • Actively reusing, reducing, and recycling design patterns and components.

You just might be on the right track.

If you are doing UX right, you should be routinely killing multiple pages, features, and entire projects worth of crap and bloatware that should never see the light of day.

Get out there and Kill, Kill, Kill!

Some product leaders have already started to use LLMs like ChatGPT to come up with more ideas… As though there was ever a shortage of ideas! There has never been a shortage of ideas – there is, and always was, however, a severe shortage of good ideas and execution.

And this has never been more true than it is today. 

85% of AI projects fail. 

A clear evidence of overabundance of bad ideas, if there ever was one.

Today, companies in all manner of industries are trying to do more with less. You need to get on board with saving money and effort, and that means specifically focusing on killing bad ideas, negative ROI efforts, and bloatware. One of the best ways (and a requirement for AI projects IMHO) is to start with a vision prototype.

You also need to work closely with users and stakeholders: run 6-8 user research studies every week and have multiple weekly four-in-a-box meetings with PMs, Developers, and Data Scientists.

You need to establish a routine process with your team to kill bad ideas as quickly, cheaply, and efficiently as possible so that your team can focus on good ideas. 

And you need to focus on killing ASAP. 

To use the inspired words of Matt Dinniman, who penned the incomparable Dungeon Crawler Carl series (and inspired the opening image of this article featuring Princess Doughnut): 

“Get out there and kill, kill, kill!"


Greg Nudelman (with Daria Kempka, Contributing Editor) 

Join the conversation

or to participate.