Eric Morrison is a User Experience Research Lead based in New York, New York. He has spent more than 14 years leading research for companies including Google, TikTok, and Disney, and currently focuses on how artificial intelligence can support creativity and teamwork inside organizations. This conversation covers how he makes decisions when the research isn’t finished yet, and what he has changed his mind about along the way.
You’ve said good research takes time. What happens when you don’t have it?
That’s most of the job, honestly. People imagine research as this clean process where you run a study, wait for the data, then hand over a verdict. In practice, half my work is making a call with sixty percent of the picture because the product decision is happening this week whether I’m ready or not.
What I’ve learned is to separate the decision from the confidence level. I can tell a team, “Here’s my best read, and here’s exactly what would change my mind.” That’s more useful than waiting for certainty that may never come. Withholding a judgment because it’s incomplete is itself a decision, and often a worse one.
Is there a moment that taught you this?
Early on I treated every recommendation like a thesis defense. I wanted the sample size right, the method airtight, every alternative explanation ruled out. That instinct isn’t wrong; it’s just badly timed for how product organizations actually move.
I remember pushing back a launch decision because I wanted one more round of interviews. The team waited. The extra round basically confirmed what we already knew. That was the moment I realized rigor and speed aren’t binary, but they do trade against each other. And I’d been treating that trade as if it didn’t exist.
How do you decide when a smaller signal is enough to act on?
I ask what the cost of being wrong actually is. If a bad call means a button is in a slightly worse spot, I’ll act on a handful of interviews and a strong pattern. If a bad call reshapes how people collaborate for the next year or what portfolio investments leadership makes, I want more before I sign off.
That sounds obvious stated out loud, but it’s easy to forget in the moment, especially when a team wants an answer, and you feel pressure to sound certain. I’ve had to get comfortable saying “I don’t know yet” to smart, impatient people. It used to feel like a failure. Now I think it’s one of the more honest things a researcher can say.
What’s an example of a call you got wrong?
I once read a drop in engagement with a collaboration feature as a sign people didn’t want it. The interviews I had pointed that way. What I missed was that the feature had shipped with almost no explanation, so people were avoiding it out of confusion, not disinterest. I recommended cutting something that just needed a better introduction.
Has that changed how you frame recommendations now?
Yes. I try to attach an expiration date to my own conclusions. Something like, “This is my read based on what I know at this moment, and this is the new information or context that would make me revisit it.” That keeps me from treating a finding as permanent just because it was hard-won.
It also makes it easier to say something today and update it tomorrow without it feeling like a reversal. You have to give yourself the latitude to adjust your course when the facts change. Research isn’t a verdict handed down once. It’s closer to a running conversation with the people using the product, and conversations change.
What kind of judgment call do people outside research not realize you’re making constantly?
Which voices in a room weigh more heavily? Not every stakeholder opinion carries equal evidence behind it, but every opinion arrives with equal confidence. Part of my work is quietly sorting signals from conviction, without making anyone feel dismissed.
I’ve also had to learn when my own instinct is worth defending and when it’s just a preference dressed up as expertise. That took longer to get right than any research method did.
If you were starting over, what would you do differently?
I’d consider smaller pieces of evidence sooner. I spent years thinking a recommendation needed to be bulletproof before I could stand behind it. Most of the decisions I’m proudest of came from acting on a partial but honest picture, then staying close enough to the work to correct course if it turned out I was wrong.
That’s really the whole discipline: not eliminating uncertainty, but knowing how much of it you can carry on a given decision, and being upfront about the rest.
