Smarter, Faster, Scrappier: How to Run Product Experiments Without Perfectionism

Hey friends,

If Q1 was about unlearning, Q2 is about experimenting- smarter, faster, and with way less perfectionism.

When I first started running experiments, I thought they had to be polished, statistically significant, and rigorously planned.

Now I know: The best experiments aren’t always scientific, they’re scrappy, fast, and focused on learnings and outcomes.

Here are three ideas I’m keeping close this quarter 👇

🔍 Lesson 1: Not everything needs an experiment

Sometimes we default to testing because we’re unsure, or we want buy-in. But testing isn’t always the answer, and it definitely shouldn’t be a stand-in for clarity.

If you can get to an answer faster by interviewing a customer, listening to support calls, or just trusting your gut, you probably should.

📌 Where to Start:
✅ Ask: What decision are we trying to make?
✅ Could a quick conversation or prototype tell us the same thing?
✅ Don’t test just because it’s “best practice.” Test when it helps you move faster with more confidence.

🛠 Try this: Before you run an experiment, write down what you hope to learn and why now is the time to test it. If you can’t answer that in two sentences, you might not need it yet.

⚡️ Lesson 2: Speed > Perfection

The best experiments I’ve run weren’t the prettiest. They were quick, imperfect, and intentionally low-risk.

What made them effective? We learned something real fast—and it shaped our roadmap way more than overplanning ever could.

📌 Where to Start:
✅ Set a time box (1 week or less).
✅ Use tools you already have (think Notion mockups or a Figma prototype).
✅ Get scrappy. The goal is to learn, not launch.

🧠 Mini Prompt: What’s one experiment you could run this week without touching a single line of code?

🎮️ I’ve been playing around with Lovable lately. It’s an AI prototyping tool that turns rough ideas into interactive mockups in minutes. Pretty fun (and surprisingly useful) for early experiments.

🧪 Lesson 3: A failed experiment isn’t a failure

We ran a test a little while back that totally flopped. Low engagement, missed the mark.

But it helped us eliminate an entire path we were considering, and that saved us months of work.

An experiment that proves something won’t work is just as valuable as one that confirms what will.

📌 Where to Start:
✅ Treat each experiment as a decision-making tool, not a proof point.
✅ Debrief as a team: What did we learn? What will we try next?
✅ Share the results (especially the ones that “failed”).

💬 Mini Challenge: What’s one experiment you ran that flopped but helped you make a better decision anyway? Share it with your team this week.

📖 Related read: The Illusion of Validity - Farnam Street
A great reminder of why copying what worked elsewhere can backfire.

💡 Final thoughts + Free resource

If you’re stuck on what to test, I made a quick worksheet that helps you scope low-lift product experiments.
It’s helped me cut through the noise and focus on learning what actually matters. Grab it here.

Hit reply and tell me: What’s one scrappy experiment you’re planning this month?

Until next time,
Stefanie