Should I test different checkout designs?

Hey everyone, I'm pulling my hair out with our checkout process. We've got decent traffic and people are adding items to their cart, but something's breaking down right before the final purchase. Our conversion rates are hovering around 2-3%, which feels painfully low. I've been looking at our Shopify analytics, and it's like watching potential revenue just slip through our fingers. Every abandoned cart represents not just lost sales, but wasted ad spend. I know checkout design matters—I've read articles about how small friction points can tank conversions. But here's my real challenge: I don't want to mess up our current flow and potentially make things worse. How scientific can I actually be about testing checkout designs? Are there best practices that don't require a complete rebuild? I'm a small team, and we can't afford massive development resources or weeks of engineering time. I need smart, surgical improvements that can move the needle without risking our entire conversion funnel. I'm looking for a systematic approach that gives me real insights, not just guesswork.