What sample size do I need for statistically significant results?

As a Shopify store owner who's been obsessively tracking every metric, I've learned the hard way that not all A/B tests are created equal. Last quarter, I ran what I thought was a killer product page test - new images, tweaked copy, different call-to-action. I was thrilled when it looked like my conversion rate jumped 15%. But when I dug deeper with my developer, we realized the sample size was laughably small. Those 'results' were basically statistical noise. We'd only tracked about 50 visitors per variant, which meant our confidence interval was massive and our conclusions were essentially meaningless. I was basically making crucial business decisions based on a coin flip. It was a wake-up call that in e-commerce optimization, you can't just eyeball results or trust your gut. You need a rigorous, mathematical approach to determining how many visitors or transactions you truly need to make a statistically valid conclusion. My ad spend is too precious, and my margins are too tight to waste time on pseudo-scientific guesswork. I needed a systematic way to understand exactly how many data points would give me real, actionable insights.