Hey everyone, I'm Carlos. Spent the last five years at Speero scaling experimentation programs for ecommerce brands. There I ran 500+ tests for 8-figure stores, ranging from content changes to pricing and replatforming initiatives. Got deep enough into it that I ended up putting together a CXL course on ecommerce experimentation. This month I joined Intelligems to make sure practitioners like you have a real seat at the table here. G.E.M. (The Growth Experiment Mindset) is the closest thing I've seen to how good experimentation actually works in practice. "How to think about testing, not just what buttons to click" is exactly the shift I've been trying to get brands to make for years and this community is the best way to get that message to more people. Looking forward to hearing how you're driving ecomm growth through experimentation!

Most store owners pick $50 or $75 because it sounds right. But your customers are already showing you the real number. Here's how to let the data decide: 1. Graph your order history Plot your orders by value over the last 90 days. You'll see natural peaks where customers tend to land. These clusters reveal spending comfort zones you didn't set... they emerged on their own. 2. Find your abandonment cliffs Pull cart abandonment rates by value. Look for sharp drops at specific price points. That's where your current threshold creates friction that kills conversions. 3. Watch what gets added last Check which products appear in carts just above your threshold. Stickers, samples, cheapest items in your store? That's customers telling you the threshold feels arbitrary. They're buying stuff they don't want just to hit the number. 4. Test at natural pile-up points Run threshold tests where orders already cluster. But don't measure AOV alone. Profit per visitor tells the real story. A lower threshold with higher conversion often wins. Your data already has the answer. You just have to ask the right questions.

Here's one quick way to start exploring your data. This Is NOT About Generating Test Ideas That's lazy. You've heard advice like "test your shipping threshold" or "try bundling." But without knowing YOUR data, you're guessing. Generic advice doesn't help because every brand is different. This is about exploring and understanding. Understanding comes first. Testing validates what you see in your data signals. I highly encourage you to go to our explore model and dig deeper into this. Personal opinion, those who learn how to master the exploration data phase are the ones who really unlock growth. https://www.skool.com/intelligems-academy-1535/classroom/90ab5960?md=605f4d08374d41e4a58b069e22ca4757

Every time someone tells me their A/B test killed their CPCs, I ask the same question: did you check what your CPCs were doing before the test? Had a merchant convinced their test caused a spike. When we actually pulled the data, their account was doing the exact same thing the week before. No test running. Same pattern. The "spike" was just their normal account behavior they'd never noticed. CPCs are noisy. They move around constantly. Most of the time when you launch a test and see costs go up, it's correlation not causation. Before you kill a test, pull your hourly data from the same time period last week. Compare like to like. You might be surprised what you find. Anyone else blamed a test for something that turned out to be normal account behavior?

Every time I test discounts I'm reminded of how hard it actually is to make discounts make sense and overcome the loss in margin. We're testing not having an offer, vs 15% off here. It increased CVR a bunch, but barely any profit gained. We'll keep testing it for longer, but just a reminder for us to be stingy when it comes to giving away margin