I loved running sales. Flash sales, seasonal promotions, email-exclusive discounts—you name it. My customers loved them too. Order notifications would flood in, and I'd watch the dashboard light up with conversions. It felt like I'd discovered a magic button for revenue.
But here's the thing about magic buttons: they're usually too good to be true.
It started innocently enough. A modest 10% off for first-time buyers. Then 15% off for email subscribers. Before long, I was running "limited-time" promotions every other week. Black Friday became Black November. Cyber Monday stretched into Cyber Week.
The dopamine hit was real. Every time I sent out a discount code, orders would spike. Revenue would climb. I'd screenshot the numbers and send them to my co-founder with celebratory emojis. We were crushing it.
Or so I thought.
Six months into this discount strategy, something started nagging at me. I'm a data scientist by training, but I'd been running my e-commerce operation on vibes and vanity metrics. One evening, staring at our monthly P&L, I asked myself a question I should have asked from day one:
Were these discounts actually making us money?
Not "generating revenue"—of course they were doing that. But were they profitable? Were we acquiring customers who'd come back? Or were we just training people to wait for sales?
I didn't have an answer. And that terrified me.
I pulled our Shopify data going back eighteen months. Every order, every discount code, every customer transaction. I started with what seemed like a simple question: what's the actual ROI on these promotions?
The first shock came from our 20% off summer sale—our biggest promotion of the year. Yes, we'd done $47,000 in revenue over that weekend. But when I calculated the actual profit after COGS and the discount itself? We'd made $3,200. Our normal margin was 42%. During that sale, it dropped to 6.8%.
"Okay," I told myself, "but we acquired all those new customers. That's long-term value."
So I looked at repeat purchase rates. Customers acquired during that summer sale had a 12% repeat rate. Customers acquired at full price? 34%. The discount customers who did come back waited an average of 127 days—and 78% of them used another discount code.
We'd trained them to only buy on sale.
Then I discovered something even worse: cannibalization. This is when discounts don't actually create new demand—they just shift purchases that would have happened anyway into the discount window.
I ran a comparison of our "email subscriber exclusive" 15% discount. When we sent it out, orders would spike for 48 hours, then crater for the next week. The total weekly order volume stayed roughly the same. We weren't growing the pie; we were just moving it around and giving away 15% margin in the process.
I looked at customer purchase histories. Sarah, who bought from us every 6-8 weeks like clockwork, started timing her purchases to our promotions. Jason, a whale customer who used to drop $300+ orders monthly, now waited for sales. We were literally paying our best customers to buy less often.
The worst part? Our average order value was dropping. Without discounts, our AOV was $127. With discounts, it dropped to $89. Customers weren't just paying less per item—they were buying fewer items. The psychology makes sense: if you're getting a deal, you feel less compelled to hit free shipping thresholds or add extras.
Here's where it got interesting. Not all discounts were disasters. When I segmented the data, clear patterns emerged.
Our free shipping threshold promotion (spend $75, get free shipping) had an 89% positive ROI. It increased AOV by $23 on average, and the shipping cost was less than the margin gain. Customers acquired this way had a 31% repeat rate—nearly identical to full-price customers.
Our product bundle discount (buy 3, save 10%) was also profitable. AOV jumped to $142, and we moved inventory that otherwise sat on shelves. The margin hit was real, but it was offset by volume and reduced holding costs.
The pattern became clear: discounts that encouraged customers to buy more worked. Discounts that just made things cheaper didn't.
I developed a simple framework for evaluating every promotion we ran:
When I applied this framework to our promotion history, the results were brutal. Only 3 out of our 14 regular promotions were actually profitable when you accounted for everything. We'd been hemorrhaging margin for months, mistaking top-line revenue growth for business health.
I didn't stop running promotions entirely. But I got strategic about it.
Out: Blanket percentage discounts. Site-wide sales. Email blast codes. Predictable monthly promotions.
In: Tiered free shipping thresholds. Strategic product bundles. Time-limited flash sales (actual 24-hour windows, not "extended by popular demand"). First-purchase discounts with second-purchase full-price incentives.
The scariest change? I stopped our monthly "subscriber appreciation" 15% discount. I was convinced our email list would revolt. Our open rates would tank. Revenue would collapse.
What actually happened? Open rates went up 8%. Revenue dipped for two weeks, then stabilized right where it had been—but now at full margin. Turns out, customers who actually wanted our products would buy them regardless. The discount hadn't been driving demand; it had just been eroding profit.
After implementing these changes, here's what happened to our business:
We were making more money on fewer orders. Our customers were higher quality. Our business was healthier. And I wasn't spending every Sunday night planning the next week's panic-discount to hit revenue targets.
Here's what I learned: discounts are a drug. They feel amazing in the moment—orders pouring in, notifications dinging, revenue climbing. But like any drug, they create dependency. Your customers get hooked on deals. You get hooked on the short-term revenue spike. And before you know it, you can't hit your numbers without them.
The hardest part wasn't the analysis. It was admitting to myself that I'd been lying about our business health. Those screenshots of revenue spikes I'd been sending? They were vanity. The real story was in the profit margins, the customer cohorts, the repeat rates. And that story was uncomfortable.
But uncomfortable truths are better than comfortable lies.
Look, I'm not here to tell you never to run promotions. E-commerce is competitive, and discounts are part of the game. But you need to know if they're helping or hurting.
Ask yourself:
If you don't have clear answers to these questions, you're flying blind. And flying blind is how you crash.
After going through this journey, I built some analysis modules to make this easier for other e-commerce operators. Because honestly? Doing this analysis manually in spreadsheets was excruciating.
The Shopify discount analysis module breaks down discount ROI, cannibalization rates, and customer quality by promotion type. It's what I wish I'd had when I started asking these questions.
For Etsy sellers (I ran a side shop for a while), the discount effectiveness module does similar analysis for that platform's specific promotion types.
And the AOV analysis module tracks how different promotions impact your average order value—which turned out to be one of the most important metrics for understanding true promotion profitability.
These aren't magic solutions. They're just tools that make the uncomfortable questions easier to answer. And sometimes, asking the uncomfortable questions is exactly what your business needs.
I still run promotions. But now I run them like a data scientist, not a desperate founder chasing vanity metrics. Some work. Some don't. The difference is, now I know which is which. And that knowledge—that uncomfortable, clarifying knowledge—has been worth more than any Black Friday spike could ever be.
If you're curious about what your own discount data might reveal, take a look at your numbers. Really look at them. Not just revenue—margin, repeat rates, AOV, cannibalization. You might not like what you find. But I promise you, knowing is better than guessing.