Latest / Startup Stories with Fexingo: Conversations About Founders, Funding, and Building Companies from Zero / How Shein Built Fast Fashion on Extreme Speed
Transcript
- Lucas: There's a number in fashion that keeps me up at night, and it's not a margin or a valuation. It's 3,000. Luna: Three thousand what? Lucas: That's roughly how many new items Shein adds to its app every single day. Zara, for comparison, launches about 1,500 new styles a year. Shein does more than that in half a day. Luna: That's a different league. How do they physically produce that many designs without drowning in inventory? Lucas: That's exactly the puzzle. The answer is a system they call 'test and respond.' Instead of predicting what will sell and making huge upfront orders — which is how Zara and H&M work — Shein starts with tiny batches, sometimes just 50 to 100 units per style. They put it on the app, measure click-through and purchase data in real time, and only reorder the ones that hit a demand threshold. Luna: So they're using customer behavior as their forecast. That's smart, but it requires a manufacturing network that can turn on a dime. Lucas: Right. And that's the second piece of the puzzle. Shein works with over 3,000 supplier factories concentrated in the Guangzhou area of China. They've built a proprietary software platform that connects these factories directly to their demand data. A factory might get an order for 200 units of a dress at 10 AM, cut the fabric by noon, and have it ready for shipment by 6 PM. Luna: That turnaround is insane. Traditional fast fashion lead times are measured in weeks, not hours. Lucas: Exactly. Zara's famous for being fast at two to three weeks from design to shelf. Shein has compressed that to about three to five days for many items. They achieve that partly by accepting higher production costs per unit — they don't get the same economies of scale as a bulk order — but they offset it with drastically lower inventory risk. They rarely end up with unsold stock. Luna: So their margin profile is different. They make less per item, but they waste less. Lucas: That's right. According to a 2024 case study in the Journal of Operations Management, Shein's inventory turnover is something like 15 to 20 times a year. Zara is around 6. H&M is closer to 3. That means Shein's capital isn't tied up in warehouses full of last season's trends. Luna: I've read that they also use algorithms to design the clothes. Is that true? Lucas: Partly. They use trend-spotting software that scans social media, runway shows, and even street style photos to identify emerging patterns, colors, and silhouettes. Their design team then creates variations based on those signals. But the real magic is in the data feedback loop: a design that gets high engagement but low conversion might get tweaked and re-tested, not killed outright. Luna: That level of granularity reminds me of how Netflix uses watch data to decide what shows to greenlight. But fashion is more fragmented — you're dealing with thousands of micro-trends. Lucas: Exactly. And that's where the scale becomes a competitive moat. The more data you collect, the better your algorithms become at predicting what will sell. And the better your predictions, the less risk you take on new batches. It's a virtuous cycle that's really hard for competitors to replicate, especially if they don't have a similar manufacturing ecosystem. Luna: But there's a dark side here, right? The environmental cost of producing so many items, the labor conditions in those factories, and the culture of disposable fashion. Lucas: Absolutely. Shein has faced repeated criticism for labor practices and environmental impact. A 2023 report from the Swiss nonprofit Public Eye found that some supplier factories had excessive overtime and wages below living standards. Shein has since launched a sustainability program and increased audits, but the model itself incentivizes overproduction. Even if they sell most of what they make, they're still churning out billions of garments a year. Luna: Is that sustainable — I mean, for the business itself? Could regulation or consumer backlash eventually cap their growth? Lucas: It's a real risk. Europe is considering tighter rules on textile waste, and the US has proposed bills around supply chain transparency. But so far, Shein's growth hasn't slowed. Their valuation hit $100 billion in 2024, and they were reportedly planning an IPO in London this year. Investors seem to believe the operational model wins in the long run, even with regulatory headwinds. Luna: That valuation implies they're already bigger than H&M and closing in on Zara in terms of market cap. Do you think the model is replicable for other brands? Lucas: Partially. You need three things: a dense manufacturing cluster, a software platform that integrates with suppliers, and a culture of data-driven decision-making. China's Pearl River Delta offers a unique combination of fabric mills, garment factories, and logistics that's hard to match. Some companies in Southeast Asia are trying, but none have achieved Shein's speed. Luna: And on the software side, Shein built its own platform from scratch. That's a huge upfront investment. Lucas: Right. Off-the-shelf ERP systems aren't designed for this kind of real-time coordination. Shein developed what they call the 'Shein Supply Chain Management System' that lets suppliers see demand forecasts, order statuses, and quality metrics on a single dashboard. That kind of proprietary infrastructure is a barrier to entry. Luna: I want to go back to the numbers for a second. You said 3,000 items a day. How many of those actually survive to become repeat orders? Lucas: Estimates suggest around 40 to 50 percent. That might sound low, but it's actually high for a test and respond model. Most traditional brands would kill for a 50 percent sell-through rate on new styles. Shein's advantage is that the failures cost very little — they only made 50 units, and those sell at a discount or get bundled. Luna: So they're hedging risk with extreme granularity. Every style is a tiny experiment. Lucas: Exactly. And the experiments feed into each other. A color that works on a dress might be applied to a top next week. A silhouette that flops gets retired quickly. The whole system is built for iteration, not perfection. Luna: It's almost like a software development approach — agile, test-driven, with rapid deployment. Just applied to dresses. Lucas: Precisely. And that's why Shein calls itself a technology company that happens to sell clothes. Their R&D spending is heavily weighted toward software engineers, not fashion designers. They have more data scientists than they do pattern makers. Luna: One thing I'm curious about: how do they handle returns? With that volume, logistics must be a nightmare. Lucas: They handle returns by discouraging them. Shein's return policy is less generous than Zara's or H&M's — often customers get store credit rather than cash refunds, and returns aren't free in many markets. That reduces reverse logistics costs. But it also puts more pressure on getting the product right the first time, which the data helps with. Luna: So the entire model is engineered to minimize the two biggest cost sinks in fashion: inventory waste and returns. Lucas: That's the thesis. And it's worked so far. But the question is whether it can scale without breaking. As Shein gets bigger, regulatory scrutiny increases, and consumer expectations around sustainability shift. They'll have to adapt. Luna: I've been thinking about something you mentioned earlier — the ecosystem dependency. If Shein really does rely on that Guangzhou cluster, could they survive a geopolitical disruption? Say, tariffs or a supply chain decoupling? Lucas: That's a huge risk. The US has already imposed de minimis rules changes that affect Shein's duty-free shipping for low-value packages. And there's talk of broader tariffs on Chinese goods. Shein has been diversifying — they opened a supplier network in Brazil and are building a fulfillment center in the US — but the vast majority of production still comes from China. A serious trade war would compress their margins significantly. Luna: It's fascinating how a business so reliant on speed and data is also vulnerable to forces it can't control. Reminds me of how dependent just-in-time manufacturing is on stable geopolitics. Lucas: Absolutely. The same agility that makes Shein dominant could become a liability if the sourcing base fractures. But for now, investors are betting on adaptability. Luna: Before we wrap up, I want to pause on something. You mentioned that Shein's model is built on constant iteration and tiny experiments. In a way, that's also how we think about building this show — trying new formats, listening to what resonates, and letting the data guide us. Lucas: That's a fair comparison. And it ties into something I want to share. This show exists because a small group of listeners chips in to keep it ad-free. If you've gotten value from these conversations — maybe a new perspective on a business model, or a concrete idea you've used at work — you can support us at buy me a coffee dot com slash fexingo. Luna: It's a low-key thing. No paywalls, no perks, just knowing that what we do here helps people think differently. And it's what keeps the show independent. Lucas: Exactly. So if that resonates, great. If not, no pressure. Now, let's get back to Shein. One more angle I think is worth exploring: how their pricing strategy differs from competitors. Luna: Yeah, because they're famously cheap. But is it just low margins, or is there something structural? Lucas: It's both. On the structural side, because they avoid inventory risk, they don't have to mark down products to clear shelves. That means they can offer lower everyday prices without needing to have big sales. Their gross margin is actually decent — reportedly around 60 percent — which is competitive with Zara. The trade-off is higher operating expenses due to faster shipping and more frequent production runs. Luna: So they make up for lower per-unit profit with higher velocity and lower waste. That's a capital-light model in a capital-intensive industry. Lucas: Right. And because they own the customer relationship through their app, they can use data to cross-sell and increase lifetime value. A customer who bought a dress might get recommended shoes and accessories that match, driving up average order size. Luna: That's a classic e-commerce play, but it's especially powerful in fashion where there's a lot of impulse buying. Lucas: Exactly. And Shein is very aggressive with push notifications and personalized marketing. Their app engagement metrics are off the charts — users spend an average of 30 to 40 minutes per session, which is closer to social media than to shopping. Luna: So they've effectively turned shopping into entertainment. That's a different competitive advantage than just speed. Lucas: It is. And it's why they've managed to build a loyal Gen Z customer base that doesn't see Shein as a cheap alternative, but as a discovery platform. The line between browsing and buying is intentionally blurred. Luna: I think that's the most important lesson from Shein: they didn't just optimize an existing model; they reimagined what a fashion retailer could be — a data-driven, real-time, content-first platform. Whether you love or hate what they represent, that's genuinely innovative. Lucas: Agreed. The next decade will test whether that innovation can survive its own success. But for now, it's one of the most instructive business stories in retail.