"We've seen that direct buying clients are generating more cumulative contribution profit in their first 3 and 6 months than clients 1 year ago who were largely Fix clients only. This incrementality gives us more optimism to believe that as our direct buy offering expands, client lifetime values will continue to grow" - Elizabeth Spaulding, CEO of Stitch Fix
In Q3 2021, Stitch Fix (SFIX) announced that all US shoppers can now make direct online purchases of apparel, shoes, and accessories from a personalised set of recommended items. Freestyle, SFIX’s Direct Buying offering, is a significant evolution of the company’s original ‘Fix’ subscription service which could completely change the company's growth trajectory. We interviewed a Former SFIX executive who was heavily involved in rolling out direct buying to understand the Freestyle user experience and why it can’t be replicated by competitors.
A Fix is not your typical apparel shopping experience: customers complete a survey on their body shape and style preferences which is fed through SFIX’s proprietary algorithms to automatically select 5 clothing items for delivery. The company employs ~5,000 part-time stylists to curate clothing products and outfits that, when combined with data science, offers shoppers a personalised experience.
If customers keep all the items in a Fix, they receive a 25% discount on the basket value otherwise they pay the standard upfront $20 styling fee. SFIX also offers a $49 Style Pass subscription for unlimited Fixes per year. The company believes it’s proprietary data set and algorithms enables them to truly personalise apparel shopping at scale.
After skimming the S1, it’s very easy for investors to conclude either churn is too high or SFIX has no advantage because ‘every retailer has data science teams’ or ‘Amazon can easily crush them’.
Historical data shows that ~20% of Fix customers remain after 12 months. This is partly because shoppers are inpatient and the first few Fixes have low keep rates because the algorithm is learning. Churn is also high because it's inherent to apparel as a product category:
"People got to a phase, especially with Covid, where they no longer needed more clothes. There have been many articles written about this for all companies. It is called closet fatigue. It wasn't that they didn't like the service, but the thing that resonated the most for me is that every client goes through phases. They might not need something new right now, and might pause their Fixes. I am not sure if that was the top reason but it makes the most sense to me. It wasn't them saying I don't want more from Stitch Fix, but I don't want anything." - Former Strategy Director at Stitch Fix
Apparel is not great for a subscription service. there is high natural churn because the product isn’t perishable and wardrobes have limited space. It’s no surprise that the 80% one-year churn figure scares most investors away.
However, we believe churn is the incorrect metric to analyse SFIX. We are reminded of Hello Fresh (HFG). Meal-kit churn is similar to SFIX and many investors were short HFG based on unsustainable LTV metrics. However, what we and other investors misunderstood was that meal-kits are not a pure subscription business; it’s a combination of a transaction-based business and a subscription service.
Like SFIX Fix customers, meal-kit users would activate and deactivate the service based on their lifestyle throughout the year; activate when they were busy at work in March, churn when they went on holiday in July, and then reactivate in September when back at work. It’s this mix between ecommerce and subscription which warps the LTV economics and misleads investors. The high APRU and contribution margin plus the natural reactivation embedded in the meal-kit service means HFG unit economics are far stronger than it may seem from the outside looking in.
This could potentially be the case for SFIX as it rolls out direct buying.
The company now allows all existing and new customers to buy items directly on their website similar to a traditional apparel retailer. The only difference is that SFIX personalises the assortment available to each individual shopper based on their shape and style preferences. This has the potential to reactivate many previous Fix customers and combine a natural transaction-based shopping experience with the historical personal shopping Fix service.
This is a significant strategic move and one that is only possible after a decade of collecting data from customers. The Fix subscription is a relatively high-friction service that limits the addressable market to those who hate shopping or have no confidence in picking fashionable items. Customers had to sign up, complete a survey, and then pay an upfront $20 styling fee without seeing the items that were being delivered. The friction of this process is the price SFIX paid to collect the data set that underpins the future of the direct buying offering. The last decade has paved the way for SFIX to have a great shot at building a truly personalized apparel store.
If SFIX is no longer a pure subscription business, we should evaluate traditional retail metrics such as the total value of apparel sold, inventory turns, and unit profitability. Unlike the Fix service, Direct Buying removes the upfront $20 styling fee and reduces the friction for new customers to use SFIX. This not only increases the potential market size but also improves the working capital as customers pay upfront rather than after the Fix products are received. Direct buying also offers a wider assortment and more fluid access to merchandise compared to the 5 Fix items received in a regular cadence. This should not only increase conversion for new customers but also reactivate previous Fix customers.
In online retail, it’s typically believed that the wider the assortment, the higher the sales conversion. SFIX is taking the opposite position relative to competitors; a smaller, personalised assortment that has a higher hit rate to drive conversion and frequency. SFIX has insourced the merchandising problem to their data science team rather than outsourcing it to customers by offering hundreds of thousands of items. Over the last five years, inventory turns have declined from nearly 10x to ~6.8x although this is inventory built ahead of potential supply chain disruption and in preparation of a wider Freestyle assortment. This inventory turnover is still much higher than Nordstrom’s ~4.8x and ASOS or Zalando which is even lower.
The SFIX model could also lead to superior net intake margins due to a higher full-price sell through rate and lower returns. Historically, the Fix offering has been a full price model unless the shopper purchases all 5 items and receives the 25% discount. This was SFIX CEO on why the high full price sell through rate is transferable to direct buying:
"On the full price front, our differentiation is really around helping clients discover things that they otherwise would not have on their own. And so it's fit, it's discovery, it's convenience that has historically characterized the real value proposition of Stitch Fix and that eye towards personalization versus promotions. And so I would expect that set of characteristics to absolutely continue. I would view a lot of this brand expansion as making sure we're matching consumers with what they love and some of those brands may be brands that they've never discovered before. But of course, ensuring we've always put a lot of focus into price comparing and making sure that our price points are competitive in the marketplace" - Elizabeth Spaulding, Baird Conference, June 2021
A higher full-price sell through rate helps SFIX avoid promotional activity, typically the most destructive element of fashion retail. Historical data on fit and size should also reduce the return rate. We estimate ASOS and Zalando have 50-60% return rates. SFIX has all the ingredients to be best in class on both full-price sell through and return rates.
This also highlights why SFIX could be attractive for third-party brands relative to other distribution channels. A higher full-price sell through rate drives higher margins for brands and is a more effective customer acquisition tool relative to pushing discounted products through Zalando or traditional department stores. If SFIX’s data is really that unique, it could be an effective acquisition tool for brands launching new products. SFIX could offer various advertising products or sponsored listings to help brands target customers. The company is also experimenting with capital-light inventory models such as consignment and drop-shipping which could materially improve the ROIC on incremental units. There are many reasons direct buying is a huge opportunity for SFIX; it improves the service not only for customers but also for brand suppliers. However, none of this matters unless you believe SFIX truly has a unique data set and algorithm to match merchandise with shoppers. The personalised assortment needs to convert customers to a sale otherwise SFIX has zero differentiation.
ASOS and Zalando effectively offer an unlimited supply of products across all brands and have an average order value of ~44 EUR with 3-4 purchases per year per shopper. The wide assortment drives the basket size and order profitability which is reinvested to acquire more customers. If the quality of SFIX’s data doesn’t drive conversion, even with high full-price sales, it will be difficult for SFIX to outbid scaled competitors for customers.