Disclaimer: This content is for educational and entertainment purposes only and should not be construed as investment advice or a recommendation to buy or sell any security. I own shares of Stitch Fix (NasdaqGS:SFIX).
Stitch Fix is an innovative online apparel retailer. By combining data science and human judgement, they have reinvented the shopping experience and democratized personal styling. Unlike other ecommerce sites that put the burden of search on consumers, Stitch Fix is a recommendation engine focused on getting relevant merchandise in front of clients. The value proposition is rooted in deep personalization, curation, and discovery.
To start using Stitch Fix, new customers fill out a detailed profile sharing measurements and preferences on style, fit, and price points. A stylist then selects individualized merchandise to send customers (typically five items) based on recommendations from proprietary algorithms. You try everything on in the comfort of your home, keep what you want, and return everything else in a prepaid shipping bag. A $20 styling fee is paid in advance of the shipment but is credited against any merchandise purchased. Shipping is free both ways and a 25% discount is offered if all items are purchased. Stitch Fix keeps track of the items kept, returned, and the feedback provided about each item to improve over time. Deliveries can be scheduled on an automated cadence (e.g. monthly or quarterly) or on an as-needed basis.
Although it won’t appeal to everybody, this model resonates with a subset of consumers. Some are fashion conscious but hate to shop. Others are hard to fit (e.g. shorter than average, long arms, short legs/long torso, etc). And some just want to discover new brands and styles. Scott Galloway sums up the value proposition, “We have mistakenly assumed that consumers want choice. No they don’t. They want less choice. They just want to be more confident in the choices they make.”
Current CEO Katrina Lake founded Stitch Fix in 2011. She began testing the idea while still enrolled at Harvard Business School — style profiles were filled out on Google Docs, she kept track of return policies at local boutiques, financed inventory purchases on a $6K limit credit card, and shipped boxes out of her apartment. Lake received seed funding and quickly moved the company to San Francisco to tap into a broader pool of data science and engineering talent.
Stitch Fix had instant product market fit. Growth was completely organic for the first four years as the company scaled to ~1m clients and ~$340 million in revenue with zero marketing spend (See Exhibit A).
Despite the traction, Stitch Fix only raised $42.5 million as a private company. The constraint was not by design as Lake self-describes herself as a terrible fundraiser. However, it forced the company to commit to operational excellence and financial discipline. Purchasing and inventory management became core competencies, they focused on unit economics, and utilized a waitlist to grow in a controlled manner. Unlike many other startups in its cohort, Stitch Fix has been profitable since late 2014.
Additionally, Lake has proven an ability to attract and retain talent. Mike Smith joined in June 2012 and is currently President and COO overseeing operations, client experience, and merchandising. He also launched Stitch Fix Men in 2016. Previously, Mike was COO at Walmart.com. Lake also convinced Eric Colson to join the team in August 2012 as Chief Algorithms Officer overseeing the team of data scientists. Eric was previously VP of Data Science and Engineering at Netflix.
There have been numerous ecommerce 2.0 flameouts over the past decade (e.g. Gilt Groupe, Fab.com, Birchbox, Shoedazzle, Beachmint, One Kings Lane). Venture capitalists flocked to these businesses due to large addressable markets and strong top-line growth. To be fair, there have been some big winners (e.g. Wayfair) which can justify the VC game. But as Bill Gurley points out, innovations around pricing or distribution — think flash sales and subscriptions in a box — don’t represent core differentiation or sustainable competitive advantages. Additionally, these startups had access to hundreds of millions of VC funding and therefore weren’t forced to prove out the unit economics before scaling rapidly.
Stitch Fix is unique. The business is capital efficient and data science capabilities create a durable competitive moat. The company employs over 100 data scientists to develop proprietary algorithms to improve the customer experience and increase operational efficiencies. Data science is deeply embedded in the company culture and is used across business units. Here are a few examples:
- Styling and item selection: algorithms optimize the client-stylist pairing and recommend relevant inventory to stylists
- Inventory planning: algorithms help to forecast demand and optimize merchandise across categories and styles to increase inventory turns and reduce excess supply
- Logistics: algorithms assign shipment requests to one of five warehouses based on proximity to the client and relevant inventory; they also determine the most efficient warehouse route for employees to fill boxes
- New style development: algorithms determine underserved and unmet client preferences to help develop new styles in the Exclusive Brand portfolio
- Additional detail can be found here
One of my favorite investing analysts, David Kim, highlights three important questions when evaluating the competitive advantages of a data-based business (subscription required). Let’s apply his checklist to Stitch Fix.
1. Is the data proprietary? ✓
Stitch Fix has a very unique relationship with users. The vast majority of client data is provided directly and explicitly by the clients themselves. On average, each client provides over 90 meaningful data points through his or her style profile (e.g. style, size, fit, and price preferences). Intimate details are often shared, such as early pregnancy details or body parts they like to flaunt. Clients trust that providing detailed and accurate data will make their experience better. The one to one relationship with a stylist also improves data collection as clients instinctively feel more comfortable sharing detailed and personal information.
The company also collects extensive merchandise data. Stitch Fix records 30–50 attributes for each item, including less traditional measurements such as button height and sleeve opening. This is obviously not proprietary but it’s built into their processes and is extremely valuable when paired with client data.
Stitch Fix has also been experimenting with product features to collect more data. Style Shuffle, launched in early 2018, is a Tinder-like game serving up merchandise with the prompt, “Is this your style?” Users have only two options to rate each item: thumbs up or thumbs down. The innovative, light-weight feature has been very successful. More than 75% of customers have tried Style Shuffle, providing Stitch Fix with more than a billion ratings. It’s an extremely effective way to learn more about style preferences.
2. Are the insights from the data critical? ✓
Fashion shoppers are often looking for curation and discovery. Window shopping has yet to be solved on the web. Instead, we have a search bar and an endless sea of merchandise. Sure, influencers may be the new curators. Perhaps you decide to buy a dress on Revolve because you saw your favorite Bachelorette cast member sponsor a post on Instagram. But it doesn’t solve the fit issue. Even with free shipping both ways, the experience can be a hassle (shipping time, repackaging, returning) and inferior to a dressing room at a boutique.
Stitch Fix has tremendous potential if they can use data to nail both the style and fit to help clients discover new products.
3. Does the data fuel a feedback loop that deepens the data moat? ✓
Perhaps the most valuable data set Stitch Fix receives is feedback on the items sent to customers. 85% of shipments result in direct client feedback allowing Stitch Fix to become more personalized over time.
This virtuous flywheel is substantiated by the results. Keep rates (average number of items purchased per Fix) were up 22% in 2017 vs. 2014. Keith Rabois describes this data network effect as an accumulating advantage. He goes on to pinpoint trust as another powerful accumulating advantage, “When you start a company…people may discount your ability to perform the desired task accurately and successfully… Presumably it gets incrementally easier to convince people that what you’re doing actually works.” To be clear, Rabois is not talking about Stitch Fix specifically, but I believe both frameworks apply.
Like all commerce companies, Stitch Fix has two primary growth variables: increase customers and increase average spend per customer (ARPU). The company has over 3.1 million active clients, meaning they checked out a Fix in the past 12 months. In addition to increased penetration in their core market, user growth has come from new categories (petite, plus, maternity, men) and new geographies (United Kingdom launched in May 2019). Despite this market expansion, the year-over-year active client growth has decelerated from ~30% last year to ~17% in the most recent quarter (See Exhibit A). Stitch Fix found instant product market fit with a subset of clients, but it seems to be difficult to rapidly expand beyond this core group. This is a new way to shop and changing user behavior is hard.
However, I believe the company has a big opportunity to increase wallet share and ARPU. Here are some potential drivers:
- More brands and merchandise: Stitch Fix carries merchandise from over 1,000 brands, including established and emerging names. This is up from 700 brands at the time of the November 2017 IPO. Brands view Stitch Fix favorably — it’s one of the only full-price growth channels, they’re accessing a highly relevant pool of potential new customers, and they receive tangible feedback to improve product designs. As Stitch Fix continues to onboard new brand partners, the variety of merchandise will continue to increase across categories (e.g. lifestyle, workwear, activewear, special occassion, etc).
- Exclusive brands: Stitch Fix has been developing a private label portfolio to address underserved and unmet client preferences. In addition to higher margins, these brands can lead to increased spend and client retention since they’re filling a gap in the market and only available on Stitch Fix. Exclusive brands have been particularly successful with men, comprising more than one-third of total men’s revenue. Historically, private label has been less than 30% of the overall business mix.
- Extras: In February 2018, Stitch Fix launched a feature allowing clients to select certain items to add to their next Fix. The curated assortment primarily consists of basics (e.g. socks, bras, underwear, undershirts).
- Repeat purchases: Stitch Fix recently began allowing customers to browse and buy previously-purchased styles in additional colors, prints, or sizes on an a la carte basis.
In addition to increasing wallet share, the Extras launch was particularly interesting. It forced Stitch Fix to build new fulfillment capabilities to ship boxes with a varying number of items.
Even before the IPO, Katrina Lake explained that the company is rooted in personalization — getting relevant merchandise in front of clients. And they’re willing to evolve the model beyond five items in a box to realize that bold vision.
the personalization capabilities are so good now that what we’re really thinking about is how do we bring those to life in more ways, and how do we make it so that you can engage with Stitch Fix in ways that isn’t just in that box and five things? How can we take these amazing capabilities that we have and help it to expand our wallet share with you and our addressable market? I think you’ll see more to come on that front… it doesn’t have to be five things, it doesn’t have be in the box, and to your point, it doesn’t even have to be apparel. I think that’s the really special part about this company…we’ll be thinking about how do we use that capability to grow over the next 10–20 years.
Evolving the model is the 10x+ opportunity that can dramatically increase the client base and average spend per user.
The Amazon Effect
To recap, Stitch Fix has an impressive leadership team, proven unit economics, a long growth runway, and data network effects. So why does the stock trade like an offline branded apparel retailer (See Exhibit B)?
In late July, Amazon announced a styling service for Prime subscribers called Personal Shopper by Prime Wardrobe (PSPW). The announcement didn’t come as a surprise given prior rumors and the fact that Jeff Bezos once told Jason Del Ray he believes the future of fashion will be mixing technology with personal touch. The service is largely inspired by Stitch Fix — a customer fills out a profile and stylists, aided by technology, curate a selection of personalized merchandise. Here are the main differences:
- Subscription: PSPW is a monthly subscription service. Stitch Fix deliveries can be scheduled on a regular cadence or on an as-needed basis.
- Styling fee: PSPW charges $4.99 per month on top of an Amazon Prime membership. Stitch Fix charges a $20 styling fee that is credited towards the cost of any items kept.
- Preview: PSPW allows users to preview stylist picks and select which items they want shipped. Stitch Fix does not allow clients to preview merchandise as they trust stylists to occasionally introduce users to items that may be out of their comfort zone.
- Categories: PSPW currently only caters to women. Stitch Fix is available in women, petite, maternity, men, plus, and kids.
Amazon is terrifying with unprecedented market dominance. They have over 100 million Prime subscribers and capture nearly half of all US ecommerce sales in the US. Additionally, Amazon is a fierce competitor (remember Quidsi) due to massive scale, a very low cost of capital, and a multi-decade time horizon. They have also successfully entered entirely new markets (e.g. Amazon Web Services and Echo/Alexa). Based on this track record, the market punishes competitors seemingly in their path. Grocers and health insurers are recent examples.
It shouldn’t be a surprise that Stitch Fix is down over 30% since the Amazon announcement. Tariff fears may have also exacerbated the selling pressure. Although it’s typically a money losing proposition to bet against Amazon, I believe Stitch Fix can win this battle.
Let’s review Amazon’s formula to dominate retail: (i) vast selection due to unlimited shelf space, (ii) maximum convenience — fast delivery, one-click checkout, detailed product information and reviews, (iii) low prices, and (iv) great customer service including hassle free returns.
But as Zack Kanter points out in his brilliant essay on Amazon, “From infinite shelf space comes a problem: how do customers discover new products?” Kanter explains that Amazon relies heavily on product reviews and sales velocity to rank search results. This model works great for standardized products (e.g. toilet paper, batteries) or when the customer knows exactly what they want. Convenience and price drive these purchase decisions. But apparel is nuanced and deeply personal. Fashion lovers are seeking curation and discovery. The perfect fit and style are often prioritized over delivery speed or price.
Emily Weiss (CEO of Glossier) nails it, “Amazon really solved buying but it killed shopping in the process.”
Amazon does a really good job on mainstream products that are all about convenience, and when you really know what you want — toiletries, things for your kitchen. But when it comes to something that’s going to define who you are, or it starts touching on your identity and the way you’re represented, it’s not really a core competency of theirs. That’s really where we spend our time, thinking about: OK, if Amazon really wins in the mainstream, what are the other areas that are more relevant to the way a person defines themselves?
So then how does amazon capture nearly 30% of US online apparel sales? Well, apparel is a commodity for some shoppers. For example, my mom buys her jeans at Coscto. You should see the satisfaction on her face when she describes the bargain price paid — value investing runs in the family. Amazon’s formula is perfect for these consumers.
These commodity-type purchases are a big reason why Amazon’s apparel strategy is heavily focused on private label brands and basics. Here is some color from the Q4 2017 earnings call, “Amazon Fashion is one area where you’re seeing us offer a number of private apparel brands. Some of the more sort of popular lines with customers have been things like Goodthreads, Amazon Essentials, which is men’s and women’s basics.”
Amazon has launched over 100 private label brands in the US for apparel and accessories. However, over 85% of these products have price points below $50. Playing to their strengths, it seems they’re attracting apparel shoppers optimizing for cheapest and fastest.
Other consumers are more fashion conscious. I believe Stitch Fix can beat Amazon in this niche. Stitch Fix has built relationships rooted in trust and they’re laser focused on personalization, curation, and discovery.
I’ll once again lean on Keith Rabois to alleviate the threat:
You have to struggle to find examples over the last 30 years where a large competitor has beat a high growth startup. It’s worth being conscious when an incumbent has unusual leverage and understanding why your counter-leverage may offset that or not — being paranoid is smart. But more typically, a focused and talented team with an ownership mentality and incentive alignment will out execute a very large entity that on paper looks very threatening.
As shown in Exhibit B, Stitch Fix currently trades at ~0.8x CY 2020 revenue. Assuming 10% steady state operating margins (low-end of their target range) and a 25% tax rate, this implies a pro forma, unlevered P/E of ~11x.
Skeptics may question the steady state margin assumptions given declines since 2015 (See Exhibit A). Advertising costs have increased as it has been more difficult to expand the client base outside of the core audience. But I also believe growth investments are hitting the income statement and masking the long-term earnings power. Here are a few examples: 18 exclusive brands have been developed from scratch, men’s launched in September 2016, kids launched in July 2018, and the UK launched in May 2019.
I think 1.8x forward revenue (pro forma, unlevered P/E of ~24x) is more appropriate given the growth runway, long-term margin profile, and competitive moat. It’s in line with relevant comps (Zalando, boohoo, and Revolve) and precedent transactions (See Exhibits B and C). This translates to a $38 price target, or close to 100% upside from current levels.
If the company continues to evolve the model and executes on their grand vision to personalize retail, the stock has 10x+ upside.
Exhibit A: Summary Financials
Exhibit B: Public Comparables
Exhibit C: Precedent Transactions