When Stitch Fix went public in November of 2017 it was largely viewed as a broken IPO. Priced at $15 – well below the range of $18-$20. What people did not realize at the time is that the model as it existed in 2016/17 was the best that SFIX had to offer. The company captured the sweet spot of its TAM at ~1.5mm customers, who were spending $550-$600 with SFIX per year. The core customer here is someone who needs style advice, and is comfortable paying for a shipment (or a ‘Fix’) of apparel sight unseen, but curated for them specifically by extensive algorithms combined with (live) stylists. Keep what you like, send the rest back. If you send it all back you pay $20. Gross margins were 45-46%, and importantly its customer acquisition cost (CAC) was at trough levels of ~3% of revenue. The company turned its inventory at 18x, which is like lightning in a bottle for a retailer (apparel asset turns generally range between 3-4x), and it generated 60% return on capital and 80% ROE. The company had just earned $1.47 per share in 2016, though the fact that CAC was rising in 2017 and therefore pressuring margins scared people in what was a very different market than what we’re living in today. Goldman got paid on the deal, but better advice would to have been to not go public at all. We think it’s headed back to its IPO price, and then lower.
The problem here is that with the IPO came extremely high growth expectations, and in chasing them the company ran out of TAM way too quickly. The rapid-fire launch schedule of new initiatives has brought it too far beyond its core TAM, and way out on the risk curve into a territory that is simply destroying the (once attractive) financial characteristics of the business. It started selling to Women with the launch of the site in Nov 2011 after its first seed round in February of that year. By 2015 it had 867k customers at an operating margin of 10.6%. That was solid. The company should have frozen growth at that point. But it didn’t. In its peak earnings year of 2016, it launched Men’s, after which the company saw a 12% decline in spending per customer (while the company was in the IPO process). Men spend less than women on premium apparel, and they cost more to acquire. Period. Then, less than a year later, it launched Kids – which in and of itself is a big step down from Men as it relates to spending per capita (the company included 8-12 items per Kids Fix as opposed to 5 items for Men and Women to make up for the lower price point). By this time operating margins had fallen to 3.5%. Then, just two months after launching Kids, SFIX announced that it was launching into the UK in 2019 – the complexity of which cost SFIX another 200bps in margin. So over a 5-year time frame of spending more in CAC for a lower-yielding consumer, EBIT declined from $64mm to $23mm, despite adding $1bn in revenue and 1.5mm net new customers.
By our estimates, this brought the business model to a point where 50% of customers don’t make it past 2-3 months and up to 75% won’t make it through the year. If the company pulls back on sales and marketing spend then the net consumer additions will go solidly negative given the high churn characteristics of this business model. The top line would follow this trend and this company would lose its 4x Sales multiple in a heartbeat. This company NEEDS to spend in order to give the market what it’s paying for – top line growth.
Perhaps the biggest admission that SFIX had run out of TAM was in December of 2019 when it introduced ‘Shop Your Looks’, otherwise known as ‘Direct Buy.’ This offering allows shoppers to go online and pick their own assortment instead of a computer algorithm and a stylist doing it for them. So let me get this straight, a company that was built on custom personalized curation of apparel and charging a premium to the consumer that needs the styling advice decides to abandon that model and put the product selection decision in the hands of the consumers. How is this different from going to www.nordstrom.com? It’s not. Actually, Nordstrom has 10x the assortment that SFIX has – with the inventory and the discounting mechanism to compete aggressively for the business. SFIX tested Direct Buy for four months, and then rolled it out nationally in March – spot on with the start of the pandemic. Since the start of Direct Buy, the market cap of this company has gone from $1.6bn (March trough) to $10bn. It’s been a six-bagger – at the same time people have walked away from wearing the exact type of dressy/occasion apparel that SFIX is known for. By our estimate, Direct Buy is now up to 20% of sales for SFIX – an astonishing shift in just 12 pandemic-impacted months. To say that the market is excited about this initiative is a severe understatement.
Our sense is that Direct Buy will be 50%-60% of the business within 2-years, and we don’t think people quite understand the dramatic impact it will have on the company’s financials. Consider the following…
- Revenue: Customer count should accelerate as the company continues to step on the gas with its customer acquisition costs, but the Direct Buy customer and the lower-spending incremental Fix customer both should drive down the spend/customer. The net result is that what was a 25-30% top line growth rate pre-pandemic should be cut in half within 2-years, and end up at a Kohl’s-like low-single digit rate by year 5 of our model.
- Gross Margin: This will be the a major shock to the model. With Direct Buy the company opens itself up to severe competition, something that its ‘personalized curated’ AI model has been isolated from in the past. We think the company competes away 500-600bp of Gross Margin over a TAIL duration.
- CAC: Remains elevated at 9-10% to drive incremental customers – keeping in mind that the company’s long-term guidance calls for a 9-11% sales & marketing ratio.
- Other SG&A: We should see some leverage here, as the company will not have to invest in incremental data-science resources as it puts the curating decisions in the consumer’s hands. We also think that the company will see the degradation in the business model and will cut costs where it can (it cut 1,500 stylists in mid-2000, which we think was driven in part by Direct Buy).
- EBIT Margin: Was -3% in ’20 (FY End Aug) due to the aforementioned dilutive growth initiatives, plus a 200-300bp hit from Covid’s lower spending per customer. Longer-term, a -3% margin is about right as it relates to where this model is headed.
- EPS: The EPS profile of 2016/17 when it solely served its 1.5mm core Fix customers is gone forever. It has a shot this year to be EPS positive as we see a snap back in high-end apparel spend, but as the more powerful business model changes occur over a TAIL duration, the company falls into the ‘will never make money – ever’ bucket. That’s a massive variance from the consensus, which has SFIX getting back to peak EPS by FY25.
- Inventory Turns: Should get cut from about 7x today to sub 4x by the end of our model. Instead of selling the consumer the inventory SFIX has, it needs to stock apparel that consumers might or might not want.
- Net Debt/EBITDA: We think the company will need to tap capital markets over the next 2-3 years to fund meaningfully higher working capital requirements. If the new CFO from Amazon is everything the market thinks he is, he should do a deal soon – from a position of strength -- while the company benefits from the reopening and return to upscale apparel. We’ve got net/EBITDA going from zero today to 3.0x in our model.
SFIX | KEY MODELING ASSUMPTIONS
VALUATION AND TIMING CONSIDERATIONS
As we see revenue decelerate despite aggressive customer acquisitions, we think that people will finally abandon EV/Sales as the barometer here and look at real financial metrics like EBITDA, margins, earnings (though there won’t be any), capital intensity, leverage, ROIC and ROE. The stock is currently trading at over 200x EBITDA – clearly a metric people don’t care about today. But as the model plays out – especially as debt is added to the equation -- we think that the EBITDA multiple will come into focus. Are we being aggressive with an 8x multiple for this business? It implies 90% downside in the stock. So yeah, we’d call that aggressive. But based on the underlying financial health and trajectory of the business, we think it’s perfectly fair.
Timing on this name is critical. The fact is that we’re early in making this call, which is why it sits on our Short Bias list – i.e. not a Best Idea yet despite the massive downside we see in the name over a TAIL duration. We’re likely to see a sharp rebound over the next two quarters in both customers added and spending per customer. Importantly, this will likely come at a lower customer acquisition cost given the sheer need for the product as people have to abandon the sweat pants and hoodies and start dressing like adults again – so the potential to actually print a profit is real, albeit temporary. Short interest on the name is at 16% -- roughly at all-time lows and down from 40% over just two months. Clearly, the shorts have capitulated after getting beat up by a 6-bagger over the past year. But the same way we think that the company should tap capital markets in the coming two quarters is the same time we’d be shorting it aggressively. Once we’re out of Macro Quad 2, the same time we de-risk reopening strength in demand, is when this name has Best Idea Short written all over it.
