SFIX - Stitch Fix

[TBW]

I think this is a bit of rorschach test. I have had a small long for a while now, and can somewhat see both sides. In summary, I think SFIX is doing better than you think but unless the new initiatives work out, scepticism here is warranted.

 

To start, SFIX business has been to mail clothes to a person who then keeps or sends it back. The secret sauce, if it exists and I think it does, is the AI that selects the clothes. The market for people that want this service is somewhat limited and I think that, and the pandemic, explains the lack of scale on the ad spend side.

 

But now SFIX is branching out and is creating personalised online shopping where the AI selects what you see. The big idea is that online shopping can be more interesting and a chance to find new stuff like the traditional store browsing experience, that due to the AI you are picking from items that you are likely to like, this is different than the current online model where typing in socks into amazon yields many generic options and you tell the store what you are looking for. The market for this is magnitudes higher than the current market SFIX is targeting. There is also adjacent areas where other brands may pay SFIX to advertise amongst their curated offerings. The lack of scaling in SGA costs is due to these investments. This is similar to many tech companies where capex flows through income statement. There is other interesting 'flywheel' type aspects in their relationship with suppliers, but this is too long already to get into that, but there are many aspects to their business model that I think are really interesting.

 

Management at SFIX has been excellent. The whole operation has been mostly self funded with all money raised to date mostly sitting in cash on the balance sheet. The CEO, who I think is awesome, has a thorough deliberate approach to rolling out new projects constantly testing along the way. That also is part of the lack of scaling as there is a few year gap between capex and product roll-out.

 

From this perspective, the SGA and ad spend may not be as bad as it seems as the ad spend is building the brand and making the new 'platform' that they are building bigger when it is launched. Also the valuation isn't terrible at 2x 2022 sales (ex cash), given their recent growth, high gross margins, and future growth potential (altho I would say that long term 11% EBIT margins don't really justify high p/s ratios...).

 

If however, the new initiatives aren't successful then everything you pointed out is a real concern.

 

Other things I don't like are SBC is super high. I think it's annoying that companies exclude SBC from ebitda and speak about being profitable.

 

But it is overlooked that working capital helps fund this business.

 

Taken all together I come on the side that SFIX is not that expensive at its current price, and it is a company that could be significantly bigger in the future with an undemanding valuation (especially relative to most tech companies).

 

With some many large and dubious tech companies, that unlike SFIX are completely reliant on capital markets, I do find it strange that there is so much short attention on this name. So I do wonder what I am missing, so if there are other bear cases I should consider I would love to hear them.

[KJP]

I did leave out some possible explanations for the lack of SG&A scaling.  As you note, they have been developing "direct buy," which if successful ought to raise margins by cutting out "stylists," i.e., replacing them with artificial intelligence.  They also have been expanding by category (women to men to kids to plus size and so on) and geography, so it's possible that they have some profitable scaled businesses subsidizing losses in the "start up" sections of the business.

 

So, I understand that there are plausible future worlds in which the company more than grows into its valuation.  But going back to my original question:  Is there anything in the historical financial statements of the company that would provide evidence that the optimistic outcome is likely?  Or instead is the current fundamental bull case built on a belief in Katrina Lake and the power of artificial intelligence to remake shopping for clothes (i.e., traditional online shopping with better and individualized curation)?

 

Note that I'm not saying such a bull case is wrong or unsound or anything else.  I'm just trying to pin down what inferences can reasonably be drawn from the historical financials relative to qualitative assessments of management, etc.

 

Regarding current valuation, I agree with you that it's not completely outlandish.  For example, let's assume the suggested 11% EBIT margin and account for taxes, making that about an 8.5% net margin.  At 20x earnings they'd need about $3 billion in sales to justify the current market cap ($5 billion/20 = $250 million/.085 = $3 billion).  Of course, I don't think the current model, with its significant non-scalable SG&A, would come close to those margins at $3 billion in sales.  But there is at least a coherent case for margin improvement and sales growth that could get you there.

 

EDIT:  A few more questions:  You referred to CapEx through the income statement.  I hear that alot, and it's certainly true to an extent.  For example, the salary of a data scientist or engineer you're paying to build a new digital offering gets expensed, whereas the construction costs of a brick-and-mortar factory get capitalized.  But how much of Stitch Fix's SG&A is actually this type investment?  More broadly, how do outsiders determine whether these types of claims by management (of Stitch Fix or any other business) are true?  Is it an assessment of who is being hired and the cadence of new product rollouts?

 

[TBW]

I think those are very good questions and ones I have for many SaaS type companies.

 

The simplest, but largely accurate, answer is that investors don't know.

 

This is far from a usual investment for me, so I have struggled with some of the same questions, but I came away comfortable given the history. This is a company that has been very good with capital, that usually goes hand-in-hand with decent business sense. Operating within their means is very different ethos than most tech companies, and thus I think the capex spend is largely reasonable too. I also think that SGA cost increases since IPO largely have to be new investments, as what else is it? SFIX leases its facilities, it's just people, computers and data storage, in terms of costs. The big unknown to me is how much shipping costs run through SGA. The new business should see far less in terms of returns and that should help overall margins.

 

In my mind this is a bit of a story stock, where I happen to believe the story. Previously it was cheap, I thought, for the potential upside 'direct buy' had, and the likely stable, but growing, economics of the current business model. I don't know that there is direct proof that SGA difference is capex, as I stated, but I haven't been convinced of the converse yet either.

[DTEJD1997]

Hey all:

 

I have not been following this too close, but I do online sales.

 

There are some BIG changes/problems coming to online sales.

 

Packaging costs (boxes, tape, foam peanuts) are going up.  Some sub-categories have gone up 4 times in the past year or so.  The price increases have been small each time, but if something goes up in cost 2 or 3 or 4 percent each time, it adds up.  My suppliers have told me to expect more of the same.

 

Shipping costs, especially at USPS have gone up.  This is especially true on smaller, light weight packages.

 

Perhaps the biggest problem of all has been labor.  Here in Detroit, the Post Office is breaking down.  A lot of postal employees are single mothers/parents.  When Covid-19 hit, and skools closed, who is going to look after the children?  They are stuck at home.  So huge numbers of postal workers are not able to come into work.  My local post office sometimes does not even up to the public.  Most days, most of the front counter clerks do not show up.  So the management has to man the counter.  Mail is not being delivered on every route, on every day.  Some routes have gone to DELIVERY EVERY OTHER DAY.

 

Perhaps the worst of all, is that the distribution centers (for outbound packages) are completely overwhelmed.  Packages will come in and not get sent out for 3-4 days, sometimes the wait time to go out is OVER a week.  This is a bottleneck at the distribution center, which results in ultimate delivery times of sometimes 2 weeks.

 

To be fair to the Post Office, I have had tremendous delays, but no package has been lost OR not ultimately delivered.

 

I have now had to move a good percentage of my shipping to Fed-Ex or other providers.  Frequently, the cost is higher than USPS.  Fed-Ex, so far, has done a fairly good job, but even they have delays.  The delays at Fed-Ex are no where near as bad at USPS, but they are there.

 

I think ALL delivery services are going to have labor issues moving forward.  The cost of labor will be going up, and the price of delivery is going up.

 

Obviously SFIX gets a better deal than I do....but rates for EVERYBODY will be going up.  Couple that with increased cost of packing materials.  That is going to be a difficult head wind going forward.

[KJP]

What is the bull case here, and can it be justified by the financial results to date? 

 

The business is currently slightly unprofitable but claims to have a big data advantage that creates a highly scalable business.  [see here:  https://investors.stitchfix.com/static-files/2b87b6c8-0ca7-4c48-a8de-b024ac0da5ef]  If that were true, margins ought to expand as the company grows.  But the reverse appears to be happening on every metric.  For example, SG&A ex-advertising has increased from 32% of revenue in 2016 (and in 2018 for that matter) to 37% of revenue in 2020, despite sales growing from $730 million to $1.7 billion.  [Note that I refer throughout to fiscal years, which end in July.]  Drilling down a bit more, from 2018 to 2020, SG&A ex-advertising increased from $389.9 million to $637.2 million, or about a $250 million increase.  The 10-Ks say that the increase primarily comes from additional compensation for data scientists and engineers.  But over those two years, the company added only 60 engineers and 45 data scientists.  Those 105 additional employees can’t be the main driver of $250 million of increased costs. 

 

I could not find any detailed breakdown of SG&A in the 10-K.  But they did disclose that from 2018 - 2020 they added at least 1,000 additional “stylists” and fulfillment employees.  Importantly, the work of these employees does not appear to be nearly as “scalable” as the work of a data scientist.  Instead, they appear to be much more of a variable cost, unless the stylists are eventually replaced by artificial intelligence.

 

Advertising expense also appears to be becoming less efficient as the company has to push harder to identify new customers.  I didn’t see disclosure about churn, so here are the numbers on total advertising spend and year-over-year change in active customers:

 

2016:  $25 million; 807,000

2017:  $70.5 million; 520,000

2018:  $102.1 million; 548,000

2019:  $152.1 million; 494,000

2020:  $167.8 million; 286,000

 

Every year they are starting from a larger base and thus larger absolute churn, so they have to spend more absolute dollars each year just to stand still.  But I don’t know the churn numbers, so I cannot determine the number of “new” customers and the unit economics of each.  On their face, however, those advertising numbers suggest to me that for several years the company has been finding it harder and harder to grow efficiently. 

 

So, we have a company that is around breakeven or slightly unprofitable as things stand.  The company would have us believe that scale will solve that, but the line items that ought to be scaling (SG&A ex-advertising) are going in the wrong direction and advertising spend appears to be getting less efficient.  Moreover, in the company’s own “long-term model,” it projects ~11% EBIT margins, essentially all of which it projects to come from SG&A ex-advertising going from 37% of revenue to 25% of revenue, exactly the same line item that, as discussed above, appears to be going backwards and appears to have significant components that are not particularly scalable.  [see slide 18 of the presentation linked to above]

 

Despite all of that, the stock soared to what I believe is an all-time high today because they had a good quarter of adding new clients (but no margin improvement) and a rosy revenue projection for 2021.  So, I’m back to where I started:  Can the existing public information justify paying $50/share for this?  If so, what am I missing?

 

SFIX went to over $113/share shortly after I suggested it was expensive at $50.  This will not go in the pitchbook for my future long/short fund. 

 

Now one quarterly report later it's back to $53 after hours today.  I don't get what caused the massive buying on the way up or the reversal.  It looks like the same company to me.  Slide 13 of today's presentation illustrates what has always concerned me about this company:  https://investors.stitchfix.com/static-files/680c14b5-3676-4507-a810-921ff538da97 

[Borgesian]

What is the bull case here, and can it be justified by the financial results to date? 

 

The business is currently slightly unprofitable but claims to have a big data advantage that creates a highly scalable business.  [see here:  https://investors.stitchfix.com/static-files/2b87b6c8-0ca7-4c48-a8de-b024ac0da5ef]  If that were true, margins ought to expand as the company grows.  But the reverse appears to be happening on every metric.  For example, SG&A ex-advertising has increased from 32% of revenue in 2016 (and in 2018 for that matter) to 37% of revenue in 2020, despite sales growing from $730 million to $1.7 billion.  [Note that I refer throughout to fiscal years, which end in July.]  Drilling down a bit more, from 2018 to 2020, SG&A ex-advertising increased from $389.9 million to $637.2 million, or about a $250 million increase.  The 10-Ks say that the increase primarily comes from additional compensation for data scientists and engineers.  But over those two years, the company added only 60 engineers and 45 data scientists.  Those 105 additional employees can’t be the main driver of $250 million of increased costs. 

 

I could not find any detailed breakdown of SG&A in the 10-K.  But they did disclose that from 2018 - 2020 they added at least 1,000 additional “stylists” and fulfillment employees.  Importantly, the work of these employees does not appear to be nearly as “scalable” as the work of a data scientist.  Instead, they appear to be much more of a variable cost, unless the stylists are eventually replaced by artificial intelligence.

 

Advertising expense also appears to be becoming less efficient as the company has to push harder to identify new customers.  I didn’t see disclosure about churn, so here are the numbers on total advertising spend and year-over-year change in active customers:

 

2016:  $25 million; 807,000

2017:  $70.5 million; 520,000

2018:  $102.1 million; 548,000

2019:  $152.1 million; 494,000

2020:  $167.8 million; 286,000

 

Every year they are starting from a larger base and thus larger absolute churn, so they have to spend more absolute dollars each year just to stand still.  But I don’t know the churn numbers, so I cannot determine the number of “new” customers and the unit economics of each.  On their face, however, those advertising numbers suggest to me that for several years the company has been finding it harder and harder to grow efficiently. 

 

So, we have a company that is around breakeven or slightly unprofitable as things stand.  The company would have us believe that scale will solve that, but the line items that ought to be scaling (SG&A ex-advertising) are going in the wrong direction and advertising spend appears to be getting less efficient.  Moreover, in the company’s own “long-term model,” it projects ~11% EBIT margins, essentially all of which it projects to come from SG&A ex-advertising going from 37% of revenue to 25% of revenue, exactly the same line item that, as discussed above, appears to be going backwards and appears to have significant components that are not particularly scalable.  [see slide 18 of the presentation linked to above]

 

Despite all of that, the stock soared to what I believe is an all-time high today because they had a good quarter of adding new clients (but no margin improvement) and a rosy revenue projection for 2021.  So, I’m back to where I started:  Can the existing public information justify paying $50/share for this?  If so, what am I missing?

 

SFIX went to over $113/share shortly after I suggested it was expensive at $50.  This will not go in the pitchbook for my future long/short fund. 

 

Now one quarterly report later it's back to $53 after hours today.  I don't get what caused the massive buying on the way up or the reversal.  It looks like the same company to me.  Slide 13 of today's presentation illustrates what has always concerned me about this company:  https://investors.stitchfix.com/static-files/680c14b5-3676-4507-a810-921ff538da97

 

It was a short squeeze that fueled the prior run, short interest was around 40% before the last earnings report.

[KJP]

What is the bull case here, and can it be justified by the financial results to date? 

 

The business is currently slightly unprofitable but claims to have a big data advantage that creates a highly scalable business.  [see here:  https://investors.stitchfix.com/static-files/2b87b6c8-0ca7-4c48-a8de-b024ac0da5ef]  If that were true, margins ought to expand as the company grows.  But the reverse appears to be happening on every metric.  For example, SG&A ex-advertising has increased from 32% of revenue in 2016 (and in 2018 for that matter) to 37% of revenue in 2020, despite sales growing from $730 million to $1.7 billion.  [Note that I refer throughout to fiscal years, which end in July.]  Drilling down a bit more, from 2018 to 2020, SG&A ex-advertising increased from $389.9 million to $637.2 million, or about a $250 million increase.  The 10-Ks say that the increase primarily comes from additional compensation for data scientists and engineers.  But over those two years, the company added only 60 engineers and 45 data scientists.  Those 105 additional employees can’t be the main driver of $250 million of increased costs. 

 

I could not find any detailed breakdown of SG&A in the 10-K.  But they did disclose that from 2018 - 2020 they added at least 1,000 additional “stylists” and fulfillment employees.  Importantly, the work of these employees does not appear to be nearly as “scalable” as the work of a data scientist.  Instead, they appear to be much more of a variable cost, unless the stylists are eventually replaced by artificial intelligence.

 

Advertising expense also appears to be becoming less efficient as the company has to push harder to identify new customers.  I didn’t see disclosure about churn, so here are the numbers on total advertising spend and year-over-year change in active customers:

 

2016:  $25 million; 807,000

2017:  $70.5 million; 520,000

2018:  $102.1 million; 548,000

2019:  $152.1 million; 494,000

2020:  $167.8 million; 286,000

 

Every year they are starting from a larger base and thus larger absolute churn, so they have to spend more absolute dollars each year just to stand still.  But I don’t know the churn numbers, so I cannot determine the number of “new” customers and the unit economics of each.  On their face, however, those advertising numbers suggest to me that for several years the company has been finding it harder and harder to grow efficiently. 

 

So, we have a company that is around breakeven or slightly unprofitable as things stand.  The company would have us believe that scale will solve that, but the line items that ought to be scaling (SG&A ex-advertising) are going in the wrong direction and advertising spend appears to be getting less efficient.  Moreover, in the company’s own “long-term model,” it projects ~11% EBIT margins, essentially all of which it projects to come from SG&A ex-advertising going from 37% of revenue to 25% of revenue, exactly the same line item that, as discussed above, appears to be going backwards and appears to have significant components that are not particularly scalable.  [see slide 18 of the presentation linked to above]

 

Despite all of that, the stock soared to what I believe is an all-time high today because they had a good quarter of adding new clients (but no margin improvement) and a rosy revenue projection for 2021.  So, I’m back to where I started:  Can the existing public information justify paying $50/share for this?  If so, what am I missing?

 

SFIX went to over $113/share shortly after I suggested it was expensive at $50.  This will not go in the pitchbook for my future long/short fund. 

 

Now one quarterly report later it's back to $53 after hours today.  I don't get what caused the massive buying on the way up or the reversal.  It looks like the same company to me.  Slide 13 of today's presentation illustrates what has always concerned me about this company:  https://investors.stitchfix.com/static-files/680c14b5-3676-4507-a810-921ff538da97

 

It was a short squeeze that fueled the prior run, short interest was around 40% before the last earnings report.

 

Right.  Spek previously mentioned a high short interest.  I thought the covering might have helped propel the stock from $35 on 12/7 (before the last earnings report) to ~$70 on 12/22.  The price then came back down to $55-$60 for a few weeks, and then shot from $56 on January 11 to $106 on January 27.  I don't have a good source for historical short interest through time.  Were both of the very big, quick moves ($35 - $70 from December  7-22, and then $56 - $106 from January 11 - 27) both short covering?

 

EDIT:  Nevermind my question.  Nasdaq publishes the bimonthly data:  https://www.nasdaq.com/market-activity/stocks/sfix/short-interest

The immediate post-earnings move doesn't look like huge short covering, but there was a big drop in short interest between 1/15 and 1/29, which overlaps with the second blowout.

[KJP]

Since I've been bad-mouthing Stitch Fix, some balance may be in order.  Here's a podcast discussing a bullish perspective:  https://podcasts.apple.com/us/podcast/special-surviving-20-years-stitch-fix-deep-dive-mario/id1526125544?i=1000515864338

This particular bull case is based on changes to Stitch Fix's traditional business model (direct buy, previewing selections before they are mailed) and faith in the company's data science efforts.

I'm not sure this particular bull disagrees with much of what I've noted, because he acknowledges that the bull case cannot be derived from historical financials.

Edited April 7, 2021 by KJP