5xing the worst biz ever: Stitch Fix

🧠 The Takeaways

Today, we’re laying out the turnaround blueprint for the worst DTC model of the last 20 years, Stitch Fix.

1. ChatGPT their style quiz.

2. Outsource all Operations.

3. 80% RIF.

+ Any interest in my LLM research assistant?

LBAB! Community - Do you want to see my Research?

I built an LLM research assistant + use it weekly for LBAB! to:

1) Find key info.

2) Figure out my ideas.

3) Test my hypotheses.

I use it as my research assistant. 

Do you want direct access to it?

Reply and let me know.

You'd get access to:

- The questions I ask to break down each business

- The analysis framework I use

- The thought process behind identifying opportunities

If there's interest, I'll start including them in the newsletter.

Let’s Examine This Biz

Note: As always, none of what follows is legal, tax, investing, financial, or any other sort of advice. And I was never here.

Stitch Fix, the “Personal Stylist delivered to your door” fashion box subscription biz, is dead.

• Share price: $4

• Market Cap: $564m

• L5 Performance: -87%

• P/E Ratio: N/A

After multiple pivots, they have the worst biz model I’ve ever seen for a public company. 

They have insanely high human + tech + operational costs. And want to trade like a tech stock. 

What started out as a smart consumer value prop is a disaster of a biz that won’t admit defeat.

Financial Summary

2024 Financial Statements (YoY Comparison)

Sales: $1.3B (-18%) 👎👎
Gross Profits: $592m (-14%) 👎
OPEX: $725m (-17%) 😰

Net Income: -$128m (+23%)  🤮

Link to Company’s earnings

TLDR Analysis: Debt is their only lifeline.

• 4th straight year of Rev decline. ⚰️

• Clinging to 44% Gross Profits. 😰

• Net Income is still 9 figs in the negative. 😰🤮⚰️

A $1B topline biz should be producing $100m in Net Income a year.

This is pretty simple math. 

OPEX can’t be $725m when Gross Profits are $592m. SG&A spend (other than $111m towards marketing) is $614m.

They’re unable to make more than they spend.

The biz model has completely broken down, proving their original hypothesis wasn’t that valuable. And its stock performance tells you everything you need to know. 

Let’s TLDR This Biz

Founding:

• Katrina Lake started Stitch Fix in 2011 at age 28 while getting her MBA at Harvard Business School

• Used SurveyMonkey to collect style preferences

The "Aha" Moment:

• Busy people wanted personalized shopping without the time investment of browsing stores / websites

• Marrying human styling expertise + algorithmic personalization became their secret sauce

Explosive Growth:

• Reached $730 million in sales by 2016-17

• Hit peak of 4m active clients in Q1 of fiscal 2022

• Stock reached ATH of $106.41 in January 2021

The Model:

• Customers complete detailed style profiles. $20 styling fee (credited toward purchases)

• Algorithms and human stylists collaborate to curate personalized, 5-item "Fixes"

• Customers try items at home for 3 days, keep what they want, return the rest

The Collapse:

• 2021: Lake steps out of the CEO role

• 2022: Stock plummets 95% from ATH

• 2022-2023: Multiple RIFs

The stock hasn’t recovered since.

Let’s Fix This Biz

Here are the 3 ways we’d gut costs to give them a chance of survival.

1) Rip out the Style Quiz and put ChatGPT on the site.

Brands love onsite embedded quizzes that are basically just branded SurveyMonkey/Typeform surveys. 

They were a huge step up in “personalization” in the 2010s because customers were willing to tell you what they’d buy inside of fun quizzes.

But just like Buzzfeed, these UXs died in 2022.

Customers want to chat with a humanlike experience instead of clicking 50 buttons over 10 minutes to get a couple recommendations.

The fact Stitch Fix wasn’t the first biz to jump on this shows how asleep at the wheel they are.

They employ thousands of Stylists who’ve spent 10 yrs having real conversations helping customers find the right products.

This is the easiest + most obvious “train a GPT + embed it on your site” use case.

1. Train a chatbot on the millions of conversations + billions of data points from previous conversations.

2. Embed it on your site.

3. Route all traffic through it.

If they really wanted, they could have an escalation where a stylist takes over. 

We can set this up over a month and rip out 80% of their tech costs.

Takeaway: Always be ahead of the core tech trends driving the biz.

2) Outsource All Operations.

Behind Stylists, Operations is far and away the largest team, with 970 (23%) team members.

Stitch Fix doesn’t break out their Operations/Fulfillment costs in the 10-K. If they’re not good enough to brag about, they’re horrible.

• In 2024, Stitch Fix turned all of their stylists into freelancers. 

• Now it’s time to turn their Ops into a 3PL.

This is a losing game for a biz with so many other costs. Their core model (try on at home + return anything) is a money pit operationally.

They need to either stop taking returns or find way cheaper ways to support this model.

Takeaway: Always outsource what you aren’t the best at.

3)  RIF 80% of the workforce.

Technology should have already reduced headcount + pushed them to profitability.

1. They never delivered on the real value of removing Stylists from the process by creating the AI stylist.

2. The recommendations aren’t tailored enough to reduce returns (the largest fashion issue this model was supposed to address).

3. What do 236 Devs work on all day for a simple site, survey funnel, + customer portal? 

Every department must be smaller. 

If we 100% outsource Fulfillment and Operations, we could run this biz on <1k employees with little Rev impact.

For an SF-based biz, seems like they missed the memo on the 2022 efficiency playbook.

Takeaway: Always do more with less.

Final Thought

For the young, hungry entrepreneur out there, this seems like the perfect biz to recreate in the AI era.

It’s a much simpler biz than Stitch Fix has made it out to be:

1. Create a GPT that interacts with customers + recommends products (real value here).

2. Spin up a Shopify subscription site to process orders.

3. Work with a 3PL to deliver products.

Don’t make it a subscription service.

The key is nailing the conversation the AI has with customers to ensure it recommends products customers actually want + don’t return.

That company can be built with 20 people, and I bet it could hit $1B in sales within 5 yrs.

The majority of spend would be on CAC.