🎢 #48 - Zara meets Netflix
🎢 #48 - Zara meets Netflix
+ cool readings like Product Hunt strategy, how to bootstrap, and more
Hey there! Welcome to my email newsletter. My name is Leo Luo, a student entrepreneur at the University of Michigan. I write about startup stories, consumer trends, and unique behaviors in the early-stage B2C space.
All my previous posts can be found here.
🔥 Startup Story
Zara meets Netflix
(Image credit: Finesse)
Recently, I have been learning about machine learning (ML) through Andrew Ng’s deeplearning.ai. I am fascinated by just how much these algorithms have changed our perception of analytics and decision-making. I am now even more bullish on how machine learning can be leveraged to revolutionize many industries. A few weeks ago, I wrote about the disruptive power of ML in the music industry and how Boomy is using it to help more beginners create and monetize their own music.
Similar to music, fashion is an industry historically untouched by radically disruptive technologies. That’s about to change. Finesse is a fast-growing fashion startup that uses ML to predict fashion trends and produce hit sell-outs every time they release a line. I had the opportunity to speak with Ramin, the founder of Finesse, to learn more about his vision, or as he terms it, ‘Zara meets Netflix’.
🌱 Genesis
“I grew up as a gay non-binary brown kid in the south of Germany, and I didn't see another brown person until I was 16. Fashion was my way of fitting in. It's culturally relevant to whatever identity you're trying to express. It was a way for me to pretend I was straight or white as I was exploring my identity. Eventually, I went to a boarding school in the UK and started to become more comfortable in who I am, and so I started to wear more female-leaning clothing to express my identity. Fashion is an incredibly powerful and incredible asset. And the best part is that we all have access to it.”
Unique insight -
“However, the industry itself hasn’t been touched by technology in a very long time and is super inefficient. Fashion is set up to allow you to produce merchandise cheaply without really knowing what will actually sell. It's like throwing spaghetti at the fridge and hoping it sticks. This is also the number one reason behind fashion being the second-largest polluter in the world - 85% of everything produced ends up in landfills.
(Image credit: Common Object)
On the other hand, there's access to data that we haven't had for a very long time. We have social media data that is way more expressive than any data we’ve had before. With social media, consumers are telling us exactly what they want. From a computational perspective, this data also provides us an opportunity to address inefficiencies in the industry.”
🚗 Product Journey
First experiment -
Ramin’s first attempt at Finesse was very different from what Finesse is today. The initial experiment was a social e-commerce app where people could post fits and make money through affiliate links. However, Ramin decided to pivot away from a platform-based to a product-based company.
“We knew there was a huge opportunity in fashion, but we didn’t know what course to take. Our first experiment was a video-driven social commerce app focused on the fashion community. Within just a couple of days of the launch, there was a user segment in the UK that kept using us, and they loved talking about which fits they liked. They wanted to vote on them. We observed this phenomenon on other platforms as well, such as Instagram and Reddit communities.”
Pivot -
“I took a step back and felt that this was a strong signal I couldn’t ignore. I thought back on my experience as a shopper passionate about fashion, but always let down by the way the industry operates.
I looked into the industry and found that the tech stack was absolutely abysmal. I looked specifically at the 2018 H&M and Zara reports - there were 80 pages, but only one page talked about technology, and it was about IT.
The idea came after. We thought ‘let’s not create a platform, let’s create the product’. The timing was also right because it had become possible for anyone to find and vet manufacturers through companies like Alibaba.”
MVP -
The team at Finesse developed a two-part ML algorithm that first identifies what fits to design, and then finds the right influencer-audience duo to collaborate with for the drop. The algorithm does so by scanning across social media for signals.
“For example, Kylie Jenner posts an image on Instagram and people go crazy in the comment section saying, ‘love this fit’, ‘fire’, ‘where can I get this from’. We treat these data points as a signal in the same way that a quant trader on Wall Street would treat any kind of signal as a signal to trade on.
The only difference with us is that our asset class is fashion instead of stocks, and our quantity of investment isn't directly tied to a financial amount, but rather how many units we produce. Our data is more expressive and less ephemeral than the stock market, which is, in essence, a bunch of numbers jumping up and down.”
[3D Sampling]
Besides the ML algorithm, another key part of Finesse’s tech stack is 3D models. They use 3D-modeled samples during the early stages of garment development, which not only reduces fabric waste but it also streamlines the production process.
(Image credit: Finesse)
“Our whole sampling process is completely done in Clo 3D, so we can do what usually takes months in a day. Our main partner factory overseas also uses Clo, which makes it easier for us to translate misalignments between us. At the end, it all comes down to getting the fabrics and cutting them together. This allows us to produce things within a 25-day timeline when others are still stuck on a five-month timeline.”
Finesse launched their first apparel drop in 2019 and collaborated with an influencer identified by the algorithm - the drop sold out within 4 to 5 hours. They did another drop shortly thereafter and the drop sold out again within 45 hours. These initial successes have given Ramin and his team the confidence to go on to fundraise and build out a team.
(example of influencer collab)
🤔 Challenges
1. Working with legacy players
“This is a very interesting, but incredibly outdated space so we hit a lot of challenges in terms of who we work with. We’re very lucky to have found partner factories that work with us at the scale and the speed that we want to, but we had to do a lot of the legwork in the beginning to explain to them how we do things and how we could save them time and money.”
2. The lack of diversity
“I am a diverse founder, and I wish we could see more investors and founders in the industry from a more diverse background, but unfortunately, this industry is very homogeneous. It's not necessarily a problem. It's more something that I wish the industry would have because when it comes to explaining how important diversity is for a product today, it's very easy to do that to another minority community.
All of our investors are minority investors and that has been great for us to bounce off ideas and figure out what we want to be as a consumer product. Every consumer product nowadays needs to have diversity ingrained in them.”
🚀 Vision
“The way that we think of scaling is not necessarily the way a typical fashion company scales which is making more SKUs, but also not in the way that Supreme has scaled which is creating more and more developed drop structures. We see ourselves scaling somewhere in between by leveraging technology.
This is where the Zara meets Netflix analogy comes in. Imagine you're going on our website, and you see the four drops that we think you're gonna like. This is how we see ourselves grow as this Zara meets Netflix analogy comes true.”
Check out Finesse!!
👨💻 What I’ve been reading
How Superpowered (YC W21) leveraged Product Hunt to get hundreds of paying customers (shout out to my friend Ali for writing this amazing article!)
How startups should think about their capital efficiency using the Burn Multiple according to David Sacks - hint, you want to aim for <1X and keep it <2X
The evolving commerce stack - from the storefront, checkout/cart, to backend subcategories
How an audio learning app bootstrapped itself to 10K users - marketing playbook for private Beta, public Beta, and product launch
😍 Jobs & Internships
Full-time:
General Catalyst - Early Stage Investment Analyst (SF, Boston, NYC)
Precursor Ventures - Analyst (Bay Area)
Umamicart - Biz Ops Analyst (NYC)
M13 - Growth & Data Strategy Analyst (NYC)
Belong - Operations Associate (Remote)
Neeva - Product Manager (Remote)
Internship:
Silicon Valley Bank - VC Intern (Bay area)
Riot Ventures - VC Intern (Remote)
Compound VC - Research Intern (Remote)
Flexport - PM Intern (SF)
Daily Harvest - Innovation Commercialization Intern (NYC)
Twilio - Developer Education Intern (Remote)
👀 How did you like today’s Consumer Startups?
Loved | Great | Good | Meh | Bad
Stay steezy team - see you next Sunday!