Rohan Deuskar

Rohan Deuskar runs Stylitics, a fifteen-year-old New York software company that most shoppers have interacted with without knowing it. The Stylitics widget is the thing that suggests the belt for the pants and the pants for the jacket on hundreds of retailer websites: Macy's, Walmart, Kohl's, Puma, Revolve. It appears roughly 140 million times a month. It is not a consumer brand and does not want to be.

The business sells to merchandisers, not shoppers. A retailer with 40,000 SKUs and a small styling team pipes its catalog into Stylitics; Stylitics returns outfits, bundles, and "shop the look" carousels tagged to real inventory. Machine learning does the pairing, computer vision does the tagging, human stylists set the taste. Deuskar's pitch to the buyer is unromantic and easy to underwrite: more items per basket, more categories discovered, more revenue per session. He has said the company now does more revenue in a single week than it did in its first five years combined.

The founding sentence is the tell. Deuskar was not a fashion insider. He was a Wharton MBA who kept buying variations of the same shirt and thought a database could help. The first version of Stylitics tried to solve that problem for consumers. The current version solved it for the stores selling the shirts. Same code, roughly. Different customer. Very different economics.

Founders talk about pivots the way sailors talk about storms - after the fact, with structure. The Stylitics pivot took years. The company launched in 2011 as a digital closet, a place to upload your clothes and get outfit suggestions. It picked up press. It did not pick up a business model. Deuskar has said publicly that the company nearly went bankrupt before figuring out that the real customer was the retailer, not the shopper.

The move from B2C to B2B is not glamorous. It is spreadsheets and demo decks and quarterly reviews with buying offices. It is the un-fun part of fashion tech. It also, in Stylitics' case, worked. By 2020 PeakSpan Capital led a growth round. By March 2022, PSG wrote an $80 million Series C check, bringing total funding to approximately $105 million. The company reports roughly $40.6 million in ARR and 110 employees.

The lesson, if there is one, is patience. Most consumer-fashion startups from the 2011 cohort are dead. The ones that survived either got acquired for parts or, like Stylitics, quietly walked over to the enterprise side of the room and started billing per SKU.

Deuskar grew up planning to be a generalist. Northwestern for undergrad, economics, 2000 to 2004. A stint at a Chicago mobile messaging startup - the kind of early-2000s tech job that taught operators how to sell into carriers before smartphones ate everything. Then Wharton, 2009 to 2011, for an MBA in entrepreneurial management. He rowed for the Wharton Crew Team and ran through the Entrepreneurship Club. Between years, he interned at Amazon, which is where the idea he still runs on today first arrived. Retail, data, and technology were starting to collapse into one thing. Amazon had already figured it out. Fashion had not.

That gap - Amazon-level data discipline arriving in fashion late - is the entire Stylitics thesis. Deuskar built the company around the belief that a merchandising team could not manually style tens of thousands of SKUs, but a machine could, if you fed it enough taste. He co-founded the company with Zach Davis, still on the board. In 2014, the City of New York named him to the inaugural class of NYC Fashion Fellows, a program built to keep operator talent in the New York fashion ecosystem.

The nominal product is outfit recommendations. The actual product is discovery. Deuskar's argument to a retail buyer is that a product detail page in isolation converts poorly; a product wrapped in an outfit converts better, and each outfit surfaces two or three additional items the shopper had not searched for. Average order value goes up. Category exposure goes up. Return rates, in the retailers' own reporting, tend to go down.

Stylitics stitches all of this into the retailer's existing site with the retailer's own inventory and pricing. The engine handles style profiles, "shop the model," email personalization, and marketing content. There is a stylist team on the backend calibrating the machine, so it does not decide that a wedding guest wants a fleece vest.

Deuskar keeps a low profile for a founder who has raised nine figures. His Twitter handle is @RohanD, dormant enough that most of what you learn about him has to come from long-form interviews and Wharton alumni features. He has written on Medium and bylined at Ad Age. He is an angel investor in Obsess, a virtual-store platform, and he is a member of YPO Manhattan. He has, by his own admission, never been the most stylish person in any room he has ever been in - a fact he uses as a kind of credential when explaining why he thought outfitting needed software in the first place. If you know what you own and cannot dress yourself, you are the perfect customer for a machine that can.

The Stylitics roadmap that Deuskar has described publicly reads like the interior of every fashion catalog turning into a Pinterest board with a checkout button. More AI-generated outfitting. More "shop the model" and shoppable galleries. More content automation for retailers who do not have the internal team to produce it. On the technology side, the company's stack now leans on modern LLMs - Anthropic Claude, OpenAI, Gemini appear in the company's public technology signals - which suggests the next iteration of the styling engine will be less rule-based and more model-based.

The bet is the same one Deuskar has been making since Wharton. A closet with a brain, at retailer scale. Everything else is execution.